Predictive Asset Management
Quick Links
Optimizing Satellite Constellations: A Case Study on ATS MDP Model Application
Probabilistic Risk Analysis of a Substation
- Case Study 4
Predictive Asset Management
Hospital Operations and Personnel Management
Drug and Medical Technology Evaluation
Pandemic Preparedness and Response
All organizations that operate complex systems manage their systems through an asset management program. Asset management is a complex discipline, with many facets. Understanding asset probabilities of failure, scheduling inspections, and prioritizing capital replacements are all important aspects of asset management.
The ultimate goal of asset management, and one which ATS can help actualize within organizations, is predictive asset management. Most organizations already have the data needed, and ATS can unlock the information within those data, to predict failure before they occur.
ATS has built predictive asset management capabilities for our customers in the electric and gas utility industry. In one effort, ATS worked with a West coast based utility to perform a fault mitigation study. Specifically, ATS performed an advanced data analysis to determine mitigation effectiveness (ME) values for various fault mitigations.
Equipment: distribution poles, spans, transmission structures, and tie lines for entire utility network
Approximately 200,000 poles and structures, 200,000 spans and tie lines, 200 substations, 800 circuits
Fault mitigations: Inspections, hardening, line clearing, etc.
In calculating ME, ATS used our artificial intelligence analysis engine to find asset and circuit probability of failure values, along with Birnbaum importance measures.
Our results proved that several circuits carried a disproportionate amount of failure risk compared to others.
Birnbaum Importance Value by Circuit
By using a machine learning algorithm to identify the relationship between Birnbaum importance and future outages, ATS was able to derive a regression model that predicts future failures.
The top four circuits identified as the highest risk circuits by our artificial intelligence analysis engine contained the 10 riskiest distribution poles in the utility’s entire network. These 10 poles experienced 33-times more outages than all other poles in the network in subsequent years after the analysis.
In addition, the four riskiest circuits identified by our artificial intelligence analysis engine correctly predicted 6 ignition events. This is 10-times more accurate compared to random selection of circuits to predict future ignitions.
Our predictive asset management can save millions of dollars in avoided risk costs, including reliability costs from lost load, financial costs associated with outage recovery, and even safety costs.
