Centro Energia Teverola started operations at its 150-MW cogeneration combined cycle power plant in 1998. The plant in Teverola, Italy, has two gas turbines, two heat recovery steam generators and a steam turbine.
View related content on PlantServices.com
Although the plant ran smoothly for the most part, filters on the gas turbines underwent unexpected performance degradation from clogging. These filters clean the inlet air to the turbine’s compressor section.
The inlet air must be as clean as possible, particularly as such large volumes of air are passed through the filters, to prevent compressor blade fouling, which has a large effect on performance. Impurities such as dust otherwise build up quickly in the compressor and degrade the polytropic efficiency of air compression.
“Filter clogging often resulted in emergency downtime to install replacement filters, which was especially costly during peak production periods,” says Vincenzo Piscitelli, general manager at Centro Energia. “When the filters clogged, the plant almost always was forced to shut down a turbine because of the serious performance degradation caused by high-pressure drop across the filters.”
Centro Energia and Emerson Process Management had been collaborating on plant performance analysis for several years. Initially, Centro Energia wanted to understand what benefits Emerson could provide to improve plant operations. When the filter problem became significant, Centro Energia engineers visited an Emerson client in England, which already had successfully adopted the AMS Performance Monitor. This trip eliminated doubts about the benefits that could be obtained.
To solve the filter problem and to improve overall performance, Centro Energia installed an AMS Performance Monitor system remote analysis of performance data on two gas turbines, two heat recovery steam generators (HRSGs) and one steam turbine. “After analyzing the data remotely, Emerson could inform operators when problems existed,” explains Piscitelli. “Remote analysis would allow us to schedule filter replacements at a more convenient time, such as during low-demand periods.”
The AMS Performance Monitor provided a rapid return on investment. A single filter change, scheduled proactively instead of reactively, paid for the leased AMS service for two years.
Remote monitoring and analysis
AMS Performance Monitor takes data from the Centro Energia historian, so it’s completely independent of the plant’s control system — a legacy Bailey Infi-90 control system — and its instrumentation. The AMS Performance Monitor typically uses as inputs the hourly averaged values from the historian.
For the gas turbines, AMS monitors several variables, not just filter pressure drop, to provide a range of benefits.
- Inlet temperature — This is needed to correct the analysis to normalized performance, so comparisons are based on known conditions.
- Inlet pressure — This also is used to correct to normalized conditions.
- Shaft power — This is a measurement of the gas turbine’s power output from the associated generator.
- Shaft speed — This generally is assumed to be constant for electricity-generating turbines.
- Fuel flow — The system measures fuel flow to the machine in energy units. If mass or volume units are preferred, assumptions about fuel properties can be specified.
- Additional measurements — Air flow, compressor temperatures and other variables provide for more in depth and comprehensive analysis.
The historian provides similar data for the two heat recovery steam generators and the steam turbine.
“The Emerson data collection tool formats the historian data into Excel files, which are then transmitted by e-mail to Emerson’s Performance Centre in Teesside, England,” explains Piscitelli. “For security reasons, no explanatory information is transmitted with the values. Instead, the data is stripped of its units of measurement. The team in the UK knows the proper data format, so it can perform the analysis unhampered.”
Monitoring gas turbine variables by extracting data from the plant’s data historian allowed remote comparison of live data to a computer model to determine when filters needed service.
On receipt in Teesside, the data undergoes a thorough analysis and cleaning, which involves automatic and manual steps. Initially, data is inspected manually for obvious corruption, blank data and other problems. The data is then put through an automated SPC-type analysis to identify missing and suspect data points (outliers or data marked as bad). If necessary, agreed values, such as last good or mean values, can be substituted for missing values, or data can be re-collected.