Case study: implementation of Predictive Maintenance strategy for the generating unit
1 June 2020

Case study: implementation of Predictive Maintenance strategy for the generating unit

Project: implementation of Predictive Maintenance strategy for the generating unit

Challenge

The key task of energy companies is to ensure a stable and uninterrupted level of energy supply. Failures of machines can cause weeks of downtime and (consequently) huge financial losses. The objective of cooperation between companies was to minimize the number of failures in the generating unit (610 MW) at the power plant. Due to heavy operating conditions inside the fluidized-bed boiler, there were reductions in cross-sections of pipe walls with water under high-pressure. This led to explosion and damage in the inner part of the boiler and (as a consequence) costly downtime.

Solution

The task of specialists from ReliaSol (Reliability Solutions) was to develop a predictive model for detecting failures that cause downtime of the installation. An additional challenge was the need to locate damage places.

In the first phase, approx. 180 measurements supplying one hierarchical model that predicts 8 types of failures were used. The model locates one of the types of failure in 6 potential sectors of the boiler. The RSIMS solution has been integrated and hidden under the hood of the DCS system, and the results of the prediction are presented in the internal desktop of the power plant crew.

The implementation of the next phase of the project enables to expand the system to approximately 1000 measurements and 150 models for the monitoring of the condition of objects in various installations of the boiler. Models monitor the current technical conditions of critical machines, provide information about work anomalies and predict events such as vibrations anomalies or rapid increases of temperatures. 

As a result of works, a desktop of the RSIMS system was implemented and launched. It reports the results of the prediction on an ongoing basis. Key facilities have been planned on the property management scheme dedicated and adapted to the needs of the power plant. Appropriate predictive models have been attached to these facilities. The report desktop (along with dedicated graphics presenting the results returned by models) has been adapted in a way enabling easy operation for operators. It is intuitively connected with schemes of the DCS system. The reporting module along with e-mail notifications enables remote control of undesirable situations without a constant focus on the system.

Benefits

The PdM solution is able to predict 100% of failures within a time horizon (3-17 hours). It can indicate the place of failure. This is a faster and more accurate view of the machine’s operation than any other monitoring system.

The predictive diagnostics system in the form of the RSIMS (Reliability Solutions Intelligence Maintenance System) platform, based on artificial intelligence, reduces costs of removing failures. Thanks to additional hours, during which the generating unit may be available, it increases revenues from the operational power reserve and revenues from the energy market.

  • Constant monitoring and insight into the operation of key facilities included in the generating unit,
  • 1 TB of analyzed data,
  • Approx. 175 implemented predictive models and anomalies.

Solutions based on artificial intelligence support the functioning of modern industry – they enable to predict failures of devices, study different scenarios for dealing with problems and suggesting best practices. Cooperation with RS is a guarantee of the highest-quality services based on expert knowledge and experience in the industry.

Older post
Case study: implementation of Predictive Maintenance strategy for a steam turbine