Our gas power plant (GE Frame 9E, Mark VI) still uses manual turbine log sheets. I proposed automatic data logging for root cause analysis, but the Maintenance Manager said it would make operators less proactive. Do the benefits outweigh this risk?
The Maintenance Manager is mostly correct. Most "operators" aren't properly trained to begin with and loathe and despise having to walk around and record readings on a log sheet. They don't bother to compare the readings from the last hour or 12 hours--they just write the numbers down and throw the clipboard on the desk. Unsurprisingly sometimes the numbers recorded are outside possibility--because they were written down in the wrong row or column. And, let's face it: The "younger generation" just doesn't like to get out of their swivel chair. Period. And because they aren't properly trained to spot discrepancies or changes in readings nothing gets noticed until something bad happens--then when someone looks at the log sheets it becomes clear that something had been "brewing" for a while, but no conscious operator caught it, and that means no accountability.
Walking around and recording readings helps to spot leaks--leaks of air, water, oil, hydraulic fluid, fuel, cooling water, compartment ventilation problems, leaky seals and pipe- and tube fittings. But simply writing down a number on a piece of paper doesn't do anything. It's the TREND that matters! Comparing the readings for the last hour or two hours or six hours or twelve hours, noting decreaes or increases--and SUDDEN increases or decreases--that can help avert problems and equipment damage, and lost revenue. I have worked at power plants where the operators get bonuses based on reliability and availability--and THOSE operators work VERY HARD to spot problem trends because those bonuses can be very large. Not only do they look at their round sheets/logs, they look at everyone else's round sheets/logs--because it can affect their bonus!!! They were NOT KIND to other operators who were slackers, quite often making their work experience very stressful and ultimately quitting. But, those operators had good training to begin with--and good OJT (On-The-Job) training from their colleagues. They were constantly quizzing each other--which made the time go faster at work; they weren't always watching the clock to see how long before their shift ended.
Accountability can take MANY forms and shapes--paying for high reliability and availability is one of them. But without some kind of accountability everyone (especially the "younger" generation) is not going to put much effort into recording--and analyzing--readings taken on rounds. The plant operators, maintenance, and controls/automation staff need to all work together as a team to solve problems which can OFTEN be found by proper scrutiny and analysis of round sheets/logs.
I've seen sites where they have small hand-held recorders to use to take readings; and the operators just throw them AT the desk/console when they come back inside. They lose charge; they get damaged (if it's broken then no one has to take rounds/readings, right.?.?.?) The devices were always abused and in need of repair.
Automation doesn't solve every problem. It might, but it's going to take a LOT of programming and AI smarts to make that move toward a reliable means of operating and troubleshooting.
And then, who will need operators?
And, what is the risk in making rounds? As long as the "operator" is awake and properly aware of his surrounding, there should be no risk. If risks are identified, the Safety Officer/team should be involved in mitigating those risks. But, there's no substitute for being awake, aware and conscious. Just because entering a space that might be hot or have vapors (which can/should be mitigated!) doesn't make it dangerous. If a conscious/aware "operator" identifies a liquid on the floor when entering a compartment it should be easy enough to back out of the compartment and report the liquid to the responsible person/department.
Now what say you? You identified yourself as an engineer. At a power plant. My colleagues and I would say to you, "Quit-chur bitchin' and do the job you were hired to do."
We now return to our regularly scheduled programming (watching the clock to see when the shift ends for you).
If the log sheets you are talking about are simply recording values from the HMI display, well, then, there's still a LOT to be learned from watching the TREND of readings over time. The Mark VI HMI does have the ability to record data, at low speed, but it's not fully automatic and the data it records must be viewed on an an HMI or Engineering Workstation with Toolbox/ToolboxST running on it. NEITHER of these applications/tools (Toolbox/ToolboxST) is difficult to learn--but for some reason people are terrified of learning how to use it. THE TURBINE CAN'T BE TRIPPED WHEN REVIEWING/ANALYZING A TREND RECORDING. FULL STOP. PERIOD. The turbine MIGHT trip when someone was reviewing or analyzing a trend recording--but IT WON'T TRIP BECAUSE THE PERSON USING TOOLBOX/TOOLBOXST DID SOMETHING "WRONG"!!!
