Parham Paydavosi, Mohammad Saied Dehghani, Sue McNeil
{"title":"Evaluating the accuracy of predicted bridge condition using machine learning: the role of condition history","authors":"Parham Paydavosi, Mohammad Saied Dehghani, Sue McNeil","doi":"10.1080/15732479.2023.2274878","DOIUrl":null,"url":null,"abstract":"AbstractEffective maintenance decisions for bridges depend on accurate performance prediction. Machine learning (ML) models use historical bridge performance data to learn and predict performance. However, in many agencies, the condition history of bridges is limited and does not go beyond a few years. The question, therefore, is, to what extent does condition history help us make better predictions? To address this question, a ML model was developed that analysed more than 600,000 bridge decks with 27 years of condition history. Two data selection methods were designed: non-overlapping and overlapping data. The non-overlapping data are typically used to train the model. The overlapping data introduced in this study uses the data more efficiently for model training recognising that strings of historical data convey more information. Longer term predictions were found to be positively impacted by every additional year of condition history. Short-term condition prediction (one or two years) does not need significant historical data. It was also found that overlapping data, compared to non-overlapping data, produced larger training samples and had higher prediction accuracy in the majority of experiments, but at the cost of higher running time due to a larger sample size.Keywords: Artificial intelligencebig datainfrastructure asset managementmachine learningbridge structure deteriorationbridge conditionneural networkperformance prediction Disclosure statementNo potential conflict of interest was reported by the author(s).","PeriodicalId":49468,"journal":{"name":"Structure and Infrastructure Engineering","volume":null,"pages":null},"PeriodicalIF":2.6000,"publicationDate":"2023-11-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Structure and Infrastructure Engineering","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1080/15732479.2023.2274878","RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"ENGINEERING, CIVIL","Score":null,"Total":0}
引用次数: 0
Abstract
AbstractEffective maintenance decisions for bridges depend on accurate performance prediction. Machine learning (ML) models use historical bridge performance data to learn and predict performance. However, in many agencies, the condition history of bridges is limited and does not go beyond a few years. The question, therefore, is, to what extent does condition history help us make better predictions? To address this question, a ML model was developed that analysed more than 600,000 bridge decks with 27 years of condition history. Two data selection methods were designed: non-overlapping and overlapping data. The non-overlapping data are typically used to train the model. The overlapping data introduced in this study uses the data more efficiently for model training recognising that strings of historical data convey more information. Longer term predictions were found to be positively impacted by every additional year of condition history. Short-term condition prediction (one or two years) does not need significant historical data. It was also found that overlapping data, compared to non-overlapping data, produced larger training samples and had higher prediction accuracy in the majority of experiments, but at the cost of higher running time due to a larger sample size.Keywords: Artificial intelligencebig datainfrastructure asset managementmachine learningbridge structure deteriorationbridge conditionneural networkperformance prediction Disclosure statementNo potential conflict of interest was reported by the author(s).
期刊介绍:
Structure and Infrastructure Engineering - Maintenance, Management, Life-Cycle Design and Performance is an international Journal dedicated to recent advances in maintenance, management and life-cycle performance of a wide range of infrastructures, such as: buildings, bridges, dams, railways, underground constructions, offshore platforms, pipelines, naval vessels, ocean structures, nuclear power plants, airplanes and other types of structures including aerospace and automotive structures.
The Journal presents research and developments on the most advanced technologies for analyzing, predicting and optimizing infrastructure performance. The main gaps to be filled are those between researchers and practitioners in maintenance, management and life-cycle performance of infrastructure systems, and those between professionals working on different types of infrastructures. To this end, the journal will provide a forum for a broad blend of scientific, technical and practical papers. The journal is endorsed by the International Association for Life-Cycle Civil Engineering ( IALCCE) and the International Association for Bridge Maintenance and Safety ( IABMAS).