Research on Relationships among Different Distress Types ...

Research ArticleResearch on Relationships among Different Distress Types ofAsphalt Pavements with Semi-Rigid Bases in China UsingAssociation Rule Mining A Statistical Point of View

Jing Li Guoqiang Liu Tao Yang Jian Zhou and Yongli Zhao

School of Transportation Southeast University Nanjing Jiangsu 210096 China

Correspondence should be addressed to Yongli Zhao yonglizhao2016126com

Received 8 November 2018 Accepted 26 December 2018 Published 11 February 2019

Academic Editor Xu Yang

Copyright copy 2019 Jing Li et al +is is an open access article distributed under the Creative Commons Attribution License whichpermits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

Distress types are significant for asphalt pavement maintenance decision and relationships among them can greatly influence thedecision outcomes In this study to analyze the relationships among different distress types from a statistical point of view 282asphalt pavements with semirigid base structures in 23 regions of China were surveyed to identify 12 distress types which weresubsequently investigated by association rule mining Results show that the distress types can be categorized into independentdistress types (IDDTs) dependent distress types (DDTs) and rutting secondary distress types (RSDTs) based on the relationships+e relationships among IDDTs are the strongest and those between IDDTs and DDTs between rutting and RSDTs and amongdepression pumping and raveling are also strong while the others are weak +e weak relationships should be ignored while thestrong ones should be considered to reduce cost and guarantee accuracy of maintenance decision +e IDDTs are the mostimportant distress types so preventing them from occurring or maintaining them immediately can greatly preserve goodperformance of asphalt pavements To help analyze the relationships a distress-type system was established and verified Inconclusion this work provides new insights to understand distress types

1 Introduction

Asphalt pavements play an important role in transportationand those with good conditions can provide drivers witha safe and comfortable ride However asphalt pavementscontinuously suffer from combined effects of heavy andrepeated traffic loadings and the natural environment duringservice lives [1]+is will bring about different distress typessuch as map cracking caused by repeated traffic loadingsmoisture damage transverse cracking at low temperaturesand rutting at high temperatures [2ndash5] +e distresses cansignificantly reduce performance and shorten service lives ofasphalt pavements [6] For example cracks not only con-tribute to a rough and noisy ride but also allow water topenetrate pavement structures which accelerates the de-terioration of asphalt pavements [1 7] Hence maintainingdeteriorated asphalt pavements is an urgent problem forroad agencies

To keep asphalt pavements at a high service level roadagencies should immediately make maintenance decision toselect some maintenance activities which consist of main-tenance needs and measures +e most common method formaking maintenance decisions is the pavement manage-ment system (PMS) a tool that comprises the network andproject level first introduced in 1960s [8 9] +e networklevel focuses on maintenance needs for road networks whilethe project level aims at maintenance measures for an in-dividual project [9] +e PMS determines maintenance ac-tivities by considering management factors (such asmaintenance budgets and costs) and engineering factors (suchas pavement performance detection and evaluation) [10 11]Specially the pavement performance is evaluated by someindices For instance the first pavement index calculated bysubtracting deduct values of distress types namely thepavement condition index (PCI) introduced by the USArmy Corps of Engineers [12]

HindawiAdvances in Civil EngineeringVolume 2019 Article ID 5369532 15 pageshttpsdoiorg10115520195369532

+e PMS determines maintenance activities in thenetwork- and project-level steps sequentially For the twosteps the work flows are similar but the latter is moredetailed and focused [8] In the network-level step roadagencies detect pavement condition and evaluate pavementperformance mainly using PCI and the riding quality index(RQI) According to the evaluation results they providesome alternative maintenance strategies whose pavementperformance and cost are subsequently predicted After-ward by considering the maintenance budgets environ-mental factors and short- and long-term needs they allocatelimited maintenance resources to road network reasonablyand identify the projects need to be funded first [8 13] Afternetwork-level models have determined pavement sections tobe financed road agencies detect these sections in the projectlevel more accurate and detailed than the network level [14]Subsequently they evaluate pavement performance com-prehensively by many indices such as the pavementstructure strength index (PSSI) in addition to PCI and RQIBased on the evaluation data some alternative maintenancemeasures are suggested then the pavement performance foreach alternative is predicted and the corresponding cost isestimated +rough comparing the cost and effect of everyalternative the most cost-effective maintenance measure isdetermined for the pavement sections +rough the aboveprocess the PMS performs well at the network level forallocating maintenance funds but inadequately for particularpavement sections at the project level [15 16] It is becauseasphalt pavements with different distress types maybepossess the same value of the pavement performance indicesindicating the same maintenance measure may be selectedfor the asphalt pavements However the same one measuregenerally cannot produce good effects for different distresstypes For example two asphalt pavements namely onewith transverse cracking and the other with pothole canown the same value of PSSI As a result slurry seal may beselected for the two asphalt pavements but it is not specificenough for the above two distress types

To maintain asphalt pavements accurately and effec-tively road agencies should consider not only pavementperformance but also particular distress types Manyscholars have explored diverse distress types especiallydifferent cracking for pavement maintenance Due to thecomplexity of cracking Ouma et al [1] analyzed the for-mation and relationships of multiple cracks which werecharacterized by longitudinal transverse diagonal blockand alligator through the wavelet morphology It shows thatthe diverse cracks possess some connections and the re-lationships are important for pavement maintenance butthe detailed relationships are not discussed +e Indianadepartment of transportation (DOT) found numerous statesin the USA cannot manage reflective cracking effectively dueto the lack of systematic procedures for selecting appropriatemaintenance measures Hence they studied the current stateof reflective cracking by a survey of all states DOTs anddeveloped a decision-making process to enhance mainte-nance measures selection [7] In addition to investigatethermal cracking of asphalt pavements based on the datafrom 46 long-term pavement performance (LTPP) sections

Dong et al [17] selected six important influence factors+en the relationships among the six factors and betweenthe six factors and the thermal cracking development ratewere analyzed through association rule mining and a dataminer whose supports and confidences can reflect thestrength of relationships Although this research only fo-cuses on thermal cracking it shows that the association rulesare very suitable for investigating probabilities of mutualinduction between different items namely relationshipsbetween different items Apart from cracking potholes andthe corresponding maintenance methods were exploredthrough a questionnaire survey of six provincial trans-portation agencies in Canada by Biswas et al [18] Mean-while this study also briefly discussed the relationshipsamong potholes map cracking longitudinal crackingbleeding and raveling Furthermore Sollazzo et al [19]probed the relationships between roughness and the pave-ment structural condition with an artificial neural networkmodel which is built and trained based on a huge number ofdata from the LTPP program +is model can give manyclues for road engineers to determine maintenance measuresat the project level in which the pavement structure strengthis very important for selecting maintenance measures

+e above studies about diverse distress types can assistroad engineers to further understand asphalt pavementconditions and select appropriate maintenance measuresHowever the relationships among different distress typeswere not thoroughly analyzed or deeply discussed Differentdistress types possess many relationships with each othersuggesting the existence of one distress type may induceother distress types to occur because of distress causes It canbe further illustrated by an example of three distress typesand two distress causes as presented in Figure 1 For thisexample distress type α occurs on an asphalt pavement dueto distress cause A meanwhile distress cause A can bringabout distress type β as well as distress cause B furthermoredistress cause B can lead to distress type c In summarydistress types α β and c occur on the asphalt pavementtogether Traditionally distress types that exist on pave-ments are considered for pavement maintenance whileother possible distress types are not For this reason aftersome maintenance measures being conducted for the ex-istent failures other distress types may occur at once Asa result the serviceability of asphalt pavements is still poor

+e above analysis shows that the relationships amongdifferent distress types are significant for pavement main-tenance In addition the association rule mining a dataminer that was first introduced by Agrawal et al [20] isa promising tool to analyze relationships between differentitems [17 20 21] Hence this study mainly aims to in-vestigate the relationships among different distress typesusing association rule mining to make pavement mainte-nance accurate and effective

2 Objective and Scope

+emain purpose of this study is to analyze the relationshipsamong different distress types Considering the complexityof distress causes and the limited space this study only

2 Advances in Civil Engineering

focuses on the relationships from a statistical point of viewDetailed objectives of this study are as follows

(1) A total of 282 asphalt pavements with semirigid basestructures in 23 regions of China are surveyed and12 major distress types are identified

(2) Association rules among the 12 distress types aremined to analyze the relationships

(3) To help analyze the relationships a distress-type systemof asphalt pavements is established and then verified bythe occurrence probabilities of the distress types

3 Distress Types Survey

To investigate distress types a total of 282 asphalt pavementsin China that were not maintained were inspected in thisstudy All the asphalt pavements were freeways withsemirigid base structures the major pavement structures inChina [4] In addition the surveyed asphalt pavements arelocated in the 23 regions of China as shown in Figure 2+isfigure displays that the inspected asphalt pavements covermost of China so they are representative

Meanwhile the surveyed 23 regions belong to the sixclimate zones in China that possess nine climate zones intotal +e surveyed climate zones are 1-2 1-3 1-4 2-2 2-3and 2-4 while those not inspected are 1-1 2-1 and 3-2 [22]For the climate zones the first number is the high-tem-perature index and the second is the low-temperature indexMoreover the smaller the first number the higher the av-erage maximum temperature of the hottest month and thesmaller the second number the lower the extreme minimumtemperature For the inspected six climate zones 1-2 in-cludes Beijing City Hebei Province Shanxi Province andTianjin City 1-3 contains Anhui Province Chongqing CitySichuan Province Henan Province Hubei ProvinceShandong Province Hunan Province Shanghai City andJiangsu Province 1-4 involves Jiangxi Province FujianProvince Guangxi Province Zhejiang Province andGuangdong Province 2-2 covers Gansu Province andLiaoning Province the 2-3 covers Shaanxi Province and the2-4 includes Guizhou Province and Yunnan Province +eother three climate zones not surveyed 1-1 2-1 and 3-2are situated in Tibet and Sinkiang Autonomous Regions

which possess few freeways Given this the surveyedasphalt pavements are also representative from a climateperspective

In addition the details of the surveyed asphalt pave-ments including climate zones regions and numbers aresummarized in Table 1 for easy access

Furthermore the service lives of all surveyed asphaltpavements ranged from 6 to 10 years because asphaltpavements are easily damaged at this stage +is means thedifference in the service lives is small Hence the influence ofthe service lives on different distress types can be ignoredwhen analyzing the relationships among them Furthermorethe design service life of freeways in China is 15 yearsfollowing the Chinese standard JTG D50-2017 [23] in-dicating all investigated distresses occurred within the de-sign service life

To collect pavement distress data two main ap-proaches are usually used by road agencies manual andautomated distress data collections [24] +e automatedapproach is costly because not only advanced equipment isrequired but also subsequent photographs need to beprocessed Meanwhile the equipment is so complicatedthat road engineers should be trained in advance [25] In

