Enterprise Asset Management is an emerging strategic tool enabling optimal utilization of assets contributing to business goals of an organization. Come join us as we uncover the hidden potential of Asset Management with our deep insights spanning industry domains and technology

« Role of Supply Chain in Effective Maintenance and Asset Management - Part II | Main | Successful EAM Implementation - Prepare & Plan »

Predicting Asset Failure Using Similarity Based Modeling - Part I

Similarity Based Modeling in conjunction with Predictive Analytics Solution helps asset intensive industries achieve and exceed asset performance, improve reliability by providing prior warnings and notifications on upcoming failures, resulting in ample time for planning, scheduling and executing maintenance jobs. Similarity Based Modeling compares a calculated feature set to a modeled feature set over all operating conditions and thus enables the identification of particular fault modes early in their progression. Overall it optimizes maintenance activities and reduces maintenance cost expenditures helping organizations in gaining competitive advantage .Organizations will then be able to focus on proactive maintenance rather on reactive maintenance.


Asset Management: The Utilities Industry Perspective
Privatization in the early 20th century opened the doors for a flurry of reforms in the utilities industry. Asset intensive industries like utilities started to tighten focus on asset management areas to improve reliability and availability of assets, enhance safety and better compliance to statutory/regulatory needs to provide better and cheaper services to customers while also satisfying the shareholders and regulators. Traditionally in the utilities industry, the asset related information was held in different parts of the business and each part was functioning as individual entities operating in silos.. Asset Management encompassed the need for integrating the available information into a single coherent system. The focus on plant efficiency, reliability of equipment, reducing the downtime and the total cost of operating the plant formed the basis of asset management in utilities industry. Harnessing the sheer volume of data generated and making informed decisions will provide enormous avenues for growth of asset intensive industry as a whole. With the advent of technology this data can add value by helping the industry to gain strategic insights by combining real time data with historic data and also take action to apply what has been learned from its cumulative experience. The effective utilization of this information brings about substantial tangible benefits to the bottom line, thereby improving the operational and financial performance. Integrating asset management with other business processes allows end-to-end asset management, proper risk analysis and targeting the maintenance programs to areas where the need really exists, thus enhancing the assets performance.

Operational Challenges

Predictive Analytics Blog-2.png 

Business Challenges

Predictive Analytics Blog-1.png 

Predictive Analytics in EAM
The maturity level of asset management in organizations is evident from its focused shift from Reactive Maintenance strategy to Proactive Maintenance strategy. In the era of technological advancements, organizations focus on continuous improvements by driving innovation, automation, and digitalization. The amount of data being retrieved, recorded, and trended is growing at a fast pace due to centralized monitoring of the plants. The plant engineer has the uphill task to make sense from this ever increasing data to identify the failure chances to critical equipment's of the plant. The central step in increasing the availability and reliability of an asset is by detecting the abnormal behavior of any operational parameter before it affects the operation. The early detection of emerging threats, which could be precursors to potential failure if allowed to progress further, enhances the reliability and availability of the asset. This is made possible by the combination of predictive analytics with the maintenance methodology which enables the plant engineers and maintenance personnel to move into proactive mode. The advanced notification gives more time to evaluate and plan corrective action that can mitigate the failure of the components and alleviates the need to perform reactive maintenance. Automation in areas of asset operation & maintenance, data capture and monitoring has paved way for smarter technology solutions in Enterprise Asset Management. A Smarter Technology solution makes organizations smarter and sustainable in delivering client value by overcoming the business challenges. Predictive Analytics in Asset Management helps to improve the plant performance, availability and reliability by utilizing statistical techniques. Refined Maintenance strategies combined with Predictive Analytics tools will help asset intensive organizations to overcome the business and operational asset management challenges.
Predictive Analytics uses certain models that combine the real time data with historical data to predict potential asset failure and drive from reactive to proactive maintenance. The Similarity Based Modeling is one methodology which uses predictive analytics to predict equipment failures and this in concurrence with the maintenance methodology offer better asset management to asset intensive industries.

Similarity Based Modeling (SBM)
Similarity Based Modeling (SBM) uses predictive analytics, sensor information around an asset and determining the changes in behavior from the historical operation of that asset. Specialized software packages that have predictive analytics capabilities are used to analyze in near real time, the data collected from different sensors in the asset. Identification of a normal pattern or expected pattern in the behavior of signals from the sensors under set operating conditions is the foremost step in SBM. This is achieved by grouping related signals and then analyzing the historical operating data of the asset under consideration. Once this is achieved, the identified pattern is then uploaded in the software system and is used for analysis of each sample of data values collected from the sensor. From the pattern an estimate of each signal's behavior is generated and compared to the actual real-time value. An empirical threshold that has been defined by the OEMs is also made available in the software system. The difference between the actual and the expected value is constantly compared to empirical thresholds defined in the system. During comparison if the difference in the actual value and expected value overshoots the empirical thresholds, we have a pattern of abnormal behavior. When the software identifies the abnormal behavior under ambient operating conditions, preset rules are fired which alerts the plant engineer of impending failures in the asset. The plant engineer focuses his attention to the abnormal behavior to find the cause and formulate the best response which is relayed to maintenance personnel to perform maintenance activity before the failure occurs.

We will cover more about SBM with an example in the next part of this blog.

Post a comment

(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)

Please key in the two words you see in the box to validate your identity as an authentic user and reduce spam.