We recently talked with Mike Harris, Generation Performance Manager at CPS Energy, to gain 5 insights in 5 minutes about his role, the value of data analytics and evolution of power plant operations. Here’s what he had to say.
Describe CPS Energy. What do your customers value most about what you do for them? What is your title there, and what are your primary responsibilities?
CPS Energy is the local power and natural gas company. We are municipally owned, since 1942 when the City of San Antonio purchased the power company outright from the then-local electric company and the then-existing local gas company. So much for history…we have been evolving from a 100% natural-gas powered company in the 1970s, to adding coal power in the late 1970s, to adding nuclear in the 1990s, to building combined cycles in the 2000s. Then we added some peaking gas turbines also in the 2000-2004 timeframe and now we have a lot of PPAs for wind and solar.
We’re doing Purchase Power Agreements (PPAs) because (as a municipality) we are a non-tax-paying entity. So the only way to take advantage of tax credits is to have a third party company do the wind and solar buildout.
My title is 'Generation Performance Manager', and that puts me squarely in charge of heat rate information, heat rate improvement, etc.
You mentioned that you’ve taken a larger role in the budgeting process, deciding which budget items get funded and which ones don’t. Can you tell me more about that?
"I’m in the Reliability Department. Reliability is just another way to say 'risk mitigation'. We have a certain number of dollars every year to spend on plant maintenance. Where should we spend it? Obviously, there are some things that are imperative to keep running our units. But there are many projects that are 'elective'. These projects we rank in terms of 'likelihood of failure' and 'consequence of failure'. So projects that mitigate higher risks (both high likelihood and high consequence) rank highest. We have been using this tool to evaluate elective projects for several years now."
How has the art, the practice, of power plant testing changed over the past 10-15 years?
Power plant testing has gone through big changes. Twenty years ago, we did the whole major setup with calibrated instruments placed all over the plant, taking the data using a datalogger, then
using a computer to analyze and produce a formal report. Then things got easier – smart transmitters, better dataloggers, a move from FORTRAN to Excel. Then things got worse – our group was whittled down from 5 to 4 to 3 to 2…to ONE. Now the only 'testing' that happens is using PI data.
What issues did you catch with the help of data analytics capabilities, which might otherwise have gone undetected until damage or safety incidents occur?
Data analytics has had several nice 'catches'. I’ll draw attention to two big ones.
One, we had a strange control logic going on with Spruce1 (575-MW coal fired unit) superheat and reheat control logic. Long story short, we were CONTROLLING a temperature based on an EXPECTED temperature curve. Put another way, we were forcing the temperature down when the unit was at low load. This cost us a LOT of money over the years since the control logic was original with the unit! We would have never realized that that was happening without Atonix Digital.
Two, the Black & Veatch M&D center caught a weird increase in temperature for our Spruce1 condenser during the wintertime. The temperature wasn’t unusually high so we would NOT have noticed until lake temperature rose later in the year. But M&D center noticed that it was much higher than EXPECTED. Sure enough, on investigation we discovered that one of our circulating water pumps had snapped its driveshaft. Weirdly, it wasn’t vibrating or alerting us in any way that there was a problem! But, if we had continued operating as is for a couple of months, summer would have arrived, and the unit would have been load limited! That would have been a PROBLEM. As it was, M&D alerted us that there was a problem. We found the problem, sent the pump off to get maintenance while it was still cool, and got the pump back before summer arrived.
What, in your experience, is the proper role of data analytics in power plant operations? Does this role differ when you’re working with coal-fired units, gas turbines, or renewables?
The role of data analytics – we’re essentially using Black & Veatch’s M&D center like a plant operations engineer. We are fairly short-staffed nowadays, so plant engineers have too many responsibilities to focus on finding and understanding plant performance. The M&D center is doing this and it’s fulfilling a need that would otherwise go undone. We’re using Atonix software with two coal-fired, two combined-cycle gas, and eight peaking gas turbines. Our renewable fleet is all PPAs, so we’re not monitoring any of our renewables.