Distress cause A

Distress cause B

Asphalt pavement

Distresstypeγ

Distress typeα

Distress typeβ

Figure 1 Relationships among different distress types

Regions surveyed

Regions not surveyed

Figure 2 Locations of surveyed 23 regions in China

Advances in Civil Engineering 3

addition some certain early age distresses could not bedetected by the automated approach [14] Hence thismethod cannot be fulfilled by many road agencies Incontrast the manual approach is cost-effective because itjust needs road engineers to inspect failure areas andcollect distress data [13] +is task is very familiar for theroad engineers in China because they have been workingon it for many years according to the Chinese standardJTG H20-2007 [26] +is standard provides detailed de-scriptions and clear regulations for each distress types sothat the road engineers can evaluate distress conditionaccurately Given the above analysis the manual in-spection method was applied to collect distress data andidentify distress types in this study

+is study aims at the relationships among differentdistress types so the kinds of distresses are the most sig-nificant while the severity levels of distresses are relativelyless important As a result the severity levels were dis-regarded here For example rutting includes both seriousrutting and slight rutting In terms of this principlethrough the manual inspection this study identified 12major distress types of asphalt pavements +ese distresstypes are transverse cracking rutting map cracking pot-hole longitudinal cracking raveling depression pumpingbleeding bump poor skid resistance and roughness +eacronyms of these distress types are listed in Table 2 forconvenience

4 Preliminary Analysis of Relationships

41 Preliminary Classification of Distress Types To facilitatethe analysis of relationships among different distress typesthe distress types should be classified at first Before

classification the proportions of asphalt pavements withdifferent distress-type numbers were investigated and theresults are presented in Figure 3 +e method for calculatingthe proportions is as follows

(1) Inspect distress types that existed on each asphaltpavement

(2) According to the inspection results count the dis-tress-type number i on each asphalt pavement where1le ile 12 since distress types that exist on a pave-ment are at least one and up to 12

(3) Count the quantity of pavements with the samedistress-type number+e quantity of pavements withdistress-type number i is qi For instance q2 denotesthe quantity of pavements with two distress types

(4) Add the above quantities qi and obtain the quantityqall 1113936

12i1qi of all asphalt pavements which is 282 in

this study(5) +e proportion of asphalt pavements with distress-

type number i in all pavements is given asproportioni qiqall qi282

From Figure 3 7021 of asphalt pavements possess twoand more distress types indicating combinations of multipledistress types deserve much attention from road agencies Inaddition although 7021 of asphalt pavements ownmultipledistress types 2979 of asphalt pavements possess only onedistress type For these pavements only four distress typesnamely rutting transverse cracking map cracking andpothole can occur on them independently as displayed inFigure 3 Hence these four distress types are called the in-dependent distress types ie IDDTs Conversely the othereight distress types cannot occur independently and theymust occur with other distress types so they are called thedependent distress types ie DDTs +e DDTs include LCRA DE PU BL BU PSR and RO

42 Preliminary Distress-Type System of Asphalt PavementsAs analyzed previously a DDT has to occur with otherdistress types but it is not clear whether a DDTmust occur

Table 1 Details of surveyed asphalt pavements

Climate zone Region Number

1-2

Beijing City 10Hebei Province 6Shanxi Province 8Tianjin City 6

1-3

Anhui Province 6Chongqing City 6Henan Province 12Hubei Province 12Hunan Province 14Jiangsu Province 20

Shandong Province 12Shanghai City 16

Sichuan Province 6

1-4

Fujian Province 10Guangdong Province 34Guangxi Province 20Jiangxi Province 8Zhejiang Province 30

2-2 Gansu Province 8Liaoning Province 20

2-3 Shaanxi Province 8

2-4 Guizhou Province 6Yunnan Province 5

Table 2 Acronyms of distress types

Distress type AcronymTransverse cracking TCRutting RUMap cracking MCPothole POLongitudinal cracking LCRaveling RADepression DEPumping PUBleeding BLBump BUPoor skid resistance PSRRoughness ROIndependent distress type IDDTDependent distress type DDTCombinations among DDTs C-DDTsRutting secondary distress type RSDT


with other DDTs orand some IDDTs +is will be discussedin this section

From Figure 3 approximately 70 of asphalt pavementspossess combinations of multiple distress types so theIDDTs and the DDTs were combined here To simplifyanalysis the combinations among the DDTs are treated asa whole and defined as the C-DDTs +e C-DDTs includesome DDTs maybe only one DDT or maybe all DDTs Interms of the above definition the proportions of asphaltpavements with different combinations of the IDDTs and theC-DDTs were calculated and the results are presented inFigure 4 In addition the calculation method is as follows

(1) Inspect distress types emerging on each asphaltpavement

(2) Based on the inspection results record the distresstypes occurring on each asphalt pavement one ormore DDTs are marked as the C-DDTs For exampleregarding two asphalt pavements one with ruttingand bleeding and the other with rutting bleedingand roughness both of the pavements are marked asRU+C-DDTs

(3) Count the quantity Q of asphalt pavements with thesame combinations of distress types For instancethe quantity of pavements with RU+C-DDTs isQRU+CminusDDTs

(4) Add the above quantitiesQ and obtain the quantityQallof all asphalt pavements which is 282 in this study

(5) +e proportion of asphalt pavements with differentdistress types is given as proportion QQall Q282eg proportionRU+CminusDDTs QRU+CminusDDTs282

From Figure 4 the asphalt pavements with C-DDTs arezero indicating the combinations among the DDTs cannot

occur independently so a DDT must emerge with someIDDTs In other words if one DDT exists on an asphaltpavement then some IDDTs must exist on the pavementand otherwise they will occur When some IDDTs haveoccurred on asphalt pavements the C-DDTs own highpercentage As plotted in Figure 4 the asphalt pavementswith the C-DDTs and all IDDTs (or except map cracking) arebigger than 10 and the asphalt pavements with the C-DDTs and transverse crackingruttingtransverse crackingand rutting are just below 10

As analyzed previously the DDTs have to emerge withsome IDDTs so the DDTs are wholly attached to theIDDTs For this reason a preliminary distress-type systemof asphalt pavements was established as shown in Figure 5In Figure 5 a DDT is attached to the entirety of the IDDTsand it has to occur with some IDDTs For instance lon-gitudinal cracking can occur when transverse crackingexists on an asphalt pavement If longitudinal crackingexists on an asphalt pavement that possesses no IDDTsthen some IDDTs (maybe transverse cracking) will occurFurthermore the relationships which exist between theIDDTs and the DDTs individuals of the IDDTs and in-dividuals of the DDTs are not clear +e details on theserelationships will be discussed by association rule miningin the later sections

5 Mining Association Rules

To further investigate relationships among all the distresstypes the above statistics were analyzed using the associa-tion rule mining because it can identify relevance and de-pendency of different items [21]

51 Fundamental Concepts of Association Rule MiningIn association rule mining three fundamental conceptsare important including an item an item set and a trans-action [27]

(1) An item is the minimum unit in a database In thiswork an item is a distress type such as transversecracking rutting and pothole Since a total of 12major distress types were identified 12 items wererecognized

(2) An item set is a collection of different itemsIn this study it is a combination of different dis-tress types such as transverse Cracking rutting1113864 1113865+e 12 distress types comprise the total item setTIS TC RU MC PO LC RA DE PU BL BUPSR RO

(3) +e essence of a transaction is an item set In thiswork a transaction is the distress types occurring onan asphalt pavement meaning that one pavementcorresponds to one transaction For instance iftransverse cracking and rutting occur on a pavementtogether then transverse cracking rutting1113864 1113865 is thecorresponding transaction for this pavement Since282 asphalt pavements were surveyed 282 trans-actions were identified making up the transactiondatabase of distress types

Asphaltpavements

with two and

more distresstypes7021

Asphaltpavements

with onedistress type2979

RU1489

TC 851

MC

355

PO 284

Figure 3 Proportions of asphalt pavements with different distress-type numbers


By an association rule it means an implication of theform X⟶ Y where X and Y are item sets the propersubsets of the TIS and XcapY empty

52 Evaluating Association Rules To evaluate associationrules two indicators are necessary the support and theconfidence of association rules

521 Support of Association Rules +e support of an as-sociation rule is a significant measure that indicates thefrequency of occurring patterns and reflects the importanceof rules +e support of X⟶ Y is calculated by the fol-lowing equation

support(X⟶ Y) N(XcupY)

N(ALL) P(XcupY) (1)

where N(XcupY) is the number of transactions includingboth X and Y and N(ALL) is the number of all transactionsin the distress-type database and it is 282 here Furthermorethe supports of X⟶ Y and Y⟶ X are equal and theyare the probability of X and Y occurring together whichmeans the greater the supports the stronger the relation-ships among different items

522 Confidence of Association Rules +e confidence ofX⟶ Y indicates the reliability of the association rule andit is expressed by the following equation

confidence(X⟶ Y) N(XcupY)

N(X) P(Y | X) (2)

where N(X) is the number of transactions containing X+e confidence represents the probability of Y occurringwhen X exists +e confidence also affects relationshipsamong different items but it is based on the support

523 Example for Support and Confidence To further il-lustrate the support and the confidence of association

All IDDTs + C-DDTs

TC + RU + PO + C-DDTs

TC + MC + PO + C-DDTs

RU + MC + PO + C-DDTs

TC + RU + MC + C-DDTs

MC + PO + C-DDTs

RU + PO + C-DDTs

RU + MC + C-DDTs

TC + PO + C-DDTs

TC + MC + C-DDTs

TC + RU + C-DDTs

PO + C-DDTs

MC + C-DDTs

RU + C-DDTs

TC + C-DDTs

C-DDTs

0 5 10Probability ()

Asp

halt

pave

men

t with

dist

ress

type

s

0

Figure 4 Proportions of asphalt pavements with different combinations of IDDTs and DDTs

Independent distress types (IDDTs)transverse cracking ruttingmap cracking and pothole

Dependent distress types (DDTs)longitudinal cracking ravelingdepression pumping bleeding

bump roughnessand poor skid resistance

Figure 5 Preliminary distress-type system of asphalt pavements


rules a transaction database is provided in Table 3 Aspresented in Table 3 three transactions are No 1 No 2and No 3 Meanwhile one transaction corresponds toone asphalt pavement with different distress types+e total item set of the transaction database isTIS rutting transverse cracking pumping1113864 1113865 Many as-sociation rules can be generated from this database andtwo typical ones of them are selected to illustrate thesupport and confidence of association rules

Regarding the association rule transverse cracking⟶rutting the transaction including both transverse crackingand rutting is only one ie No 3 Owing to three trans-actions altogether in the database the support is calculatedby 13 3333 the transactions containing transversecracking are two namely No 1 and No 3 so the confidenceis given by 12 50 +e confidence means rutting ownsa 50 occurrence probability when transverse crackingexists on asphalt pavements and the probability of thissituation is 3333 according to the support Meanwhile thesupport also indicates the probability of transverse crackingand rutting occurring together is 3333

Concerning the association rule pumping⟶ transversecracking the transaction incorporating pumping andtransverse cracking together is only one ie No 3 Becauseof this database possesses a total of three transactions thesupport is 13 3333 as the transaction includingpumping is also only one still No 3 the confidence iscalculated by 11 100 +e confidence suggests whenpumping emerge on asphalt pavements transverse crackingmust exist on the pavements if transverse cracking does notexist it must occur in the future Meanwhile the probabilityof this situation is 3333 in terms of the support In ad-dition the support also means transverse cracking andpumping possess a 3333 probability of emerging together