Operating a GE-design heavy duty gas turbine is mostly boring--especially when things are running at steady-state and are relatively stable and normal. It's the start-ups and shutdowns that really need to be understood--everything that happens and when it happens and what it looks like on a trend recording. Knowing how to determine if a Process Alarm is important, and when it's not, takes experience and teamwork--but it can and has happened in many control rooms around the world. But, it's not learned from some book or training course--it takes learning to read and understand P&IDs and learning what the alarm text messages mean. For example, L.O. TEMPERATURE HIGH: when will that alarm sound, and what happens if the L.O. temperature continues to rise, especially if it's rising fairly quickly. Will the machine eventually trip? Yes. The alarm occurs when the L.O. bearing header temperature reaches 165 deg F, and the machine will trip if the L.O. Bearing header temperature increases above 175 deg F. The normal L.O. Bearing Header temperature is 125-130 deg F, and if conscious operators were paying proper attention to the L.O. Bearing Header temperature it would never reach 165 deg. What to do when the L.O. Bearing Header temperature reaches 165 deg F? Well, someone needs to check the cooling water system, the L.O. Bearing Header Temperature regulating valve, and someone should probably initiate the process of reducing the load on the machine to try to reduce the temperature of the L.O. Bearing Header temperature (not that lowering load will work if the oil temperature is climbing quickly--it's probably too late at that point.
That's the kind of training operators should be getting from their colleagues. What each and every alarm means and what to do when it is annunciated. That's NOT in any book GE ever produced--even the vaunted GE Belfort. That kind of knowledge keeps the plant running--reliable and available. Which most operators don't even understand that THAT is their job: Keeping the machines running. Knowing what to do when the turbine control detects something wrong--or doesn't. Conscious and aware operators, with good training and support can do this. And should do.
The Maintenance Manager is mostly correct. Most "operators" aren't properly trained to begin with and loathe and despise having to walk around and record readings on a log sheet. They don't bother to compare the readings from the last hour or 12 hours--they just write the numbers down and throw the clipboard on the desk. Unsurprisingly sometimes the numbers recorded are outside possibility--because they were written down in the wrong row or column. And, let's face it: The "younger generation" just doesn't like to get out of their swivel chair. Period. And because they aren't properly trained to spot discrepancies or changes in readings nothing gets noticed until something bad happens--then when someone looks at the log sheets it becomes clear that something had been "brewing" for a while, but no conscious operator caught it, and that means no accountability.
Walking around and recording readings helps to spot leaks--leaks of air, water, oil, hydraulic fluid, fuel, cooling water, compartment ventilation problems, leaky seals and pipe- and tube fittings. But simply writing down a number on a piece of paper doesn't do anything. It's the TREND that matters! Comparing the readings for the last hour or two hours or six hours or twelve hours, noting decreaes or increases--and SUDDEN increases or decreases--that can help avert problems and equipment damage, and lost revenue. I have worked at power plants where the operators get bonuses based on reliability and availability--and THOSE operators work VERY HARD to spot problem trends because those bonuses can be very large. Not only do they look at their round sheets/logs, they look at everyone else's round sheets/logs--because it can affect their bonus!!! They were NOT KIND to other operators who were slackers, quite often making their work experience very stressful and ultimately quitting. But, those operators had good training to begin with--and good OJT (On-The-Job) training from their colleagues. They were constantly quizzing each other--which made the time go faster at work; they weren't always watching the clock to see how long before their shift ended.
Accountability can take MANY forms and shapes--paying for high reliability and availability is one of them. But without some kind of accountability everyone (especially the "younger" generation) is not going to put much effort into recording--and analyzing--readings taken on rounds. The plant operators, maintenance, and controls/automation staff need to all work together as a team to solve problems which can OFTEN be found by proper scrutiny and analysis of round sheets/logs.
I've seen sites where they have small hand-held recorders to use to take readings; and the operators just throw them AT the desk/console when they come back inside. They lose charge; they get damaged (if it's broken then no one has to take rounds/readings, right.?.?.?) The devices were always abused and in need of repair.
