Generation assets in competitive markets must perform at a high level year in and year out to meet their contractual obligations. Failure to embrace a philosophy of “continual improvement” in plant operations means it’s only a matter of time before the competition will eat your lunch. The focal point of continual-improvement efforts typically is the gas turbine in the industry segment served by CCJ. It’s certainly a priority, but failure to pay close attention to support systems and equipment can cause problems you’re not prepared for and don’t want.
As plants get beyond 15 years of age, things that you took for granted begin reminding you of their presence. An analogy might be your home refrigerator, TV, washing machine, and/or hot-water heater. Problems with support systems and equipment will keep you up at night because there generally are no easy answers.
What you hear more and more at user-group meetings is that the plant person who knew all about the system of current concern has retired, or the manufacturer no longer makes the model you have (or, worse, has gone out of business), or you can’t find spare parts, etc. It’s important to conduct system reviews every couple of years, at least, and budget for upgrades and replacements before your plant’s balance sheet suffers.
What follows is a case history on the replacement of an obsolete vibration monitoring system which illustrates how quickly advancements in I&C technology, monitoring, and diagnostics can impact your plant’s risk profile. CCJ ONsite recently published a similar article concerning LCI starting systems for gas turbines, and fogging and wet compression.
The original machine protection system (MPS) for Ferndale Generating Station’s gas and steam turbine/generators became obsolete in 2014, the year the plant turned 20 years old. The legacy supplier no longer supported the MPS and owner Puget Sound Energy (PSE) planned to replace it and implement a cost-effective machine condition-monitoring strategy (Fig 1). This strategy involved correlating other process data—including bearing temperature, oil temperature, and active/reactive power—with vibration signals to increase the reliability of the diagnostics.
The original MPS did not have online condition-monitoring capability, and ad-hoc diagnostic services were performed on the generating units through the legacy MPS supplier. A specialist traveled to Ferndale with a portable data analyzer to diagnose machine vibration whenever in-depth analysis was required.
Onsite visits, plus follow-on processing of event data, are costly and time-consuming. Installing a new condition-monitoring system from the legacy MPS supplier, currently owned by the gas-turbine OEM, would have been expensive and required maintenance and IT infrastructure. Limited IT infrastructure resources and no onsite machinery diagnostic expertise at Ferndale encouraged PSE to consider other alternatives.
The utility needed a proven and reliable MPS and online condition-monitoring capability with no changes to the existing DCS interface. PSE also chose to replace existing transducers with upgraded ones while retaining the original transducer mountings, reuse the field wiring, and fit the new MPS in the same panel cutout as the legacy system.
PSE investigated the market for machine condition-monitoring systems that provided the desired capabilities yet were cost effective to replace its legacy MPS. One particularly-relevant solution investigated was Brüel & Kjær Vibro’s VC-8000 Setpoint MPS/Condition Monitoring Software (CMS). It offered modern protection and reliable advanced condition-monitoring capabilities without the need for a dedicated CMS server.
The CMS is based on the OSIsoft PI data historian used at Ferndale. VC-8000 eliminated the need for separate, special vibration databases. PI’s user interface visualizes all vibration data including current values, alarm status, and trends.
Specialized measurements and plots (for example, time waveforms, orbits, shaft centerline, waterfall, cascade spectrum, etc) are provided via the CMS’s visualization tools that augment PI’s native visualization capabilities. Other third-party products and services—such as statistical data analysis, automatic decision support, and thermodynamic performance monitoring of the generating units—can be integrated as well.
Another benefit of the VC-8000 MPS is its capability to continuously back up data on the local Secure Digital (SD) card. If PSE’s network went down, no data would be lost. Data can be manually retrieved, used, and retroactively stored in PI. Monitored data can be remotely uploaded by File Transfer Protocol (FTP) or emailed to an analyst at any time without a remote connection to the database—benefitting the utility with the capability to quickly make decisions, remove firewall navigation efforts, and reduce costly site visits.
The SD function proved immediately useful. The PI connection to the MPS was delayed, but no data were lost because of the connectivity interruption.
PSE benefits from correlating process and vibration data—improving the reliability of fault analysis and improving root-cause analysis (RCA) results. Coincidentally, the first diagnostic evaluation performed using VC-8000 occurred during a training session when a machine fault was suddenly detected: An imbalance occurred on Bearing 3 of one gas turbine. Personnel were notified of the issue immediately, enabling an expeditious assessment (Fig 2).
The bottom line: The VC-8000 provided comprehensive condition-monitoring capabilities at a lower cost than what would be achieved by upgrading the system using the legacy MPS supplier.