524 4resholds of Support and Confidence +e thresholdsof support and confidence namely the minimum valuesshould be determined before mining association rules +eminimum support (support threshold) based on projectrequirements must be determined to ensure the importanceof association rules Generally an association rule is im-portant when the support reaches 5 [17 20] As mentionedin Section 521 the greater the supports the stronger therelationships Meanwhile the minimum confidence (con-fidence threshold) is needed to assure the reliability of as-sociation rules since the confidence also has an impact onrelationships Association rule mining is generating strongrules meeting the support and confidence thresholdsconcurrently

It is noteworthy that this study aims at investigatingrelationships among different distress types meaning allassociation rules among the 12 distress types are requiredFor this reason both the support and confidence thresholdswere taken as zero to mine strong association rules

53 Miners of Association Rules In many association rulesminers the most widely used is the a priori algorithm firstintroduced by Agrawal et al [28] +e a priori algorithm

performs in two steps First it scans the database and dis-covers all item sets with an occurrence not less than thesupport threshold Second according to the discovered itemsets a set of strong association rules are mined +e detailson the a priori algorithm can be found in Reference [27 28]Additionally the a priori algorithm was executed by Wai-kato Environment for Knowledge Analysis (WEKA) anopen-source tool universally applied in business educationand research [29]

6 Analysis of Relationships Based onAssociation Rules

61 Relationships among IDDTs Regarding the four IDDTsC14 middot C1

3 12 association rules exist among the IDDTs +esupports are presented in Figure 6(a) where TC harr RUrepresents the association rules TC⟶ RU and RU⟶ TCso do others Meanwhile the confidences are depicted inFigure 6(b)

From Figure 6(a) all supports are bigger than 20 andthe average value is 2506 meaning the average probabilityof two IDDTs occurring together is around 25 indicatingpowerful relationships among the IDDTs Because of thepowerful relationships among the IDDTs the IDDTs couldnot only occur independently but also induce each other toemerge which can be demonstrated by the confidences inFigure 6(b) +is figure shows that the confidences of allassociation rules change little with an average of 5247meaning the average probability is over 50 that one IDDTinduces another to occur

Of the 12 association rules TCharr RU possess the largestsupport over 30 suggesting more than 30 of asphaltpavements own both transverse cracking and ruttingTransverse cracking is widespread because most freeways inChina are semirigid base structures which can easily bringabout drying shrinkage and thermal shrinkage and result intransverse cracking [30] Meanwhile rutting is prevalent asa consequence of the increase in traffic volume tire pressureand axial load [4]

Actually transverse cracking is not a serious damage forasphalt pavements since it just decreases the pavementsmoothness However it usually leads to other distress typesthat are serious for pavements such as map cracking andpotholes Typically transverse cracking breaks the integrityof the surface course so that moisture can penetrate into thesurface and accumulate between the lower surface courseand the semirigid base course whose voids are almost zeroWhen traffic loadings pass through the transverse crackingthe moisture can generate hydrodynamic pressure +ehydrodynamic pressure can also decrease the bonds betweenthe base course and the lower surface course so that thelower surface course has to bear large tensile stress When

Table 3 Transaction database of distress types an example

Transaction number TransactionNo 1 Transverse crackingNo 2 RuttingNo 3 Transverse cracking rutting pumping


the tensile stress is bigger than the tensile strength fatiguedamage will occur in the lower surface course and spreadinto the map cracking in the upper surface course with therepeated action of numerous and great traffic loadingsGiven this the support of TC harr MC is high reachingapproximately 20 Meanwhile the hydrodynamic pressurecan diminish the adhesion bonds at the aggregate-masticinterface and peel asphalt films from aggregates [31] con-tributing to the occurrence of potholes at the transversecracking As a result the support of TCharr PO is very largegreater than 30 In the same way the moisture existing inmap cracking will also produce the hydrodynamic pressurewith the effect of traffic loadings and give rise to the potholeHence the support of MC harr PO is also big around 20

In addition moisture can usually gather in rutting areassince rutting depth is usually greater than 10mm [4 26] Itcan bring about hydroplaning when plentiful vehiclespassing through the accumulated moisture which is not safefor road users [4] Meanwhile with the repeated effect of thevehicle compressive stress and the vacuum suction gener-ated by high-speeding driving asphalt films are stripped offaggregates After that the aggregates are pulled out by ve-hicle wheels so that pothole occurs on asphalt pavementsHence the support of RUharr PO is great approximately 25Furthermore because the strong relationships betweenpothole andmap cracking as analyzed earlier the support ofRUharr MC is also large whose value is approximately 20

According to the above analysis the powerful relationshipsamong the IDDTs should be considered to make asphaltpavement maintenance decision accurately Specifically for anasphalt pavement with only one IDDT this existent IDDT aswell as other IDDTs that maybe occur should be considered Itcan be further illustrated by the example of an asphalt pave-ment with only transverse cracking In this example transversecracking and pothole should be considered for pavementmaintenance because pothole owns an approximately 50occurrence probability when transverse cracking existsaccording to the confidence of TC⟶ PO in Figure 6(b)

Meanwhile the probability of the above situation is around30 in terms of the support of TC harr PO in Figure 6(a)

62 Relationships between IDDTs and DDTs Because of thefour IDDTs and the eight DDTs there are 2 times C1

4 middot C18 64

association rules between the IDDTs and the DDTs Allassociation rules are provided in Figure 7 where TC⟶DDTs and DDTs⟶ TC represent the association rulesbetween transverse cracking and the eight DDTs so doothers

According to Figure 7(d) it is noteworthy that fourconfidences are 100 whose association rules are BLBUPSRRO⟶ RU+ese association rules possess supports ofapproximately 5 so they cannot be disregarded Hence therelationships between the four distress types and rutting arestrong +e 100 confidences mean when BLBUPSRROexists on an asphalt pavement rutting must exist if ruttingdoes not emerge then it will occur on the pavement in thefuture However the reverse is not true because the confi-dences of association rules RU ⟶ BLBUPSRRO aresmall than 100 as shown in Figure 7(c) For example theconfidence of association rule RU⟶ BL is about 10 sobleeding just owns an about 10 occurrence probabilitywhen rutting exists

+e strong relationships between the four distress typesand rutting may be because some distress causes can inducethem at the same time Traditionally the poor constructionof asphalt pavements can lead to aggregate gradation seg-regation namely some areas with concentrated coarse ag-gregates and some areas with concentrated fine aggregates[32] For the areas with concentrated fine aggregates the fineaggregates are so many that they can push the coarse ag-gregate skeleton of asphalt mixtures open +is will declinethe shear strength of asphalt mixtures and bring aboutrutting and bump [33] Meanwhile the overmuch fine ag-gregates make air voids so small that they may be full ofasphalt binder+e asphalt binder will move to the pavement

TC harr RU

TC harr MC

TC harr PO

RU harr MC

RU harr PO

MC harr PO

0 10 20 30 40Support ()

Average is 2506

Ass

ocia

tion

rule

(a)

TC rarr RUTC rarr MCTC rarr PORU rarr TCRU rarr MCRU rarr POMC rarr TCMC rarr RUMC rarr POPO rarr TCPO rarr RUPO rarr MC

0 20 40 60 80 100Confidence ()

Ass

ocia

tion

rule

Average is 5247

(b)

Figure 6 Association rules among IDDTs (a) Supports in IDDTs (b) Confidences in IDDTs


20 100

80

60

40

20

0

15

10

5

0BL BU PSR RO LC

Confidence

Supp

ort (

)

Con

fiden

ce (

)

Support

RA DE PUDDT

(a)

20 100Average confidence is 6282

80

60

40

20

0

15

10

5

0

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

BL BU PSR RO LC RA DE PUDDT

(b)20

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

BL BU PSR RODDT

LC RA DE PU

100

80

60

40

20

0

15

10

5

0

(c)

Average confidenceis 5842

20100

80

60

40

20

0

15

10

5

0

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


(d)

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

DDTBL BU PSR RO LC RA DE PU

20 100

80

60

40

20

0

15

10

5

0

(e)


Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


20 100

80

60

40

20

0

15

10

5

0

(f )

Figure 7 Continued


upper surface with the effect of traffic loadings in hotweather contributing to the occurrence of bleeding [34]Concurrently the uneven distribution of aggregates leads tothe nonuniform distribution of compaction degree whichbrings about variable compacted depths on asphalt pave-ments with compressive stress of various vehicles Conse-quently the pavements will possess roughness In additionrutting can emerge on asphalt pavements when the steadystate of aggregate skeleton is not good [35] which may becaused by the inadequate strength of aggregates Further-more the insufficient aggregate strength can also result inpoor skid resistance of pavement upper surface course

According to the above analysis these four distress typesmust adhere to rutting when occurring so they are called therutting secondary distress types ie the RSDTs Hence theDDTs are redefined as LC RA DE and PU So far the 12distress types of asphalt pavements are classified into threecategories the IDDTs the DDTs and the RSDTs

From Figures 7(a) 7(c) 7(e) and 7(g) in general thesupports between the IDDTs and the DDTs which areslightly over 10 are much bigger than those between theIDDTs (except rutting) and the RSDTs which are around25 It is because the RSDTs must adhere to rutting tooccur while the DDTs can emerge with one or more IDDTsIn terms of the analysis the relationships between the IDDTsand the DDTs are stronger than those between the IDDTs(except rutting) and the RSDTs+e strong relationships canbe demonstrated by the confidences in Figures 7(b) 7(d)7(f) and 7(h) +ese figures suggest that the average con-fidences of DDTs⟶ IDDTs are around 60 indicating anapproximately 60 probability of the IDDTs occurring whenthe DDTs exist on asphalt pavements

In brief the relationships between the IDDTs and theDDTsRSDTs those between the IDDTs and the DDTs andthose between rutting and the RSDTs should be consideredto take maintenance decision accurately Conversely theweak relationships namely between the IDDTs (exceptrutting) and the RSDTs should be neglected to reducemaintenance decision cost

63 Relationships among DDTs and RSDTs So far the re-lationships among the IDDTs and those between the IDDTsand the DDTsRSDTs have been analyzed +e remainingrelationships including those among the DDTs among theRSDTs and between the DDTs and the RSDTs are discussedin this section For these relationships a total of C1

4 middot C13 +

C14 middot C1

3 + 2 times C14 middot C1

4 56 association rules exist among theDDTs and the RSDTs and the association rules are listed inTable 4 In Table 4 DDTs⟶ DDTs represents the asso-ciation rules among the DDTs and RSDTs⟶ RSDTs ownsthe same implication Meanwhile DDTs⟶ RSDTs andRSDTs ⟶ DDTs are the association rules between theDDTs and the RSDTs and LC⟶ DDTs is the associationrules between longitudinal cracking and the other DDTs sodo others Moreover in Table 4 S and C denote the supportand the confidence of association rules respectively

From the bottom-right of Table 4 all supports of RSDTs⟶ RSDTs are far less than 5 revealing the probabilitythat two RSDTs occur together is very low not to mentionthree or four RSDTs In other words the RSDTs can hardlyinduce each other to occur Hence the relationships amongthe RSDTs are extremely weak Similarly from the top-rightand bottom-left of Table 4 all supports of DDTs⟶ RSDTsand RSDTs ⟶ DDTs are below 5 indicating the re-lationships between the DDTs and the RSDTs are very weakas well Concerning an asphalt pavement with only bleedingin terms of the support of BL⟶ PSR PSR should beneglected because the probability of PSR and bleeding oc-curring together is only 071 as shown in the down-right ofTable 4