Automation doesn't solve every problem. It might, but it's going to take a LOT of programming and AI smarts to make that move toward a reliable means of operating and troubleshooting.
And then, who will need operators?
And, what is the risk in making rounds? As long as the "operator" is awake and properly aware of his surrounding, there should be no risk. If risks are identified, the Safety Officer/team should be involved in mitigating those risks. But, there's no substitute for being awake, aware and conscious. Just because entering a space that might be hot or have vapors (which can/should be mitigated!) doesn't make it dangerous. If a conscious/aware "operator" identifies a liquid on the floor when entering a compartment it should be easy enough to back out of the compartment and report the liquid to the responsible person/department.
Now what say you? You identified yourself as an engineer. At a power plant. My colleagues and I would say to you, "Quit-chur bitchin' and do the job you were hired to do."
We now return to our regularly scheduled programming (watching the clock to see when the shift ends for you).
Thank you for taking the time to write such a detailed response. I agree with many of your points, particularly that there is no substitute for field rounds. Operators should continue inspecting for leaks, unusual noises, odors, vibrations, and other conditions that no control system can detect.
However, I think there may be a misunderstanding of my proposal. I wasn't suggesting replacing operator rounds or inspections. My proposal was specifically for the Performance Engineering Department to automatically retrieve process data already available in the GE Mark VI system for performance monitoring, trend analysis, and root cause investigations.
I also agree that walking the plant is valuable because it forces operators to observe equipment. But those observations don't depend on manually copying process values onto a log sheet. Operators can still perform the same rounds, inspect the same equipment, and report abnormalities while the process data is logged automatically in the background.
Continuous automatic logging provides far higher-resolution historical data than hourly manual log sheets and is invaluable for diagnosing issues such as compressor fouling, heat rate degradation, exhaust temperature trends, and other performance problems. That level of detail is difficult to achieve with manual records alone.
So, from my perspective, the two approaches are complementary rather than competing. Operators continue to own plant condition through field inspections and situational awareness, while the Performance Engineering Department benefits from accurate, continuous historical data for analysis. My proposal was never to replace operators, only to improve the quality and availability of engineering data.
If the log sheets you are talking about are simply recording values from the HMI display, well, then, there's still a LOT to be learned from watching the TREND of readings over time. The Mark VI HMI does have the ability to record data, at low speed, but it's not fully automatic and the data it records must be viewed on an an HMI or Engineering Workstation with Toolbox/ToolboxST running on it. NEITHER of these applications/tools (Toolbox/ToolboxST) is difficult to learn--but for some reason people are terrified of learning how to use it. THE TURBINE CAN'T BE TRIPPED WHEN REVIEWING/ANALYZING A TREND RECORDING. FULL STOP. PERIOD. The turbine MIGHT trip when someone was reviewing or analyzing a trend recording--but IT WON'T TRIP BECAUSE THE PERSON USING TOOLBOX/TOOLBOXST DID SOMETHING "WRONG"!!!
Operating a GE-design heavy duty gas turbine is mostly boring--especially when things are running at steady-state and are relatively stable and normal. It's the start-ups and shutdowns that really need to be understood--everything that happens and when it happens and what it looks like on a trend recording. Knowing how to determine if a Process Alarm is important, and when it's not, takes experience and teamwork--but it can and has happened in many control rooms around the world. But, it's not learned from some book or training course--it takes learning to read and understand P&IDs and learning what the alarm text messages mean. For example, L.O. TEMPERATURE HIGH: when will that alarm sound, and what happens if the L.O. temperature continues to rise, especially if it's rising fairly quickly. Will the machine eventually trip? Yes. The alarm occurs when the L.O. bearing header temperature reaches 165 deg F, and the machine will trip if the L.O. Bearing header temperature increases above 175 deg F. The normal L.O. Bearing Header temperature is 125-130 deg F, and if conscious operators were paying proper attention to the L.O. Bearing Header temperature it would never reach 165 deg. What to do when the L.O. Bearing Header temperature reaches 165 deg F? Well, someone needs to check the cooling water system, the L.O. Bearing Header Temperature regulating valve, and someone should probably initiate the process of reducing the load on the machine to try to reduce the temperature of the L.O. Bearing Header temperature (not that lowering load will work if the oil temperature is climbing quickly--it's probably too late at that point.
