Risk-informed Asset Management Research Articles

A Two-Stage Data-Driven Spatiotemporal Analysis to Predict Failure Risk of Urban Sewer Systems Leveraging Machine Learning Algorithms.

Risk-informed asset management is key to maintaining optimal performance and efficiency of urban sewer systems. Although sewer system failures are spatiotemporal in nature, previous studies analyzed failure risk from a unidimensional aspect (either spatial or temporal), not accounting for bidimensional spatiotemporal complexities. This is owing to the insufficiency of good-quality data, which ultimately leads to under-/overestimation of failure risk. Here, we propose a generalized methodology/framework to facilitate a robust spatiotemporal analysis of urban sewer system failure risk, overcoming the intrinsic challenges of data imperfections-e.g., missing data, outliers, and imbalanced information. The framework includes a two-stage data-driven modeling technique that efficiently models the highly right-skewed sewer system failure data to predict the failure risk, leveraging a bidimensional space-time approach. We implemented our analysis for Bogotá, the capital city of Colombia. We train, test, and validate a battery of machine learning algorithms-logistic regression, decision trees, random forests, and XGBoost-and select the best model in terms of goodness-of-fit and predictive accuracy. Finally, we illustrate the applicability of the framework in planning/scheduling sewer system maintenance operations using state-of-the-art optimization techniques. Our proposed framework can help stakeholders to analyze the failure-risk models' performance under different discrimination thresholds, and provide managerial insights on the model's adequate spatial resolution and appropriateness of decentralized management for sewer systemmaintenance.

Risk-informed asset management to tackle scouring on bridges across transport networks

Scour is one of the main causes of bridge failures resulting in significant macro-economic impacts, often beyond the direct costs of infrastructure damage. Given the pressure to increase the resilience of transport networks, ageing bridge infrastructure, constrained budgets, variable knowledge of asset conditions and limited data, mixed ownership and operation of bridges, and concerns about the risks of climate change, there is a need to implement cost-effective monitoring and maintenance strategies. To this end, this study aims to set the scene for a risk-informed approach for tackling bridge scour, while considering the socio-economic impacts of disruptions due to bridge failures or closures. This study reviews the current practices in predicting, monitoring and managing bridge scour. It discusses the development of a risk-informed approach to aid the whole-life appraisal of bridges while considering the direct and indirect costs associated with bridge failure or closures. The approach provides a rational means to enable asset managers to evaluate the factors that affect bridge failure risk, select and prioritise appropriate mitigation measures, thereby improving the allocation of scarce monitoring and maintenance resources.

A two-time-scale point process model of water main breaks for infrastructure asset management

Deterioration modelling has been a bottlenecking step towards risk-informed asset management of municipal water distribution networks. To close the gap, we proposed a two-time-scale (TTS) point process model on a pipe level for modelling and prediction of water main breaks. This paper presents the characterization, statistical parameter estimation, probabilistic features, and application of the model. Combining Poisson and renewal models into one, the proposed TTS process is characterized by a conditional intensity function of two time variables—one in a pipe clock for overall pipe aging and the other in a repair clock for local renewal. As a result, different aging patterns including the complicated bathtub-type behaviour can be modelled. A novel statistical method that combines data augmentation and Markov Chain Monte Carlo was developed for model estimation to deal with partially missing event histories. A case study using real-life data collected from a regional municipality in Canada was presented to illustrate the application of the proposed model. The modelling process ranging from model estimation, verification, validation, and updating to application in asset management was thoroughly demonstrated. This study also demonstrated that one must use the full distributions of the parameters to obtain an unbiased prediction of mean number of water main breaks. The proposed model was also compared with the Poisson process model in terms of break intensity, survival probability, mean cumulative number of breaks, and mean annual number of breaks. The implications of the different results to asset management were carefully discussed as well. Last, the ability of the proposed model to capture the maintenance effectiveness of pipe repair was proven. This work represents a solid advancement towards holistic assessment of the aging risk of a municipal water distribution network.