Compared with the above weak relationships the re-lationships between depression raveling and pumping arestronger since the supports of DE⟶ RA (RA⟶ DE) DE⟶ PU (PU⟶ DE) and PU⟶ RA (RA⟶ PU) rangefrom 5 to 10 as shown in the top-left of Table 4 Forconvenience the relationships among the three distresstypes are plotted in Figure 8 From Figure 8 the averagesupport and confidence among these distress types are 682and 3815 respectively It means the average probability of

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

DDT

BL BU PSR RO LC RA DE PU

20 100

80

60

40

20

0

15

10

5

0

(g)


20 100

80

60

40

20

0

15

10

5

0

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

DDT


(h)

Figure 7 Association rules between IDDTs and DDTs (a) TC⟶ DDTs (b) DDTs⟶ TC (c) RU⟶ DDTs (d) DDTs⟶ RU(e) MC⟶ DDTs (f ) DDTs⟶ MC (g) PO⟶ DDTs (h) DDTs⟶ PO


Tabl

e4

Associatio

nrulesam

ongDDTsa

mon

gRS

DTsa

ndbetweenDDTs

andRS

DTs

SC

SC

SC

SC

SC

SC

SC

SC

DDTs

DDTs⟶

DDTs

RSDTs⟶

DDTs

LC⟶

DDTs

RA⟶

DDTs

DE⟶

DDTs

PU⟶

DDTs

BL⟶

DDTs

BU⟶

DDTs

PSR⟶

DDTs

RO⟶

DDTs

LC284

1667

426

2500

426

2069

142

2857

213

3333

142

2222

142

6667

RA284

1538

780

4583

567

2759

284

5714

355

5556

071

1111

071

3333

DE

426

2308

780

4583

709

3448

284

5714

355

5556

000

000

071

3333

PU426

2308

567

3333

709

4167

284

5714

213

3333

142

2222

071

3333

RSDTs

DDTs⟶

RSDTs

RSDTs⟶

RSDTs

LC⟶

RSDTs

RA⟶

RSDTs

DE⟶

RSDTs

PU⟶

RSDTs

BL⟶

RSDTs

BU⟶

RSDTs

PSR⟶

RSDTs

RO⟶

RSDTs

BL142

769

284

1667

284

1667

284

1379

284

4444

071

1111

071

3333

BU213

1154

355

2083

355

2083

213

1034

284

5714

071

1111

071

3333

PSR

142

769

071

417

000

000

142

690

071

1429

071

1111

142

6667

RO142

769

071

417

071

417

071

345

071

1429

071

1111

142

2222

SandCdeno

tesupp

orta

ndconfi

denceof

associationrulesrespectiv

ely


two of them occurring together is 682 and the averageprobability of one of them inducing another to occur is3812 For example when depression exists pumping andraveling have probabilities of 4167 and 4583 occurringrespectively

+e relationships between the three distress types arestrong maybe because they possess the same distress causesAs analyzed in section 61 when transverse or map crackingexists on asphalt pavements the incomplete pavementstructure will be invaded by moisture +e moisture willerode the aggregate-asphalt interface and peel off asphaltbinder with the repetitive action of the vacuum suctiongenerated by high-speeding driving After asphalt binderbeing stripped off aggregate surface the asphalt mixtures inthe upper surface course will emerge raveling Meanwhilethe aggregates in the lower surface course will leave thepavement structure with pumping resulting in large holes inthe lower surface course and depression in the pavementstructure Given the above analysis the strong relationshipsshould also be considered for maintenance

In summary for the relationships existing among the DDTsand RSDTs only those between depression pumping andraveling should be considered for asphalt pavement accuratemaintenance decision On the contrary other relationshipsnamely between the DDTs and the RSDTs and among theRSDTs should be ignored to reduce maintenance decision cost

7 Distress-Type System of Asphalt Pavements

71 Establishment In terms of the preliminary distress-typesystem in Figure 5 (section 42) and the relationships be-tween rutting and the RSDTs in Figure 7(d) (section 62) thedistress-type system of asphalt pavements was established asdepicted in Figure 9 +e system can be used to help analyzerelationships among different distress types

From Figure 9 the IDDTs not only occur independentlybut also induce each other to emerge Meanwhile Figure 9also shows that every DDT is attached to the entirety of theIDDTs while each RSDT is attached to rutting indicatinga DDTmust occur with some IDDTs while an RSDT has toemerge with rutting In particular if a DDT occurs on anasphalt pavement with no IDDT then some IDDTs willemerge on this pavement Similarly if an RSDT appears onan asphalt pavement without rutting then rutting will occurIn summary Figure 9 displays that the interrelationshipsamong the IDDTs between the IDDTs and the DDTs andbetween rutting and the RSDTs are strong but those betweenthe DDTs and the RSDTs are weak Since the IDDTs occurindependently and mutually and the other distress typeshave to emerge with them they are the most importantdistress types for asphalt pavements

Furthermore Figure 9 implies that the four IDDTscorrespond to four important properties of asphalt mixturesSpecifically transverse cracking vs low-temperature prop-erty rutting vs high-temperature stability map cracking vsfatigue property and pothole vs moisture susceptibility+is further confirms the significance of the IDDTs Giventhe above analysis preventing the IDDTs from occurring or

maintaining them once they emerge can substantially keepasphalt pavements in good conditions

72Verification From Figure 9 the grades of the 12 distresstypes are so different that it can be speculated the occurrenceprobabilities of the distress types are also different and thechange may be very large Hence to verify the distress-typesystem in Figure 9 this study explored the occurrenceprobabilities of the 12 distress types +e occurrenceprobability of a distress types is defined as the occurrencenumber of this distress type divided by the total occurrencenumber of the 12 distress types Concerning the occurrencenumber one distress type can occur on many asphaltpavements so the sum of these pavements is the occurrencenumber of this distress type For instance transversecracking occurred on 166 asphalt pavements in the databaseof this study so it possesses the occurrence number of 166With this method the occurrence numbers of all distresstypes were counted and they were subsequently added up todetermine the total occurrence number whose value is 814+e processes for determining the occurrence probabilitiesare as follows

(1) Inspect distress types occurred on each asphaltpavement

(2) According to the inspection results count the oc-currence number nj (ge1) of every distress typewhere j is the distress type and j 1 2 12

(3) Add up the above occurrence number nj of the 12distress types to obtain the total occurrence numberntotal 1113936

12j1nj and the value is 814

(4) +e occurrence probability of distress type j is cal-culated by Probabilitj njntotal

It is noteworthy that the total occurrence number ntotal ofall distress types is different from the number of all distress

Depression

RavelingPumping

709

4167

3448

780

4583

4583

Average support is 685Average confidence is 3812

2759 3333

567

Figure 8 Relationships between depression pumping andraveling


types Traditionally the former is greater than the latterbecause one distress type can occur many times in a surveysample In this study the total occurrence number ntotal of alldistress types is 814 while the number of all distress types is12 as identified in section 3

+e above method and definitions can be further il-lustrated by the example in Table 3 +is example showsthree asphalt pavements one with transverse cracking onewith rutting and the other with transverse cracking ruttingand pumping In this database the number of all distresstypes is three while the total occurrence number ntotal arefive Among the ntotal transverse cracking and ruttingpossess two occurrence numbers each and pumping ownsone Hence the occurrence probabilities of both transversecracking and rutting are 40 and that of pumping is 20as shown in Table 5 Table 5 presents the process forcalculating the occurrence probabilities of distress types inthe example

With the above method the occurrence probabilities ofall distress types were calculated and the results are shown inFigure 10 From Figure 10 overall the occurrence proba-bilities of the distress types change largely Specifically theoccurrence probabilities of the IDDTs are the largest muchbigger than those of the DDTs which are far greater thanthose of the RSDTs+e occurrence probabilities are steppeddistribution consistent with the distress type system of as-phalt pavements in Figure 9 Figure 9 displays that theRSDTs have to emerge with rutting an IDDT so theypossess the lowest occurrence probabilities Comparedwith the RSDTs the DDTs own greater occurrence prob-abilities because they can occur with one or more IDDTsMeanwhile the RSDTs and the DDTs have to occur with theIDDTs and the IDDTs can occur independently andmutually so the IDDTs possess the highest occurrenceprobabilities

Regarding the four IDDTs transverse cracking andrutting own the highest probabilities and the values are bothbigger than 20 far greater than other distress types Itmeans transverse cracking and rutting are the most commondistress types in China so they deserve more attention thanothers from road agencies +is is consistent with thesupport of association rules TC harr RU which is more than30 as shown in Figure 6(a) indicating over 30 of asphaltpavements possess both transverse cracking and rutting

8 Conclusions

To investigate the relationships among different distresstypes this study surveyed 282 asphalt pavements withsemirigid bases in 23 regions of China and identified 12major distress types +rough the above analysis the mainconclusions are as follows

(1) +e 12 distress types were categorized into theIDDTs DDTs and RSDTs based on statisticalanalysis and association rules Specifically the IDDTsnot only occur independently but also induce eachother to emerge every DDT must occur with someIDDTs each RSDT must emerge with rutting anIDDT

(2) +e relationships among the IDDTs are the mostpowerful followed by those between the IDDTs andthe DDTs between rutting and the RSDTs andamong depression pumping and raveling Howeverthe relationships between the IDDTs (except rut-ting)DDTs and the RSDTs and those among theRSDTs are very weak +e weak relationships shouldbe ignored to reduce maintenance decision costwhile the strong relationships should be consideredto make asphalt pavement maintenance decisionaccurately

(3) +e IDDTs are the most significant distress types soavoiding them occurring or maintaining them in-stantly when they emerge can largely keep asphaltpavements in good conditions

(4) A distress-type system of asphalt pavements whichcan reflect the relationships between the IDDTsDDTs and RSDTs was established to help analyzethe relationships among different distress typesFurthermore the system is verified by the occurrenceprobabilities of different distress types which changegreatly and are stepped distribution

(5) Transverse cracking and rutting deserve more con-cern than other distress types from road agenciessince they are the most widespread distress types inChina Precisely they both have over 20 occur-rence probabilities and more than 30 of asphaltpavements possess them together

Because the surveyed asphalt pavements were allsemirigid base structures the relationships may be dif-ferent from other pavement structures such as cementconcrete pavement structures and full-depth asphaltpavement structures +erefore this study has limitations

Table 5 Calculation process of occurrence probabilities for theexample in Table 3

Distress type Occurrencenumber nj

Total occurrencenumber ntotal

Occurrenceprobability

Transversecracking 2

525 40

Rutting 2 25 40Pumping 1 15 20

Independent distress types (IDDTs)

Transverse cracking

Dependent distress types(DDTs)

Rutting secondary distress types(RSDTs)

Map cracking Rutting Pothole

BumpBleedingRoughnessPoor skid resistance

(i)(ii)

(iii)(iv)

LongitudinalcrackingRavelingDepressionPumping

(i)(ii)

(iii)(iv)

