Improving Asset Reliability with Seeq Condition Monitoring
RWE, a global leader in power generation, is transitioning to renewable energy with the goal of becoming carbon-neutral by 2040.
RWE is one of the leading European gas power providers, and the organization plans to use its existing gas generation plants to drive its transition to renewable energy. However, that puts increased demand on those plants to ensure they can provide stable backup for renewable energy sources.
With that in mind, RWE Generation wanted to optimize its gas generation plants. The company sought greater flexibility and availability across its gas fleet. RWE partnered with Seeq to build a predictive maintenance system and pinpoint problems quickly. Seeq, an AI-powered analytics platform, transformed operational data into insights, enabling rapid fault detection that yielded greater uptime and productivity.
Seeq Condition Monitoring and the Challenges of RWE Generation
RWE was already using asset condition monitoring tools to track the health of critical assets. Condition monitoring tracks equipment health and environmental factors in real-time so that maintenance teams can make repairs at the first sign of a problem. For example, if a machine overheats or begins to vibrate excessively, the condition monitoring process will alert technicians right away. Condition monitoring systems typically use vibration monitoring, oil analysis, and other forms of data analysis to track asset health.
Condition-Based Monitoring Challenges
Before partnering with Seeq, RWE used a network of sensors to monitor temperatures near their gas turbines. However, the sensors often misread the temperature and caused unnecessary shutdowns. When too many sensors failed, the system tripped and caused an extended outage.
The maintenance team also struggled to manage the workload associated with condition monitoring: in particular, collecting and interpreting data was taking too much time and labor. They needed an automated solution to speed up their workflows.
Finally, the preventive maintenance program needed to be overhauled. The team was struggling to determine exactly when to inspect its brush gear. They needed an accurate way to calculate operational hours and track when the last inspection had been carried out.
In sum, RWE needed a comprehensive analytics platform to streamline its condition maintenance program and reduce inefficiencies.
The right analytics platform would need to:
- Store and analyze condition monitoring data, easing the workload on RWE’s maintenance team and freeing them up for more complex tasks.
- Extract actionable insights to drive a predictive maintenance program, based on the use of condition monitoring data to perform maintenance when it’s actually needed.
- Monitor and analyze asset performance so that crews have time to prevent equipment failure.
- Drive effective root cause analysis to understand and prevent the root causes of common asset malfunctions.
- Plan preventive maintenance tasks like routine inspections to maximize efficiency and make the best possible use of equipment downtime.
How Seeq Unlocks Value Across RWE Generation
Seeq helped RWE detect sensor problems early on, preventing the widespread sensor issues that were taking gas turbines offline.
RWE uses GT 26 gas turbines, each with 24 thermocouples (thermal sensors) positioned in a ring to measure the temperature of the turbine’s exhaust gases. The sensors were experiencing frequent drift and returning inconsistent temperature readings, which led to frequent stoppages and shutdowns.
Correcting Sensor Failure with Seeq
RWE worked with Seeq to create a customized monitoring model to perform predictive maintenance on its thermocouple sensors. The model generates four statistical KPIs for each thermocouple, calculated over the four most recent startups. The KPIs are designed to track sensor inconsistencies. The goal is to identify those inconsistencies early, before they trip the system and cause expensive downtime.
Seeq Workbench monitors each KPI and sends out either amber alerts or critical red alerts if they exceed a predetermined threshold. The platform also identifies the specific sensors that are returning bad readings, labeling those thermocouples “BDQ” or “bad data quality.”
Failure Mode Analysis with Seeq
RWE uses a brush gear system in its power generation process. The brush gear needs regular inspections to ensure that it is functioning correctly. As the brush gear system deteriorates, the system can overheat and lose power, forcing the machine offline. It’s vital to inspect the system often enough to catch the signs of deterioration, but not so often that it interferes with regular production.