Evaluation of artificial intelligence tool performance and uncertainty for predicting sewer structural condition

The implementation of a risk-informed asset management system by a wastewater infrastructure utility requires information regarding the probability and the consequences of component failures. This paper focuses on the former, evaluating the performance of artificial intelligence tools, namely artificial neural networks (ANNs) and support vector machines (SVMs), in predicting the structural condition of sewers. The performance of these tools is compared with that of logistic regression on the case study of the wastewater infrastructures of SANEST — Sistema de Saneamento da Costa do Estoril (Costa do Estoril Wastewater System). The uncertainty associated to ANNs and SVMs is quantified and the results of a trial and error approach and the use of optimization algorithms to develop SVMs are compared. The results highlight the need to account for both the performance and the uncertainty in the process of choosing the best model to estimate the sewer condition, since the ANNs present the highest average performance (78.5% correct predictions in the test sample) but also the highest dispersion of performance results (73% to 81% correct predictions in the test sample), whereas the SVMs have lower average performance (71.1% without optimization and 72.6% with the parameters optimized using the Covariance Matrix Adaptation Evolution Strategy) but little variability.

Development of risk-informed, performance-based asset management in mining

This paper discusses possibilities of developing a holistic risk-informed, performance-based asset management in mining (RIPBAMM). This process would consist of modelling and probabilistic quantification regarding decision support performance indicators. It assists decision-makers in determining not only which mine improvement investment options should be implemented, but also how to prioritize resources for their implementation based on their predicted levels of profitability. The RIPBAMM approach will complement and integrate existing main mine activities such as exploration, ore body modelling, mine design, planning and scheduling, exploitation (all the phases of the mine life), mineral treatment, cost and market model, operational safety and health, environmental issues, mining equipment reliability and maintenance process, equipment selection model, security, etc. RIPBAMM will involve an integrated assessment of dominant influence factors and performance measures related to mining operations. This process is intended to maximize both net present value (NPV) of the mine, and long-term profitability through a continuous support to a decision-making process. It may be particularly useful while optimizing several mine sites belonging to the same mining company. Initial risk informed asset management (RIAM) applications have been initially developed for the nuclear power industry. Afterwards, this process has been adapted to provide decision-making support to other types of power stations, complex facilities (usually capital-intensive), or even groups of such facilities across a wide variety of industries. RIPBAMM is introducing numerous (stochastic) models and supporting performance metrics that can ultimately be employed in order to support decisions that affect the allocation and management of mine resources (i.e. financial support, employment, scheduling, etc.).

Using Risk-Informed Asset Management for Feedwater System Preventative Maintenance Optimization

The initial development of a South Texas Project Nuclear Operating Company process for supporting preventative maintenance optimization by applying the Balance-Of-Plant model and Risk-Informed Asset Management alpha-level software applications is presented. Preventative maintenance activities are evaluated in the South Texas Project Risk- Informed Asset Management software while the plant maintains or improves upon high levels of nuclear safety. In the Balance-Of-Plant availability application, the level of detail in the feedwater system is enhanced to support plant decision-making at the component failure mode and human error mode level of indenture by elaborating on the current model at the super-component level of indenture. The enhanced model and modeling techniques are presented. Results of case studies in feedwater system preventative maintenance optimization using plant-specific data are also presented.

ICONE11-36121 PRACTICAL APPLICATION OF DECISION SUPPORT METRICS FOR POWER PLANT RISK-INFORMED ASSET MANAGEMENT

The objective of this paper is to provide electric utilities with a concept for developing and applying effective decision support metrics via integrated risk-informed asset management (RIAM) programs for power stations and generating companies. RIAM is a process by which analysts review historical performance and develop predictive logic models and data analyses to predict critical decision support figures-of-merit (or metrics) for generating station managers and electric utility company executives. These metrics include, but are not limited to, the following : profitability, net benefit, benefit-to-cost ratio, projected return on investment, projected revenue, projected costs, asset value, safety (catastrophic facility damage frequency and consequences, etc.), power production availability (capacity factor, etc.), efficiency (heat rate), and others. RIAM applies probabilistic safety assessment (PSA) techniques and generates predictions probabilistically so that metrics information can be supplied to managers in terms of probability distributions as well as point estimates. This enables the managers to apply the concept of "confidence levels" in their critical decision-making processes.