Figure 9 Distress-type system of asphalt pavements


and cannot be expanded to other pavement structuresHowever this study aims to provide a new perspective toinvestigate relationships among different distress types sothe results and conclusions can be used as referencesMoreover this study only quantitatively analyzed therelationships among different distress types from a statis-tical point of view Although the reasons for these re-lationships are described and we will investigate distresscauses in future research to further illustrate the re-lationships among the distress types

Data Availability

+e data used to support the findings of this study are in-cluded within the article

Conflicts of Interest

No conflicts of interest exist in the submission of thismanuscript and the manuscript is approved by all authorsfor publication

Acknowledgments

+e authors would like to acknowledge all those who par-ticipated in the field surveys on the asphalt pavementsAdditionally all the authors of the following references aremuch appreciated +is study was supported by the Post-graduates Research and Innovation Program in JiangsuProvince (KYLX15_0146 and KYLX_0167) Moreover othersupport was provided by the following projects Develop-ment and Application of Multi-Functions Life Cycle Ana-lyzer (2013YQ160501) Research on Key Technology of HotIn-Place Recycling for Rubber Asphalt Pavements(BY2015070-04) and Research on Normalized ConstructionTechnology of Hot Mix Plant Recycling for Asphalt Pave-ments (Transportation Science and Technology Project inJilin Province) +e authors gratefully acknowledge theirfinancial support

References

[1] Y O Ouma and M Hahn ldquoWavelet-morphology baseddetection of incipient linear cracks in asphalt pavements fromRGB camera imagery and classification using circular Radontransformrdquo Advanced Engineering Informatics vol 30 no 3pp 481ndash499 2016

[2] G Liu Y Jia T Yang H Du J Zhang and Y Zhao ldquoFatigueperformance evaluation of asphalt mixtures based on energy-controlled loading moderdquo Construction and Building Mate-rials vol 157 pp 348ndash356 2017

[3] L Gao H Li J Xie Z Yu and S Charmot ldquoEvaluation ofpavement performance for reclaimed asphalt materials indifferent layersrdquo Construction and Building Materialsvol 159 pp 561ndash566 2018

[4] T Xu and X Huang ldquoInvestigation into causes of in-placerutting in asphalt pavementrdquo Construction and BuildingMaterials vol 28 no 1 pp 525ndash530 2012

[5] Y Gao Y Zhang F Gu T Xu and H Wang ldquoImpact ofminerals and water on bitumen-mineral adhesion anddebonding behaviours using molecular dynamics simula-tionsrdquo Construction and Building Materials vol 171pp 214ndash222 2018

[6] N-D Hoang and Q-L Nguyen ldquoAutomatic recognition ofasphalt pavement cracks based on image processing andmachine learning approaches a comparative study on clas-sifier performancerdquo Mathematical Problems in Engineeringvol 2018 pp 1ndash16 2018

[7] Y Yoon S Patel R Ji and M Hastak ldquoCurrent state ofreflective cracking in the United Statesrdquo Journal of Con-struction Engineering and Management vol 143 no 3 article04016099 2017

[8] W R Hudson R Haas and R D Pedigo ldquoPavementmanagement system developmentrdquo NCHRP Report 215NCHRP Washington DC USA 1979

[9] O Swei J Gregory and R Kirchain ldquoPavement ManagementSystems Opportunities to Improve the Current FrameworksrdquoTransportation Research Board 95th Annual Meeting vol 162015

[10] K Golab R B Kulkarni and G B Way ldquoA statewidepavement management systemrdquo Interfaces vol 12 no 6pp 5ndash21 1982

[11] X Chen Q Dong H Zhu and B Huang ldquoDevelopment ofdistress condition index of asphalt pavements using LTPPdata through structural equation modelingrdquo TransportationResearch Part C Emerging Technologies vol 68 pp 58ndash692016

[12] ASTM D 6433-99 Standard Practice for Roads and ParkingLots Pavement Condition Index Surveys ASTM InternationalWest Conshohocken PA USA 1999

[13] Y O Ouma J Opudo and S Nyambenya ldquoComparison offuzzy AHP and fuzzy TOPSIS for road pavement maintenanceprioritization methodological exposition and case studyrdquoAdvances in Civil Engineering vol 2015 Article ID 14018917 pages 2015

[14] G P Ong S Noureldin and K C Sinha ldquoAutomatedpavement condition data collection quality control qualityassurance and reliabilityrdquo Joint Transportation ResearchProgram Indiana Department of Transportation and PurdueUniversity West Lafayette IN USA 2010

[15] Z He X Qin H Wang and C Comes ldquoImplementingpractical pavement management systems for small commu-nities a south Dakota case studyrdquo Public Works Managementamp Policy vol 22 no 4 pp 378ndash391 2017

25

20

15

Occ

urre

nce p

roba

bilit

y (

)

10

5

0TC RU MC PO LC RA

DDTsIDDTs RSDTsDistress types

DE PU BL BU PSR RO

Figure 10 Distribution of occurrence probabilities of differentdistress types


[16] Y Zhang and J P Mohsen ldquoA project-based sustainabilityrating tool for pavement maintenancerdquo Engineering vol 4no 2 pp 200ndash208 2018

[17] S Dong J Zhong P Hao et al ldquoMining multiple associationrules in LTPP database an analysis of asphalt pavementthermal cracking distressrdquo Construction and Building Mate-rials vol 191 pp 837ndash852 2018

[18] S Biswas L Hashemian and A Bayat ldquoInvestigation ofpothole severity and maintenance methods in Canadathrough questionnaire surveyrdquo Journal of Cold Regions En-gineering vol 32 article 04018002-1 2018

[19] G Sollazzo T F Fwa and G Bosurgi ldquoAn ANN model tocorrelate roughness and structural performance in asphaltpavementsrdquo Construction and Building Materials vol 134pp 684ndash693 2017

[20] R Agrawal T Imielinski and A Swami ldquoMining associationrules between sets of items in large databasesrdquoACM SIGMODRecord vol 22 no 2 pp 207ndash216 1993

[21] A Telikani and A Shahbahrami ldquoData sanitization in as-sociation rule mining an analytical reviewrdquo Expert Systemswith Applications vol 96 pp 406ndash426 2018

[22] Ministry of Transport of the Peoplersquos Republic of ChinaTechnical Specifications for Construction of Highway AsphaltPavements JTG F40-2004 China Communication PressBeijing China 2004

[23] Ministry of Transport of the Peoplersquos Republic of ChinaSpecifications for Design of Highway Asphalt Pavement JTGD50-2017 China Communication Press Beijing China 2017

[24] N-D Hoang ldquoAn artificial intelligence method for asphaltpavement pothole detection using least squares support vectormachine and neural network with steerable filter-based fea-ture extractionrdquo Advances in Civil Engineering vol 2018pp 1ndash12 2018

[25] M H Yousaf K Azhar F Murtaza and F Hussain ldquoVisualanalysis of asphalt pavement for detection and localization ofpotholesrdquo Advanced Engineering Informatics vol 38pp 527ndash537 2018

[26] Ministry of Transport of the Peoplersquos Republic of ChinaHighway Performance Assessment Standards JTG H20-2007China Communication Press Beijing China 2007

[27] R Rekik I Kallel J Casillas and A M Alimi ldquoAssessing websites quality a systematic literature review by text and as-sociation rules miningrdquo International Journal of InformationManagement vol 38 no 1 pp 201ndash216 2018

[28] R Agrawal and R Srikant ldquoFast algorithms for mining as-sociation rulesrdquo in Proceedings of the 20th InternationalConference on Very Large Databases VLDB pp 487ndash499Santiago Chile September 1994

[29] M Hall E Frank G Holmes B Pfahringer P Reutemannand I H Witten ldquo+e WEKA data mining softwarerdquo ACMSIGKDD explorations newsletter vol 11 no 1 pp 10ndash182009

[30] G Zang L Sun Z Chen and L Li ldquoA nondestructiveevaluation method for semi-rigid base cracking condition ofasphalt pavementrdquo Construction and Building Materialsvol 162 pp 892ndash897 2018

[31] A Bozorgzad S-F Kazemi and F Moghadas NejadldquoEvaporation-induced moisture damage of asphalt mixturesmicroscale model and laboratory validationrdquo Constructionand Building Materials vol 171 pp 697ndash707 2018

[32] S Chun K Kim B Park and J Greene ldquoEvaluation of theeffect of segregation on coarse aggregate structure and ruttingpotential of asphalt mixtures using Dominant Aggregate Size

Range (DASR) approachrdquo KSCE Journal of Civil Engineeringvol 22 pp 125ndash134 2017

[33] Y Zhang T Ma X Ding T Chen X Huang and G XuldquoImpacts of air-void structures on the rutting tests of asphaltconcrete based on discretized emulationrdquo Construction andBuilding Materials vol 166 pp 334ndash344 2018

[34] H Khosravi S M Abtahi B Koosha and M Manian ldquoAnanalytical-empirical investigation of the bleeding mechanismof asphalt mixesrdquo Construction and Building Materialsvol 45 pp 138ndash144 2013

[35] T Ma D Zhang Y Zhang and J Hong ldquoMicromechanicalresponse of aggregate skeleton within asphalt mixture basedon virtual simulation of wheel tracking testrdquo Construction andBuilding Materials vol 111 pp 153ndash163 2016


International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018


Active and Passive Electronic Components

VLSI Design



Shock and Vibration


Civil EngineeringAdvances in

Acoustics and VibrationAdvances in



Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of



Journal ofEngineeringVolume 2018

SensorsJournal of



RotatingMachinery


Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018


Chemical EngineeringInternational Journal of Antennas and

Propagation




Navigation and Observation


Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

+e PMS determines maintenance activities in thenetwork- and project-level steps sequentially For the twosteps the work flows are similar but the latter is moredetailed and focused [8] In the network-level step roadagencies detect pavement condition and evaluate pavementperformance mainly using PCI and the riding quality index(RQI) According to the evaluation results they providesome alternative maintenance strategies whose pavementperformance and cost are subsequently predicted After-ward by considering the maintenance budgets environ-mental factors and short- and long-term needs they allocatelimited maintenance resources to road network reasonablyand identify the projects need to be funded first [8 13] Afternetwork-level models have determined pavement sections tobe financed road agencies detect these sections in the projectlevel more accurate and detailed than the network level [14]Subsequently they evaluate pavement performance com-prehensively by many indices such as the pavementstructure strength index (PSSI) in addition to PCI and RQIBased on the evaluation data some alternative maintenancemeasures are suggested then the pavement performance foreach alternative is predicted and the corresponding cost isestimated +rough comparing the cost and effect of everyalternative the most cost-effective maintenance measure isdetermined for the pavement sections +rough the aboveprocess the PMS performs well at the network level forallocating maintenance funds but inadequately for particularpavement sections at the project level [15 16] It is becauseasphalt pavements with different distress types maybepossess the same value of the pavement performance indicesindicating the same maintenance measure may be selectedfor the asphalt pavements However the same one measuregenerally cannot produce good effects for different distresstypes For example two asphalt pavements namely onewith transverse cracking and the other with pothole canown the same value of PSSI As a result slurry seal may beselected for the two asphalt pavements but it is not specificenough for the above two distress types