RWE’s previous analysis indicated that the brush gear should be inspected around every 400 hours of operation. The maintenance team was struggling to track exactly when those 400 hours of operation were completed, since the brush gear does not operate on a fixed schedule but is used in response to demand. That’s where Seeq’s predictive modeling is invaluable.
Automating Proactive Maintenance with Seeq
Seeq’s flexibility and its ability to connect to multiple data sources were critical for automating the inspection scheduling workflow and preventing brush gear failure.
Seeq ingests data from RWE’s inspection records through a maintenance app. The Seeq platform also collects process data from the plant’s AVEVA PI system to calculate the brush gear’s actual operational hours. Finally, Seeq uses RWE’s commercial data to forecast demand for the brush gear system and predict its operating hours over the next few days.
Put together, this data provides a clear analysis of:
- How recently the brush gear was inspected
- Its operating hours since inspection
- Its likely use in coming days
Based on all of this data, Seeq predicts when the brush gear will need to be inspected to function correctly and prevent deterioration. The result is that RWE performs inspections when they are actually needed, not according to an arbitrary calendar date. It’s the difference between periodic maintenance and data-driven, optimized preventive maintenance.
Using Fleet Dashboards to Monitor Asset Health
RWE’s team developed a customized MARA dashboarding tool that brings together all available data collections and integrates them with Seeq’s predictive modeling capabilities. The team uses the dashboard for all of its short- and long-term maintenance planning.
The dashboard includes machine data, reliability, statistics, commercial data, criticality assessments, and maintenance data. The predictive modeling section uses the collected data to create targeted failure mode models and maintenance analysis.
RWE’s end users and engineers interact with these dashboards on a regular basis, allowing them to make business decisions more efficiently and optimize the value of each maintenance activity.
Integrating Vibration Data
RWE intends to implement vibration-based maintenance more consistently and use it as a key element in its predictive maintenance program. Vibration monitoring is a widely-used means of increasing asset reliability by predicting machine faults and deterioration.
To that end, the team works closely with Seeq to continue developing new predictive models that integrate vibration-based maintenance across the company’s power stations. At the same time, RWE is collaborating with Seeq to maintain their existing predictive models and monitor existing failure modes.
Using Seeq, RWE assigns health scores to each asset and creates accurate predictions of maintenance needs and potential issues. Predictive maintenance typically reduces maintenance costs while increasing asset uptime.
Insights from Predictive Analytics in RWE Operation
RWE’s partnership with Seeq drives more efficient condition monitoring workflows and enables a highly effective predictive maintenance program, reducing the overall demand for reactive maintenance and increasing reliability.
The RWE plants have successfully implemented predictive maintenance models to keep their equipment and sensors up and running, greatly reducing forced shutdowns and stoppages.
Fleet dashboards now integrate predictive modeling using data from PI/DCS, MARA/Power BI, and forecast generation. Thanks to these integrations, RWE’s engineers and technicians now have daily access to asset health data and predictions about upcoming maintenance needs.
Seeq has enabled accurate predictions, and improved KPI tracking and inference making.
The results speak for themselves. As RWE Process Engineer Baris Mutlu puts it, “We use Seeq software…to visualize thermocouple data in a dedicated monitoring model, which enables us to optimize our maintenance strategy, which in the end, boosts our asset reliability.”
Empowering RWE Teams with Seeq
Shared dashboards and automated reporting enable intuitive, self-service analytics for RWE’s team of engineers. Because the dashboard brings together data and predictions in one central location, it also accelerates decision-making. The result is faster root-cause analysis and a steep reduction in manual data preparation.
Transforming Performance with Seeq
Through its ongoing collaboration with Seeq, RWE has improved its uptime, eliminating most of the unplanned stoppages that were plaguing its gas-driven power generation plants.
The team has greatly increased visibility into data and is actively using data-driven insights to plan its upcoming maintenance activities. The result is greater asset reliability and efficiency throughout the organization. Learn more about the Seeq and RWE partnership here.
Find out how Seeq can turn your data into a key tool to support your business’s health and profitability.