Understand the importance of the maintenance as a profit center

 

We need to understand the importance of the maintenance as a profit center where processes are managed and optimized, instead the idea of maintenance where invoices are paid.

 

Recently I attended an interesting webcast from plant engineering, its objective was to understand the importance of the maintenance as a profit center. I want to share some of the topics mentioned during this webcast.

 The Internet of things give us access to big data, data coming from smart machines or smart sensors. In addition to this big data, the use of predictive technologies and powerful algorithms like artificial intelligence, allow us to detect failures earlier, have more powerful analysis and understand better where the problems are coming from, instead of dealing with the symptoms that could take a lot of time. This provides a better Root of Cause Analysis (RCA).

 

The purpose of this technologies applied to the industry is to save maintenance time, and with a better RCA we can reduce the downtime.

  

Then, we can use the time saved in maintenance planning and scheduling. This planning and scheduling together with the RCA improvements produces a higher mean time between failures (MTBF), uptime and lower maintenance cost. All this increase the profits of the plant.

  

How can we handle the big data?

  

The access to data is getting bigger every year, the management of this big data has become a challenge, it is impossible for a human being to analyze this amount of data. That’s why the use of big data analytics is important to filter all this information. Data analytics starts looking for trends, for interactions between data, giving a better understanding about the data we are receiving.

  

But the information received from data analytics programs is still more than the human being can handle, if we consider that the human being can absorb in average 3 words per second, read or spoken. So, the use of spoken words or text in the industry is not enough method to communicate all the information we are getting.

  

In the other hand, a picture can express the idea of many words. The use of visual processing has become very important because it helps us understand better the information received from data analytics. 

 

That is the case of virtual reality (VR) and augmented reality (AR) technologies, with industry applications where useful virtual information is displayed on the top of real equipment using glasses, or smart phones or tablets. This provides a better understanding of what is going on the process or the equipment, making easier and more effective the maintenance job.

  

Also, the AR/VR is helping train personal on their jobs. There are VR welding simulators, that allow for hands-on experience, without the cost of learning on actual materials.

 

Smart glasses are seeing adoption on factory floors and in fields like construction and oil and gas, where hands need to be free, but tech tools can increase efficiency. The maintenance personal can make a repair on a piece of equipment, and having a guide on how to do that, right in your field of view.

 

The AR wearable lets users do things like see through faulty pipes or overlay directions onto the helmet's display. 

Some of the opportunities related to the use of big data coming from the IIoT:

 

•Collecting data without a purpose, not useful and meaningful information. Industries needs to look for the way to make the information useful.

 

•Big amounts of data should be processed by effective algorithms. Engineers need to understand the process and how to apply specific algorithms to the process. Ineffective algorithms are not going to help engineers to make right decisions.

 

•Conditional algorithms will need to be created for process and maintenance changes, for example, changes introduced by new products.

 

• The organization needs to trust or believe in the information they are receiving. If they think the information doesn’t match what they see over the years, then the organization is not going to make use of the information derived from the data.

 

•Sometimes industries are getting the data but it is not communicated or reported effectively. 

 

•The organizations don’t have the systems, procedures or processes in place to use the data correctly.

 

•Creates “tail-chasing” tweaks and sets off avalanche of changes.

 

•The industries need to be prepare to add complexity, sensor calibrations, PLC analog modules.

  

Predictive maintenance technology:

  

The use of predictive technologies allows the algorithms to see the issue before the equipment fails. The use of predictive technology with the IIoT and the appropriate algorithm can give us a clear idea about the source of the issue.

 

The statistics show that 68% of process or equipment failures are introduced from the beginning, form the start up, this is called the infant mortality curve. This means that the defect is taking place from the beginning and this defect is like a time bomb, waiting for that problem to happen.

 

With the use of IIoT to get the data, predictive technology and the appropriate algorithms, this infant mortality defects can be detected and fixed easily.

 

The I to P to F curve is a model that shows how an installation of asset could become with the time a point of failure and them to a catastrophic failure. The cost to do the repair gets higher as we go close to the catastrophic failure. So, if we can recognize and fix the issue much earlier, then there is a save of time and money. Time that can be used for maintenance planning and scheduling.

 

As we see on the curve, the predictive maintenance takes place from the first symptoms, from the point where the failure status occurs. Then the preventive maintenance takes place when the audible noise and hot to touch scenarios occur. At the end the corrective maintenance occurs when the equipment functionality fails.

 

I to P to F curve of maintenance

Benefits of big data to the maintenance:

1. Savings from the reduction of inventory costs due the early detection of failures.
2. Reduction of the preventive maintenance.
3. Better verification of the maintenance performed.
4. Better identification of the infant mortality failures.
5. Better Root of Cause Analysis.

To be successful using big data, the industry must have a good planning, efficient algorithms and processes to convert data to information. The information needs to be displayed meaningfully, and train people about how to use this information effectively. Also, the company needs to have a good MOC (Management of change) process to maintain the data going into the algorithms and avoid tweaks.

Not all the companies are ready for this kind of technologies, I want to finish this post mentioning 7 questions that Shon Isenhour, Principal, Eruditio, LLC, did to evaluate if your organization is ready:

 

• Are business processes established, documented and adhered to?

 

• Does your organization have a mature calibration program?

• Do you plan work effectively?

• What is the “size” of your scheduling window?

 

• Do you effectively manage your backlog?

 

• Do you effectively utilize Predictive Maintenance (PdM) Technologies to identify failures early on the P-F Curve?

 

• Does your processes effectively utilize PdM data and a planning & scheduling process to correct findings before they reach functional failure?