To maintain asphalt pavements accurately and effec-tively road agencies should consider not only pavementperformance but also particular distress types Manyscholars have explored diverse distress types especiallydifferent cracking for pavement maintenance Due to thecomplexity of cracking Ouma et al [1] analyzed the for-mation and relationships of multiple cracks which werecharacterized by longitudinal transverse diagonal blockand alligator through the wavelet morphology It shows thatthe diverse cracks possess some connections and the re-lationships are important for pavement maintenance butthe detailed relationships are not discussed +e Indianadepartment of transportation (DOT) found numerous statesin the USA cannot manage reflective cracking effectively dueto the lack of systematic procedures for selecting appropriatemaintenance measures Hence they studied the current stateof reflective cracking by a survey of all states DOTs anddeveloped a decision-making process to enhance mainte-nance measures selection [7] In addition to investigatethermal cracking of asphalt pavements based on the datafrom 46 long-term pavement performance (LTPP) sections

Dong et al [17] selected six important influence factors+en the relationships among the six factors and betweenthe six factors and the thermal cracking development ratewere analyzed through association rule mining and a dataminer whose supports and confidences can reflect thestrength of relationships Although this research only fo-cuses on thermal cracking it shows that the association rulesare very suitable for investigating probabilities of mutualinduction between different items namely relationshipsbetween different items Apart from cracking potholes andthe corresponding maintenance methods were exploredthrough a questionnaire survey of six provincial trans-portation agencies in Canada by Biswas et al [18] Mean-while this study also briefly discussed the relationshipsamong potholes map cracking longitudinal crackingbleeding and raveling Furthermore Sollazzo et al [19]probed the relationships between roughness and the pave-ment structural condition with an artificial neural networkmodel which is built and trained based on a huge number ofdata from the LTPP program +is model can give manyclues for road engineers to determine maintenance measuresat the project level in which the pavement structure strengthis very important for selecting maintenance measures

+e above studies about diverse distress types can assistroad engineers to further understand asphalt pavementconditions and select appropriate maintenance measuresHowever the relationships among different distress typeswere not thoroughly analyzed or deeply discussed Differentdistress types possess many relationships with each othersuggesting the existence of one distress type may induceother distress types to occur because of distress causes It canbe further illustrated by an example of three distress typesand two distress causes as presented in Figure 1 For thisexample distress type α occurs on an asphalt pavement dueto distress cause A meanwhile distress cause A can bringabout distress type β as well as distress cause B furthermoredistress cause B can lead to distress type c In summarydistress types α β and c occur on the asphalt pavementtogether Traditionally distress types that exist on pave-ments are considered for pavement maintenance whileother possible distress types are not For this reason aftersome maintenance measures being conducted for the ex-istent failures other distress types may occur at once Asa result the serviceability of asphalt pavements is still poor

+e above analysis shows that the relationships amongdifferent distress types are significant for pavement main-tenance In addition the association rule mining a dataminer that was first introduced by Agrawal et al [20] isa promising tool to analyze relationships between differentitems [17 20 21] Hence this study mainly aims to in-vestigate the relationships among different distress typesusing association rule mining to make pavement mainte-nance accurate and effective

2 Objective and Scope

+emain purpose of this study is to analyze the relationshipsamong different distress types Considering the complexityof distress causes and the limited space this study only













Distress cause A

Distress cause B

Asphalt pavement

Distresstypeγ

Distress typeα

Distress typeβ


Regions surveyed




















1-2


1-3



Sichuan Province 6

1-4
























Asphaltpavements

with two and


Asphaltpavements


RU1489

TC 851

MC

355

PO 284







N(ALL) P(XcupY) (1)




N(X) P(Y | X) (2)



All IDDTs + C-DDTs





MC + PO + C-DDTs

RU + PO + C-DDTs

RU + MC + C-DDTs

TC + PO + C-DDTs

TC + MC + C-DDTs

TC + RU + C-DDTs

PO + C-DDTs

MC + C-DDTs

RU + C-DDTs

TC + C-DDTs

C-DDTs


Asp

halt

pave

men

t with

dist

ress

type

s

0




























4 middot C18 64




TC harr RU

TC harr MC

TC harr PO

RU harr MC

RU harr PO

MC harr PO

0 10 20 30 40Support ()

Average is 2506

Ass

ocia

tion

rule

(a)


0 20 40 60 80 100Confidence ()

Ass

ocia

tion

rule

Average is 5247

(b)



20 100

80

60

40

20

0

15

10

5

0BL BU PSR RO LC

Confidence

Supp

ort (

)

Con

fiden

ce (

)

Support

RA DE PUDDT

(a)


80

60

40

20

0

15

10

5

0

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


(b)20

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

BL BU PSR RODDT

LC RA DE PU

100

80

60

40

20

0

15

10

5

0

(c)


20100

80

60

40

20

0

15

10

5

0

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


(d)

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


20 100

80

60

40

20

0

15

10

5

0

(e)


Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


20 100

80

60

40

20

0

15

10

5

0

(f )

Figure 7 Continued







4 middot C13 +

C14 middot C1





Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

DDT


20 100

80

60

40

20

0

15

10

5

0

(g)


20 100

80

60

40

20

0

15

10

5

0

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

DDT


(h)



Tabl

e4

Associatio

nrulesam

ongDDTsa

mon

gRS

DTsa

ndbetweenDDTs

andRS

DTs

SC

SC

SC

SC

SC

SC

SC

SC

DDTs

DDTs⟶

DDTs

RSDTs⟶

DDTs

LC⟶

DDTs

RA⟶

DDTs

DE⟶

DDTs

PU⟶

DDTs

BL⟶

DDTs

BU⟶

DDTs

PSR⟶

DDTs

RO⟶

DDTs

LC284

1667

426

2500

426

2069

142

2857

213

3333

142

2222

142

6667

RA284

1538

780

4583

567

2759

284

5714

355

5556

071

1111

071

3333

DE

426

2308

780

4583

709

3448

284

5714

355

5556

000

000

071

3333

PU426

2308

567

3333

709

4167

284

5714

213

3333

142

2222

071

3333

RSDTs

DDTs⟶

RSDTs

RSDTs⟶

RSDTs

LC⟶

RSDTs

RA⟶

RSDTs

DE⟶

RSDTs

PU⟶

RSDTs

BL⟶

RSDTs

BU⟶

RSDTs

PSR⟶

RSDTs

RO⟶

RSDTs

BL142

769

284

1667

284

1667

284

1379

284

4444

071

1111

071

3333

BU213

1154

355

2083

355

2083

213

1034

284

5714

071

1111

071

3333

PSR

142

769

071

417

000

000

142

690

071

1429

071

1111

142

6667

RO142

769

071

417

071

417

071

345

071

1429

071

1111

142

2222

SandCdeno

tesupp

orta

ndconfi

denceof


ely

















Depression

RavelingPumping

709

4167

3448

780

4583

4583


2759 3333

567







8 Conclusions













525 40



Transverse cracking





(i)(ii)

(iii)(iv)


(i)(ii)

(iii)(iv)




Data Availability




Acknowledgments


References
















25

20

15

Occ

urre

nce p

roba

bilit

y (

)

10

5

0TC RU MC PO LC RA


DE PU BL BU PSR RO



























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia













Distress cause A

Distress cause B

Asphalt pavement

Distresstypeγ

Distress typeα

Distress typeβ


Regions surveyed




















1-2


1-3



Sichuan Province 6

1-4
























Asphaltpavements

with two and


Asphaltpavements


RU1489

TC 851

MC

355

PO 284







N(ALL) P(XcupY) (1)




N(X) P(Y | X) (2)



All IDDTs + C-DDTs





MC + PO + C-DDTs

RU + PO + C-DDTs

RU + MC + C-DDTs

TC + PO + C-DDTs

TC + MC + C-DDTs

TC + RU + C-DDTs

PO + C-DDTs

MC + C-DDTs

RU + C-DDTs

TC + C-DDTs

C-DDTs


Asp

halt

pave

men

t with

dist

ress

type

s

0




























4 middot C18 64




TC harr RU

TC harr MC

TC harr PO

RU harr MC

RU harr PO

MC harr PO

0 10 20 30 40Support ()

Average is 2506

Ass

ocia

tion

rule

(a)


0 20 40 60 80 100Confidence ()

Ass

ocia

tion

rule

Average is 5247

(b)



20 100

80

60

40

20

0

15

10

5

0BL BU PSR RO LC

Confidence

Supp

ort (

)

Con

fiden

ce (

)

Support

RA DE PUDDT

(a)


80

60

40

20

0

15

10

5

0

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


(b)20

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

BL BU PSR RODDT

LC RA DE PU

100

80

60

40

20

0

15

10

5

0

(c)


20100

80

60

40

20

0

15

10

5

0

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


(d)

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


20 100

80

60

40

20

0

15

10

5

0

(e)


Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


20 100

80

60

40

20

0

15

10

5

0

(f )

Figure 7 Continued







4 middot C13 +

C14 middot C1





Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

DDT


20 100

80

60

40

20

0

15

10

5

0

(g)


20 100

80

60

40

20

0

15

10

5

0

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

DDT


(h)



Tabl

e4

Associatio

nrulesam

ongDDTsa

mon

gRS

DTsa

ndbetweenDDTs

andRS

DTs

SC

SC

SC

SC

SC

SC

SC

SC

DDTs

DDTs⟶

DDTs

RSDTs⟶

DDTs

LC⟶

DDTs

RA⟶

DDTs

DE⟶

DDTs

PU⟶

DDTs

BL⟶

DDTs

BU⟶

DDTs

PSR⟶

DDTs

RO⟶

DDTs

LC284

1667

426

2500

426

2069

142

2857

213

3333

142

2222

142

6667

RA284

1538

780

4583

567

2759

284

5714

355

5556

071

1111

071

3333

DE

426

2308

780

4583

709

3448

284

5714

355

5556

000

000

071

3333

PU426

2308

567

3333

709

4167

284

5714

213

3333

142

2222

071

3333

RSDTs

DDTs⟶

RSDTs

RSDTs⟶

RSDTs

LC⟶

RSDTs

RA⟶

RSDTs

DE⟶

RSDTs

PU⟶

RSDTs

BL⟶

RSDTs

BU⟶

RSDTs

PSR⟶

RSDTs

RO⟶

RSDTs

BL142

769

284

1667

284

1667

284

1379

284

4444

071

1111

071

3333

BU213

1154

355

2083

355

2083

213

1034

284

5714

071

1111

071

3333

PSR

142

769

071

417

000

000

142

690

071

1429

071

1111

142

6667

RO142

769

071

417

071

417

071

345

071

1429

071

1111

142

2222

SandCdeno

tesupp

orta

ndconfi

denceof


ely

















Depression

RavelingPumping

709

4167

3448

780

4583

4583


2759 3333

567







8 Conclusions













525 40



Transverse cracking





(i)(ii)

(iii)(iv)


(i)(ii)

(iii)(iv)




Data Availability




Acknowledgments


References
















25

20

15

Occ

urre

nce p

roba

bilit

y (

)

10

5

0TC RU MC PO LC RA


DE PU BL BU PSR RO



























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia


















1-2


1-3



Sichuan Province 6

1-4
























Asphaltpavements

with two and


Asphaltpavements


RU1489

TC 851

MC

355

PO 284







N(ALL) P(XcupY) (1)