That's the kind of training operators should be getting from their colleagues. What each and every alarm means and what to do when it is annunciated. That's NOT in any book GE ever produced--even the vaunted GE Belfort. That kind of knowledge keeps the plant running--reliable and available. Which most operators don't even understand that THAT is their job: Keeping the machines running. Knowing what to do when the turbine control detects something wrong--or doesn't. Conscious and aware operators, with good training and support can do this. And should do.
Thank you for your insights. I agree that understanding trends and knowing how to interpret alarms are essential operator skills, and I also agree that reviewing trends in Toolbox is a safe and valuable activity.
However, my proposal wasn't about replacing Toolbox or reducing operator responsibility. It was about making automatically logged process data readily available to the Performance Engineering Department for long-term trend analysis and root cause investigations, instead of relying on manually transcribed values from log sheets.
I also fully agree that operators should continue making field rounds. Walking the plant is how leaks, unusual noises, odors, vibrations, and other abnormal conditions are identified. Those observations remain essential regardless of whether process values are written on paper or logged automatically.
My question is this: If the GE Mark VI is already capable of recording process data, what additional value is gained by manually transcribing those same HMI values onto log sheets for engineering analysis? It seems to me that operators can remain fully engaged with the plant while performance engineers benefit from continuous, higher-resolution historical data.
Interesting setup! The GE Frame 9E with a Mark VI control system has a strong track record in gas power generation. I'd be interested to hear about your operating experience, maintenance practices, or any upgrades you've implemented to improve reliability and performance.
You are not explaining your proposal very well--at least not here on this forum.
If "maintaining" operator engagement with the plant (which is pretty much NOT a real thing) there are many ways to do so. A few were suggested above, but making operators manually log parameter values on a log sheet and then making the log sheets available to the Performance Engineering Department (the existence of which wasn't made known in the original post) isn't really going to accomplish much other than making work for two groups of people.
YES the Mark* VI is more than capable of providing data automatically using different means (MODBUS--though that doesn't generally include time-stamp information), and at least one other protocol which escapes my memory at this moment (which does include time-stamps information), and using one of GE's HISTORIAN HMIs--which requires a LOT of configuration, training, maintenance and experience.
Trender or Trend Recorder can be configured to capture data in several different ways--but it requires configuration to do so. There are other Toolbox/ToolboxST methods for capturing and recording data. But these methods will take some configuration, and they could be automated to a certain extent (using IFTTT or scripts, etc.). And the Performance Engineering Department can then set up some kind of data review and analysis using "triggers" and making calculations.
But, I can tell you--as you seem to be aware of--forcing operators to manually record parameter values is not going to measurably improve plant engagement. And, if anyone thinks it will--ask them to explain how the improvement in engagement will be measured and analyzed.
Best of luck! Human power plant operators have the most inertia of any entity known to man. They are HIGHLY resistant to change even if they are vocally in favor of it, some would even say they are highly ALLERGIC to change. They just don't like change. Full stop. Period. And, interrupting their "work" (as THEY perceive it) to manually record parameter values on a log sheet is pretty much a waste of money and time. One in a thousand operators MIGHT spot changes in a parameter over time and think it to be interesting--but they aren't likely to bring it to anyone's attention.
You would do better to use actual data from some kind of "event" to show operators how the event could have been predicted or even prevented--possibly getting them interested in being more proactive (a better thing than being engaged). And the more event data can be used to explain what happened and why and how it could have been anticipated the better the chances of improving proactivity (maybe even engagement).
Anyway, again--best of luck! There are much better methods of using the existing HMIs to gather data, high-speed and slow-speed, than having operators write parameter values down and handing the log sheets to the Performance Engineering Department. LOTS BETTER methods.
Facebook
Google
GitHub
Linkedin
