Thirty years before artificial intelligence (AI) became part of common public lexicon, the powerplant industry was getting steeped in it. There was always the question, however, how far you could “push” the technology and reduce staff expertise. Now, a relatively new entrant to the M&D-as-a-service space emphasizes how its operations experts remain in the loop.
A little history: Predictive analytics technology began to be applied in powerplants in the mid-to-late 1990s when the earliest software packages came on the market. Typically, these packages pulled data from the plant historian, then applied pattern recognition algorithms to detect parameters ‘out of range.”
As the technology improved, coincident with the move by large owner/operators (O/O) to remote and centralized monitoring and diagnostics (M&D) centers in the late 2000s, software with advanced pattern recognition (APR) algorithms became a standard feature in these centers, along with thermal performance monitoring, vibration monitoring, and other functions. Experts at the M&D centers usually assisted the folks at the plant in responding to anomalies.
Consolidated Asset Management Services (CAMS), a large powerplant operator, offers its asset performance management (APM) APR as a service with a different idea. The company emphasizes the operations expertise, along with the technology. While they license their APR technology from Prometheus, CAMS APM is happy to work with whatever APR technology is available at their clients’ plants.
According to David Saad, a veteran in the APR space, with years under his belt at one of the pioneer APR software providers, most of the available software packages work in similar ways today. What distinguishes a service provider is how effective they are in helping the plant respond to the alerts and anomalies.
These days, less and less actual engineering expertise resides at CCGT plants as staffing levels shrink. More expertise is outsourced, whether to third parties or engineering departments within the O/O organization. Relatively young, inexperienced engineers are replacing veterans who retire.
Says Saad: “Other firms providing these services deploy less experienced engineers to monitor the APR screens, and when they find deviations, they contact a subject matter expert (SME) to confirm what the issue is, even if it is as simple as a bad signal. Then other departments are alerted, such as maintenance and monitoring, to verify whether bad data was used to train the model. In fact, the information is bouncing between too many folks in the loop, gumming up the works.”
CAMS philosophy is that one SME is assigned to respond to and resolve issues, indicated by the software, for the client. “We call it “engineer-in-a-box,” notes Saad, a counterpoint to those who sell “software in a box.” “We are APR software agnostic,” adds Ken Earl, VP in charge of the APM team, “the experience we bring to the table as a reliable partner, as operators of 65,000 MW of power facilities, is key.”
“In other words,” Saad explains, “we don’t just dispatch a daily or weekly report to the plant.”
An example of when expertise comes in handy is knowing which subsystems need monitoring. APR typically monitors the steady state data from operations, but it is also vital to keep an eye on systems, like liquid fuel, which are “still live” even if the plant isn’t burning it. CAMS APM experts aren’t limited to the main gas turbine, HRSG, and steam turbine islands either. For example, they monitor data from dissolved gas analyzers (DGA) in the GSU transformers ranging from simple (one or two gases) to more complex DGAs looking for up to 10 gases.
CAMS experts also help their clients set up the APR monitoring system, clean up the tag list in the data historian, enhance the APR system already installed, and even suggest adding instrumentation which will make the APM more effective. Many plants have APR software, but staff do not have the time to get the most out of it. “Today, most CCGT plant O&M staff are not plant-engineering focused, because they are relegated to project management and reacting to day-to-day issues.” Earl stresses. CAMS has also enhanced the basic software they’ve licensed from Prometheus to improve modeling and visualization.
Customer feedback underscores value
CAMS APM has delivered significant value to our facility by enabling early detection of parameters operating outside historical trends, says Anner Reyes, Plant Manager, Parkway Sewaren 7 Generation Center. This has helped us prevent forced outages. The proactive approach gives plant staff ample time to source materials, allocate resources, and plan repairs during scheduled shutdowns, rather than reacting to emergencies.
Unlike traditional OEM monitoring, CAMS APM customizes its service to our station by incorporating input from plant personnel and building models around real-time concerns. Their real-time support, combined with historical data analysis and SME expertise, has been instrumental in troubleshooting complex issues, even challenging OEMs to investigate combustion problems that could have caused major damage. CAMS APM goes above and beyond by offering a tailored, collaborative approach that transforms monitoring into a powerful tool for reliability and operational excellence.
Examples of how CAM APR works with their clients follow.
Gain MW, prevent FOD. Summer of 2024, CAMS APM was investigating possible performance loss on a CCGT block triggered by the APR software. Utilizing pin charts (which quickly separate groups of data by color), a gradual shift was noted, from the previous year, of steam turbine inlet pressure vs steam turbine first stage pressure (Fig 1).
The site was able to schedule an outage, before capacity testing, during which heavy debris was found in the steam turbine main strainer, along with a sheared pin from one of the main-steam stop valves. The planned outage repairs allowed the site to recover 10 MW, which allowed for $700,000 additional revenue. Also, the timely outage prevented the valve pin from releasing downstream and causing foreign object damage (FOD) to the high-pressure turbine. A previous similar FOD incident resulted in damages costing $5-million.
Fix NOx, combustor issues. Last summer, CAMS APM noted an increase in engine exhaust NOx upstream of the catalyst at a site. Utilizing live radar plotting (Fig 2), experts identified hot and cold regions in the exhaust profile to help the site identify which combustion cans were possible culprits. These temperature variations, together with combustion dynamics changes, supported an in-depth analysis by the OEM to improve operability until the planned major inspection the following spring.
Meanwhile, the site took advantage of its fall planned shutdown to borescope the unit, subsequently discovering multiple issues with the unibody seal and one combustion nozzle. This compelled the site to extend the outage prior to the winter run season and address these issues. The work performed has returned GT exhaust NOx to appropriate levels, alleviating concerns with permit exceedances until the spring 2026 planned shutdown.
Forecast bearing trip limit. In May 2025 CAMS APM noted an increasing rate of change on the GT exhaust bearing vibration at a site. OEM reps were contacted, who suggested the behavior and values were normal. To ensure this situation was being adequately monitored, and to inform the customer well in advance of an impact on operations, CAMS APM added a forecast model (Fig 3). Five months later, the model indicated the situation was worsening. The site was notified that it had 13 days before reaching the trip limit on the exhaust bearing vibration.
The site again contacted the OEM reps. The OEM reiterated that the behavior and values were normal. However, as a precaution, the trip limit parameter was raised by the OEM. This prevented the unit from going into a runback or trip so the site could make it to the spring 2026 major inspection.
Caution on VAR. While many APM-as-a-service providers are quick to point out the huge dollar savings from one or two major “catches,” CAMS takes a different tact. “While these numbers can appear dramatic,” Earl stresses, “they are extremely site- and O/O-organization specific.” If a client wishes to use such value at risk (VAR) numbers internally or report them, we will work with them to do so, but they have to agree on the methodology used.”
Avoiding reactive response to mal-performance. The following example illustrates how the CAMS APM team goes beyond just monitoring. An alert was issued to a site on irregular vibration in an auxiliary cooling tower fan. Readings had been erratic for almost a year, spiking to 0.62 mils, almost double the historical trend. The anomaly had not affected the non-drive end of the assembly, indicating a potential gearbox issue. In response, and with CAMS recommendations, the plant initiated a work order to inspect the fan.
To mitigate further risk, the fan was designated for emergency use only to avoid further damage and potential failure until the gearbox could be replaced at the next outage. CAMS then refined the models around cooling tower fans with new data. By identifying which gearbox components should be scrapped and which could be repaired and retained, CAMS team helped the client reduce repair costs by up to 50%. CCJ