N(X) P(Y | X) (2)



All IDDTs + C-DDTs





MC + PO + C-DDTs

RU + PO + C-DDTs

RU + MC + C-DDTs

TC + PO + C-DDTs

TC + MC + C-DDTs

TC + RU + C-DDTs

PO + C-DDTs

MC + C-DDTs

RU + C-DDTs

TC + C-DDTs

C-DDTs


Asp

halt

pave

men

t with

dist

ress

type

s

0




























4 middot C18 64




TC harr RU

TC harr MC

TC harr PO

RU harr MC

RU harr PO

MC harr PO

0 10 20 30 40Support ()

Average is 2506

Ass

ocia

tion

rule

(a)


0 20 40 60 80 100Confidence ()

Ass

ocia

tion

rule

Average is 5247

(b)



20 100

80

60

40

20

0

15

10

5

0BL BU PSR RO LC

Confidence

Supp

ort (

)

Con

fiden

ce (

)

Support

RA DE PUDDT

(a)


80

60

40

20

0

15

10

5

0

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


(b)20

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

BL BU PSR RODDT

LC RA DE PU

100

80

60

40

20

0

15

10

5

0

(c)


20100

80

60

40

20

0

15

10

5

0

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


(d)

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


20 100

80

60

40

20

0

15

10

5

0

(e)


Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


20 100

80

60

40

20

0

15

10

5

0

(f )

Figure 7 Continued







4 middot C13 +

C14 middot C1





Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

DDT


20 100

80

60

40

20

0

15

10

5

0

(g)


20 100

80

60

40

20

0

15

10

5

0

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

DDT


(h)



Tabl

e4

Associatio

nrulesam

ongDDTsa

mon

gRS

DTsa

ndbetweenDDTs

andRS

DTs

SC

SC

SC

SC

SC

SC

SC

SC

DDTs

DDTs⟶

DDTs

RSDTs⟶

DDTs

LC⟶

DDTs

RA⟶

DDTs

DE⟶

DDTs

PU⟶

DDTs

BL⟶

DDTs

BU⟶

DDTs

PSR⟶

DDTs

RO⟶

DDTs

LC284

1667

426

2500

426

2069

142

2857

213

3333

142

2222

142

6667

RA284

1538

780

4583

567

2759

284

5714

355

5556

071

1111

071

3333

DE

426

2308

780

4583

709

3448

284

5714

355

5556

000

000

071

3333

PU426

2308

567

3333

709

4167

284

5714

213

3333

142

2222

071

3333

RSDTs

DDTs⟶

RSDTs

RSDTs⟶

RSDTs

LC⟶

RSDTs

RA⟶

RSDTs

DE⟶

RSDTs

PU⟶

RSDTs

BL⟶

RSDTs

BU⟶

RSDTs

PSR⟶

RSDTs

RO⟶

RSDTs

BL142

769

284

1667

284

1667

284

1379

284

4444

071

1111

071

3333

BU213

1154

355

2083

355

2083

213

1034

284

5714

071

1111

071

3333

PSR

142

769

071

417

000

000

142

690

071

1429

071

1111

142

6667

RO142

769

071

417

071

417

071

345

071

1429

071

1111

142

2222

SandCdeno

tesupp

orta

ndconfi

denceof


ely

















Depression

RavelingPumping

709

4167

3448

780

4583

4583


2759 3333

567







8 Conclusions













525 40



Transverse cracking





(i)(ii)

(iii)(iv)


(i)(ii)

(iii)(iv)




Data Availability




Acknowledgments


References
















25

20

15

Occ

urre

nce p

roba

bilit

y (

)

10

5

0TC RU MC PO LC RA


DE PU BL BU PSR RO



























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia


















Asphaltpavements

with two and


Asphaltpavements


RU1489

TC 851

MC

355

PO 284







N(ALL) P(XcupY) (1)




N(X) P(Y | X) (2)



All IDDTs + C-DDTs





MC + PO + C-DDTs

RU + PO + C-DDTs

RU + MC + C-DDTs

TC + PO + C-DDTs

TC + MC + C-DDTs

TC + RU + C-DDTs

PO + C-DDTs

MC + C-DDTs

RU + C-DDTs

TC + C-DDTs

C-DDTs


Asp

halt

pave

men

t with

dist

ress

type

s

0




























4 middot C18 64




TC harr RU

TC harr MC

TC harr PO

RU harr MC

RU harr PO

MC harr PO

0 10 20 30 40Support ()

Average is 2506

Ass

ocia

tion

rule

(a)


0 20 40 60 80 100Confidence ()

Ass

ocia

tion

rule

Average is 5247

(b)



20 100

80

60

40

20

0

15

10

5

0BL BU PSR RO LC

Confidence

Supp

ort (

)

Con

fiden

ce (

)

Support

RA DE PUDDT

(a)


80

60

40

20

0

15

10

5

0

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


(b)20

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

BL BU PSR RODDT

LC RA DE PU

100

80

60

40

20

0

15

10

5

0

(c)


20100

80

60

40

20

0

15

10

5

0

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


(d)

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


20 100

80

60

40

20

0

15

10

5

0

(e)


Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


20 100

80

60

40

20

0

15

10

5

0

(f )

Figure 7 Continued







4 middot C13 +

C14 middot C1





Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

DDT


20 100

80

60

40

20

0

15

10

5

0

(g)


20 100

80

60

40

20

0

15

10

5

0

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

DDT


(h)



Tabl

e4

Associatio

nrulesam

ongDDTsa

mon

gRS

DTsa

ndbetweenDDTs

andRS

DTs

SC

SC

SC

SC

SC

SC

SC

SC

DDTs

DDTs⟶

DDTs

RSDTs⟶

DDTs

LC⟶

DDTs

RA⟶

DDTs

DE⟶

DDTs

PU⟶

DDTs

BL⟶

DDTs

BU⟶

DDTs

PSR⟶

DDTs

RO⟶

DDTs

LC284

1667

426

2500

426

2069

142

2857

213

3333

142

2222

142

6667

RA284

1538

780

4583

567

2759

284

5714

355

5556

071

1111

071

3333

DE

426

2308

780

4583

709

3448

284

5714

355

5556

000

000

071

3333

PU426

2308

567

3333

709

4167

284

5714

213

3333

142

2222

071

3333

RSDTs

DDTs⟶

RSDTs

RSDTs⟶

RSDTs

LC⟶

RSDTs

RA⟶

RSDTs

DE⟶

RSDTs

PU⟶

RSDTs

BL⟶

RSDTs

BU⟶

RSDTs

PSR⟶

RSDTs

RO⟶

RSDTs

BL142

769

284

1667

284

1667

284

1379

284

4444

071

1111

071

3333

BU213

1154

355

2083

355

2083

213

1034

284

5714

071

1111

071

3333

PSR

142

769

071

417

000

000

142

690

071

1429

071

1111

142

6667

RO142

769

071

417

071

417

071

345

071

1429

071

1111

142

2222

SandCdeno

tesupp

orta

ndconfi

denceof


ely

















Depression

RavelingPumping

709

4167

3448

780

4583

4583


2759 3333

567







8 Conclusions













525 40



Transverse cracking





(i)(ii)

(iii)(iv)


(i)(ii)

(iii)(iv)




Data Availability




Acknowledgments


References
















25

20

15

Occ

urre

nce p

roba

bilit

y (

)

10

5

0TC RU MC PO LC RA


DE PU BL BU PSR RO



























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia






N(ALL) P(XcupY) (1)




N(X) P(Y | X) (2)



All IDDTs + C-DDTs





MC + PO + C-DDTs

RU + PO + C-DDTs

RU + MC + C-DDTs

TC + PO + C-DDTs

TC + MC + C-DDTs

TC + RU + C-DDTs

PO + C-DDTs

MC + C-DDTs

RU + C-DDTs

TC + C-DDTs

C-DDTs


Asp

halt

pave

men

t with

dist

ress

type

s

0




























4 middot C18 64




TC harr RU

TC harr MC

TC harr PO

RU harr MC

RU harr PO

MC harr PO

0 10 20 30 40Support ()

Average is 2506

Ass

ocia

tion

rule

(a)


0 20 40 60 80 100Confidence ()

Ass

ocia

tion

rule

Average is 5247

(b)



20 100

80

60

40

20

0

15

10

5

0BL BU PSR RO LC

Confidence

Supp

ort (

)

Con

fiden

ce (

)

Support

RA DE PUDDT

(a)


80

60

40

20

0

15

10

5

0

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


(b)20

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

BL BU PSR RODDT

LC RA DE PU

100

80

60

40

20

0

15

10

5

0

(c)


20100

80

60

40

20

0

15

10

5

0

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


(d)

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


20 100

80

60

40

20

0

15

10

5

0

(e)


Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


20 100

80

60

40

20

0

15

10

5

0

(f )

Figure 7 Continued







4 middot C13 +

C14 middot C1





Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

DDT


20 100

80

60

40

20

0

15

10

5

0

(g)


20 100

80

60

40

20

0

15

10

5

0

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

DDT


(h)



Tabl

e4

Associatio

nrulesam

ongDDTsa

mon

gRS

DTsa

ndbetweenDDTs

andRS

DTs

SC

SC

SC

SC

SC

SC

SC

SC

DDTs

DDTs⟶

DDTs

RSDTs⟶

DDTs

LC⟶

DDTs

RA⟶

DDTs

DE⟶

DDTs

PU⟶

DDTs

BL⟶

DDTs

BU⟶

DDTs

PSR⟶

DDTs

RO⟶

DDTs

LC284

1667

426

2500

426

2069

142

2857

213

3333

142

2222

142

6667

RA284

1538

780

4583

567

2759

284

5714

355

5556

071

1111

071

3333

DE

426

2308

780

4583

709

3448

284

5714

355

5556

000

000

071

3333

PU426

2308

567

3333

709

4167

284

5714

213

3333

142

2222

071

3333

RSDTs

DDTs⟶

RSDTs

RSDTs⟶

RSDTs

LC⟶

RSDTs

RA⟶

RSDTs

DE⟶

RSDTs

PU⟶

RSDTs

BL⟶

RSDTs

BU⟶

RSDTs

PSR⟶

RSDTs

RO⟶

RSDTs

BL142

769

284

1667

284

1667

284

1379

284

4444

071

1111

071

3333

BU213

1154

355

2083

355

2083

213

1034

284

5714

071

1111

071

3333

PSR

142

769

071

417

000

000

142

690

071

1429

071

1111

142

6667

RO142

769

071

417

071

417

071

345

071

1429

071

1111

142

2222

SandCdeno

tesupp

orta

ndconfi

denceof


ely

















Depression

RavelingPumping

709

4167

3448

780

4583

4583


2759 3333

567







8 Conclusions













525 40



Transverse cracking





(i)(ii)

(iii)(iv)


(i)(ii)

(iii)(iv)




Data Availability




Acknowledgments


References
















25

20

15

Occ

urre

nce p

roba

bilit

y (

)

10

5

0TC RU MC PO LC RA


DE PU BL BU PSR RO



























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia























4 middot C18 64




TC harr RU

TC harr MC

TC harr PO

RU harr MC

RU harr PO

MC harr PO

0 10 20 30 40Support ()

Average is 2506

Ass

ocia

tion

rule

(a)


