This group brings together the best thinkers on energy and climate. Join us for smart, insightful posts and conversations about where the energy industry is and where it is going.

Post

Energy performance monitoring enables condition-based maintenance

Vilnis Vesma's picture
Director VESMA.COM LIMITED
  • Member since 2021
  • 11 items added with 5,391 views
  • Jul 19, 2021
  • 701 views

SOME years ago I was invited to Sweden to run a training course on energy monitoring and targeting for an engineering manufacturer. To help me customize the presentation, they sent me some weekly consumption data from the factory so that I could show them tools like regression analysis and cusum applied to their own circumstances. Among the data were weekly consumption and output figures for their air compressors. The last few weeks of compressor data were missing, which was frustrating because just before they stopped there were clear signs of a significant adverse change in performance. The data gap after the change left me unable to discern what its new behaviour looked like.

On the day of the training I asked if we could have a look at the compressor house to see if there were any clues about why it was consuming excessive electricity. "Sadly not", they replied: "it burned down a few weeks ago". I leave the reader to draw their own conclusions.

--o--

On a more positive note, I once had an assignment setting up energy monitoring and targeting at a chemical works, specifically a set of distillation columns. Being a highly-automated process plant it was well-provided with temperature and flow measurements which were being logged at one-minute intervals. The process was a little more complicated than some: it was handling a mixture of three fluids, one of which was the desired product while the other two were volatile solvents which were being boiled off. How to calculate expected steam consumption? My solution was to sample the flow and temperature data at 20-minute intervals, and using the reported blend proportions, compute the latent heat implicitly absorbed by each solvent and the increase in sensible heat. Basically school physics. The 20-minute estimated heat demands were aggregated into a weekly expected consumptions and could then be compared with actual weekly steam use.

My primary purpose was to detect and quantify deviations from expected steam consumpton. Excess consumption was always a risk, and likely to arise because of degradation or fouling of the internals. Distillation columns with sufficiently-costly excess steam use could then be taken out of service early for maintenance, with production shifted to the most energy-efficient alternative units.

The columns were high-temperature, high-pressure vessels that ran continuously for nine months at a time between scheduled maintenance shutdowns. There was potentially a lot of money to be saved by postponing maintenance where it was safe to do so, but there was no way of directly observing them internally. So the owner had previously spent tens if not hundreds of thousand of pounds on an artificial-intelligence-based condition monitoring system, analysing the plant data in real time to determine if each column was healthy. This had not proved effective. However, my energy monitoring and targeting scheme gave them condition monitoring free of charge. Any internal changes within a column would change its energy demand characteristic for the worse, but conversely if a column continued to consume steam in line with expectations, it could only be because it was internally intact and did not yet need to be stripped down. 


 

Vilnis Vesma's picture
Thank Vilnis for the Post!
Energy Central contributors share their experience and insights for the benefit of other Members (like you). Please show them your appreciation by leaving a comment, 'liking' this post, or following this Member.
More posts from this member
Discussions
Spell checking: Press the CTRL or COMMAND key then click on the underlined misspelled word.
Bob Meinetz's picture
Bob Meinetz on Jul 19, 2021

"There was potentially a lot of money to be saved by postponing maintenance where it was safe to do so, but there was no way of directly observing them internally. So the owner had previously spent tens if not hundreds of thousand of pounds on an artificial-intelligence-based condition monitoring system, analysing the plant data in real time to determine if each column was healthy."

Once again, Vilnis, we see the limits of AI-based analysis. It's one thing to allow a monitoring system to present an operator with an accurate picture of column conditions, another to allow it to decide whether each one is "healthy".

Jim Stack's picture
Jim Stack on Jul 20, 2021

I worked at the Telephone company and we put in DAS systems at all utility power plants. The data was very valuable but sometimes they were not checking it. In later systems they could set automatic levels and data so they would be alerted to any times when things were not normal. It really makes a difference to be able to be any problems as soon as they happen. 

   My home solar system also has data I monitor for trends. I also own 2 large PPA Solar contracted systems and at the biggest one I no longer have the live data. A truck hit a support and caused a partial outage that I only found out about a few weeks later. Having live data is very valuable. I should be part of all power systems and loads.    

Get Published - Build a Following

The Energy Central Power Industry Network is based on one core idea - power industry professionals helping each other and advancing the industry by sharing and learning from each other.

If you have an experience or insight to share or have learned something from a conference or seminar, your peers and colleagues on Energy Central want to hear about it. It's also easy to share a link to an article you've liked or an industry resource that you think would be helpful.

                 Learn more about posting on Energy Central »