While growth in renewable energy technologies increases grid capacity, both new and existing power plants must plan to manage performance under any circumstance. The focus is on improving the availability, reliability, and efficiency of the assets that power economic growth. Thus, power generators across the globe are looking for more efficient ways to meet customer needs during a time of great market volatility and rapidly changing demands.

EtaPRO from Toshiba ESS is a leading plant performance improvement platform that meets this need. The technology helps plant operators improve plant reliability and efficiency while reducing operational costs and unplanned maintenance. This solution offers rich features deep in capability on a proven integrated platform and supported expertise from real-world industry experience.

In the US, a mid-Atlantic power station has leveraged many technologies to improve the reliability, availability and capacity of its three 660MW coal-fired supercritical units. Having experience using equipment condition monitoring systems, they were interested in a single platform that could provide thermal performance and equipment condition monitoring with a broad distribution of critical information to operations, maintenance, engineering and corporate personnel.

They decided to evaluate a new system and chose EtaPRO APR, which selects a unique subset of the historical data that most closely represents the current equipment operating mode and builds a unique model based on that data subset. This localised modelling eliminates the influence of irrelevant historical records, provides improved accuracy with fewer false alarms and eliminates the need to develop different models to address each unique mode of operation.

During their evaluation period, EtaPRO APR was executed on ten critical pieces of equipment, which were also monitored by their legacy system to ensure that no critical plant issues might be missed. In addition, the models for both methods were developed using the same historical data sets.

Shortly after installation, EtaPRO APR initiated an alarm involving the generator hydrogen temperature coming from one of the gas coolers. The same data in the customer’s legacy system indicated that the predicted temperature was tracking reasonably closely with a modelled value. Therefore, no alarm was initiated by the legacy system. While a 32°C temperature will be experienced during start-up on a cold day, continued operation at that temperature would cause the generator hydrogen seals to shrink and potentially become brittle, ultimately resulting in possible gas leakage and a unit shutdown for repair.

Plant staff investigated the alarm and found a control system fault keeping a hydrogen cooler bypass valve from fully closing. Having the predicted value track a measured value through an equipment or instrument fault is fairly common with many statistical systems. However, it occurs when a single data value overly influences its prediction [auto correlation] and possibly the predicted values of many other signals. This results in missing critical equipment issues and/or initiates false alarms on other plant data.

A technologically advanced Midwestern power utility in the US was committed to reducing emissions and improving efficiencies across their fleet. They began evaluating additional carbon reduction efforts on their existing generating fleet through efficiency improvement projects.

EtaPRO and VirtualPlant performance technologies were installed to help quantify and track the resulting performance gains. Now their coal-fired generating station is on track for a heat rate record of 9,900 Btu/kWhr, and they did it with technology and a team effort by plant staff, corporate engineering and EtaPRO.

EtaPRO provides plant personnel with feedback to effectively monitor incremental improvements in plant operation. Plant operators and engineers can see lower heat rates on the screen with every improvement. The customer is also validating the results by running EtaPRO reports and comparing them to a separate report created by the fuels department. When the heat rate results concur on both reports, this confirms that the improvements are paying off.

Based on early success, the company is accelerating the implementation of EtaPRO to its remaining fleet of six coal and combined cycle plants. The generation utility evaluated several technologies before applying the EtaPRO Performance & Condition Monitoring System across their generating fleet. They chose EtaPRO because of its fully integrated VirtualPlant models of the boiler and turbine cycles, the feature-rich online monitoring software and EtaPRO extensive experience deploying fleet-wide monitoring solutions.

Plant owners and operators want to know when their plants are at risk due to incipient failure, how to avoid the failure and how long they have to take action before it becomes critical. For example, EtaPRO Predictor, with AutoDiagnosis technology (US Patent 7,089,154 B2), provides unprecedented early detection and diagnosis of faults in gas and steam turbines, generators, pumps, fans, compressors and other critical rotating machinery. However, unlike OEM protection systems, EtaPRO Predictor predicts time-to-criticality and provides recommendations for action to avoid unplanned downtime.

Traditional vibration monitoring systems provide diagnostic information only after a high vibration incident. This means that very few faults are monitored, warnings to the operator come too late to avoid failure, and post-mortem analysis is required to understand what happened. In contrast, EtaPRO Predictor continuously and automatically evaluates multiple fault symptoms in their early stages and provides a warning of impending failure so corrective action can be taken.

How does the EtaPRO Predictor work? Vibration signals carry vast amounts of diagnostic information about machine component problems in their early development. This diagnostic information is revealed by EtaPRO’s proprietary algorithms, executing advanced time and spectral analyses.

EtaPRO Predictor turns this information, together with process and geometrical component data, into AutoDiagnosis messages. An instantaneous AutoDiagnosis reveals events such as rubbing, surge and cavitation. In contrast, a predictive AutoDiagnosis shows specific machinery faults in their early development and provides a time-to-criticality prognosis for recommended action. ESI

Now read this from Toshiba ESS Toshiba Africa: Creating innovative solutions in response to energy issues

© All content copyright 1997 - 2021 Vuka Group, unless specified otherwise.