0 20 40 60 80 100Confidence ()

Ass

ocia

tion

rule

Average is 5247

(b)



20 100

80

60

40

20

0

15

10

5

0BL BU PSR RO LC

Confidence

Supp

ort (

)

Con

fiden

ce (

)

Support

RA DE PUDDT

(a)


80

60

40

20

0

15

10

5

0

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


(b)20

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

BL BU PSR RODDT

LC RA DE PU

100

80

60

40

20

0

15

10

5

0

(c)


20100

80

60

40

20

0

15

10

5

0

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


(d)

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


20 100

80

60

40

20

0

15

10

5

0

(e)


Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


20 100

80

60

40

20

0

15

10

5

0

(f )

Figure 7 Continued







4 middot C13 +

C14 middot C1





Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

DDT


20 100

80

60

40

20

0

15

10

5

0

(g)


20 100

80

60

40

20

0

15

10

5

0

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

DDT


(h)



Tabl

e4

Associatio

nrulesam

ongDDTsa

mon

gRS

DTsa

ndbetweenDDTs

andRS

DTs

SC

SC

SC

SC

SC

SC

SC

SC

DDTs

DDTs⟶

DDTs

RSDTs⟶

DDTs

LC⟶

DDTs

RA⟶

DDTs

DE⟶

DDTs

PU⟶

DDTs

BL⟶

DDTs

BU⟶

DDTs

PSR⟶

DDTs

RO⟶

DDTs

LC284

1667

426

2500

426

2069

142

2857

213

3333

142

2222

142

6667

RA284

1538

780

4583

567

2759

284

5714

355

5556

071

1111

071

3333

DE

426

2308

780

4583

709

3448

284

5714

355

5556

000

000

071

3333

PU426

2308

567

3333

709

4167

284

5714

213

3333

142

2222

071

3333

RSDTs

DDTs⟶

RSDTs

RSDTs⟶

RSDTs

LC⟶

RSDTs

RA⟶

RSDTs

DE⟶

RSDTs

PU⟶

RSDTs

BL⟶

RSDTs

BU⟶

RSDTs

PSR⟶

RSDTs

RO⟶

RSDTs

BL142

769

284

1667

284

1667

284

1379

284

4444

071

1111

071

3333

BU213

1154

355

2083

355

2083

213

1034

284

5714

071

1111

071

3333

PSR

142

769

071

417

000

000

142

690

071

1429

071

1111

142

6667

RO142

769

071

417

071

417

071

345

071

1429

071

1111

142

2222

SandCdeno

tesupp

orta

ndconfi

denceof


ely

















Depression

RavelingPumping

709

4167

3448

780

4583

4583


2759 3333

567







8 Conclusions













525 40



Transverse cracking





(i)(ii)

(iii)(iv)


(i)(ii)

(iii)(iv)




Data Availability




Acknowledgments


References
















25

20

15

Occ

urre

nce p

roba

bilit

y (

)

10

5

0TC RU MC PO LC RA


DE PU BL BU PSR RO



























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia


20 100

80

60

40

20

0

15

10

5

0BL BU PSR RO LC

Confidence

Supp

ort (

)

Con

fiden

ce (

)

Support

RA DE PUDDT

(a)


80

60

40

20

0

15

10

5

0

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


(b)20

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

BL BU PSR RODDT

LC RA DE PU

100

80

60

40

20

0

15

10

5

0

(c)


20100

80

60

40

20

0

15

10

5

0

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


(d)

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


20 100

80

60

40

20

0

15

10

5

0

(e)


Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support


20 100

80

60

40

20

0

15

10

5

0

(f )

Figure 7 Continued







4 middot C13 +

C14 middot C1





Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

DDT


20 100

80

60

40

20

0

15

10

5

0

(g)


20 100

80

60

40

20

0

15

10

5

0

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

DDT


(h)



Tabl

e4

Associatio

nrulesam

ongDDTsa

mon

gRS

DTsa

ndbetweenDDTs

andRS

DTs

SC

SC

SC

SC

SC

SC

SC

SC

DDTs

DDTs⟶

DDTs

RSDTs⟶

DDTs

LC⟶

DDTs

RA⟶

DDTs

DE⟶

DDTs

PU⟶

DDTs

BL⟶

DDTs

BU⟶

DDTs

PSR⟶

DDTs

RO⟶

DDTs

LC284

1667

426

2500

426

2069

142

2857

213

3333

142

2222

142

6667

RA284

1538

780

4583

567

2759

284

5714

355

5556

071

1111

071

3333

DE

426

2308

780

4583

709

3448

284

5714

355

5556

000

000

071

3333

PU426

2308

567

3333

709

4167

284

5714

213

3333

142

2222

071

3333

RSDTs

DDTs⟶

RSDTs

RSDTs⟶

RSDTs

LC⟶

RSDTs

RA⟶

RSDTs

DE⟶

RSDTs

PU⟶

RSDTs

BL⟶

RSDTs

BU⟶

RSDTs

PSR⟶

RSDTs

RO⟶

RSDTs

BL142

769

284

1667

284

1667

284

1379

284

4444

071

1111

071

3333

BU213

1154

355

2083

355

2083

213

1034

284

5714

071

1111

071

3333

PSR

142

769

071

417

000

000

142

690

071

1429

071

1111

142

6667

RO142

769

071

417

071

417

071

345

071

1429

071

1111

142

2222

SandCdeno

tesupp

orta

ndconfi

denceof


ely

















Depression

RavelingPumping

709

4167

3448

780

4583

4583


2759 3333

567







8 Conclusions













525 40



Transverse cracking





(i)(ii)

(iii)(iv)


(i)(ii)

(iii)(iv)




Data Availability




Acknowledgments


References
















25

20

15

Occ

urre

nce p

roba

bilit

y (

)

10

5

0TC RU MC PO LC RA


DE PU BL BU PSR RO



























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia







4 middot C13 +

C14 middot C1





Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

DDT


20 100

80

60

40

20

0

15

10

5

0

(g)


20 100

80

60

40

20

0

15

10

5

0

Supp

ort (

)

Con

fiden

ce (

)

Confidence

Support

DDT


(h)



Tabl

e4

Associatio

nrulesam

ongDDTsa

mon

gRS

DTsa

ndbetweenDDTs

andRS

DTs

SC

SC

SC

SC

SC

SC

SC

SC

DDTs

DDTs⟶

DDTs

RSDTs⟶

DDTs

LC⟶

DDTs

RA⟶

DDTs

DE⟶

DDTs

PU⟶

DDTs

BL⟶

DDTs

BU⟶

DDTs

PSR⟶

DDTs

RO⟶

DDTs

LC284

1667

426

2500

426

2069

142

2857

213

3333

142

2222

142

6667

RA284

1538

780

4583

567

2759

284

5714

355

5556

071

1111

071

3333

DE

426

2308

780

4583

709

3448

284

5714

355

5556

000

000

071

3333

PU426

2308

567

3333

709

4167

284

5714

213

3333

142

2222

071

3333

RSDTs

DDTs⟶

RSDTs

RSDTs⟶

RSDTs

LC⟶

RSDTs

RA⟶

RSDTs

DE⟶

RSDTs

PU⟶

RSDTs

BL⟶

RSDTs

BU⟶

RSDTs

PSR⟶

RSDTs

RO⟶

RSDTs

BL142

769

284

1667

284

1667

284

1379

284

4444

071

1111

071

3333

BU213

1154

355

2083

355

2083

213

1034

284

5714

071

1111

071

3333

PSR

142

769

071

417

000

000

142

690

071

1429

071

1111

142

6667

RO142

769

071

417

071

417

071

345

071

1429

071

1111

142

2222

SandCdeno

tesupp

orta

ndconfi

denceof


ely

















Depression

RavelingPumping

709

4167

3448

780

4583

4583


2759 3333

567







8 Conclusions













525 40



Transverse cracking





(i)(ii)

(iii)(iv)


(i)(ii)

(iii)(iv)




Data Availability




Acknowledgments


References
















25

20

15

Occ

urre

nce p

roba

bilit

y (

)

10

5

0TC RU MC PO LC RA


DE PU BL BU PSR RO



























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia


Tabl

e4

Associatio

nrulesam

ongDDTsa

mon

gRS

DTsa

ndbetweenDDTs

andRS

DTs

SC

SC

SC

SC

SC

SC

SC

SC

DDTs

DDTs⟶

DDTs

RSDTs⟶

DDTs

LC⟶

DDTs

RA⟶

DDTs

DE⟶

DDTs

PU⟶

DDTs

BL⟶

DDTs

BU⟶

DDTs

PSR⟶

DDTs

RO⟶

DDTs

LC284

1667

426

2500

426

2069

142

2857

213

3333

142

2222

142

6667

RA284

1538

780

4583

567

2759

284

5714

355

5556

071

1111

071

3333

DE

426

2308

780

4583

709

3448

284

5714

355

5556

000

000

071

3333

PU426

2308

567

3333

709

4167

284

5714

213

3333

142

2222

071

3333

RSDTs

DDTs⟶

RSDTs

RSDTs⟶

RSDTs

LC⟶

RSDTs

RA⟶

RSDTs

DE⟶

RSDTs

PU⟶

RSDTs

BL⟶

RSDTs

BU⟶

RSDTs

PSR⟶

RSDTs

RO⟶

RSDTs

BL142

769

284

1667

284

1667

284

1379

284

4444

071

1111

071

3333

BU213

1154

355

2083

355

2083

213

1034

284

5714

071

1111

071

3333

PSR

142

769

071

417

000

000

142

690

071

1429

071

1111

142

6667

RO142

769

071

417

071

417

071

345

071

1429

071

1111

142

2222

SandCdeno

tesupp

orta

ndconfi

denceof


ely

















Depression

RavelingPumping

709

4167

3448

780

4583

4583


2759 3333

567







8 Conclusions













525 40



Transverse cracking





(i)(ii)

(iii)(iv)


(i)(ii)

(iii)(iv)




Data Availability




Acknowledgments


References
















25

20

15

Occ

urre

nce p

roba

bilit

y (

)

10

5

0TC RU MC PO LC RA


DE PU BL BU PSR RO



























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia

















Depression

RavelingPumping

709

4167

3448

780

4583

4583


2759 3333

567







8 Conclusions













525 40



Transverse cracking





(i)(ii)

(iii)(iv)


(i)(ii)

(iii)(iv)




Data Availability




Acknowledgments


References
















25

20

15

Occ

urre

nce p

roba

bilit

y (

)

10

5

0TC RU MC PO LC RA


DE PU BL BU PSR RO



























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia






8 Conclusions













525 40



Transverse cracking





(i)(ii)

(iii)(iv)


(i)(ii)

(iii)(iv)




Data Availability




Acknowledgments


References
















25

20

15

Occ

urre

nce p

roba

bilit

y (

)

10

5

0TC RU MC PO LC RA


DE PU BL BU PSR RO



























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia



Data Availability




Acknowledgments


References
















25

20

15

Occ

urre

nce p

roba

bilit

y (

)

10

5

0TC RU MC PO LC RA


DE PU BL BU PSR RO



























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia


























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia


Research on Relationships among Different Distress Types ...

Documents

Transcript of Research on Relationships among Different Distress Types ...