In today’s landscape of automated industrial engineering, there is an overwhelming amount of data generated. Sensors continually measure pressure, vibration, flow, speed and everything in between. All of that data can provide exceptional value in an industrial environment. But gathering such a vast quantity of data from so many collection points is challenging, to say the least. So when faced with the prospect of implementing big data collection and analysis, what should an automated manufacturing facility expect? And what is the best way to proceed?
What is Big Data?
The term big data refers to data sets so complex and massive that traditional data analysis tools and techniques are practically ineffective. In an industrial setting it can comprise data input streams from the entire production process.
A helpful context for understanding big data is the five Vs: volume, velocity, variability, veracity and value. These refer to the fact that the raw data elements constituting big data are generated at a tremendous volume, are transmitted at near instantaneous speeds, come from many different sources, constitute a complete set for reliability and provide deep value that you can’t get otherwise.
By applying a framework of big data analytics to industrial engineering, you can get answers to questions such as:
- Why has machine downtime in a given line risen by 16% this month?
- Why has production dropped 12% in one facility while production in another facility has improved by 7%?
- Which step in the production process is responsible for a certain defect?
- Why is my facility producing 10% more waste this month?
Applied effectively, big data analysis can be leveraged in an automated industrial setting to more fully understand root cause issues and improve performance. Historical process data within a plant can be collected and used to identify variable costs for various points in the line. You can then take that information and use it as a factor to inform the development of a process strategy and drive process improvement.
Applying Big Data To Derive Value in Industrial Engineering
In an automated industrial environment, you can leverage big data technologies to drive improvements in performance in a number of ways: forecasting, predictive modeling, evaluating performance metrics, revealing surprising insights and making prescriptive decisions based on real-time data analysis.
It might seem like the ideal solution then is to install as many data collection points as possible within your manufacturing environment. With more sensors, you can collect more information for, in theory, a more robust view of your facility and equipment.
I would strongly argue against that; at least, your plan should be to implement automation and monitoring on a case-by-case basis. You may hear someone say, “Yes, more data is better,” but that’s not always the case. As with all things, there is such a thing as too much automation and too much monitoring.
If someone says to you, “I want to collect this data point,” stop and ask yourself a couple of questions: “What will it do for us?” “What’s the value in it?”
Automation, monitoring and data can be incredibly valuable assets, but you have to make sure they are providing value commensurate to their cost. As a case in point, think about any public bathroom you may have been in lately — I look at that as automation gone crazy. I can flush my own toilet. I’m able to turn a faucet on and off. I can turn a knob to get a hand towel. And I don’t need to press a button on a screen to tell a facilities manager my 1-10 ranking of the experience in the bathroom.
That’s not to say that those instances aren’t helpful or convenient; they certainly can be, and in that particular case they can provide a hygienic benefit. But to me, it’s all about the return on investment. In your facility, what’s the ROI on implementing further automation and monitoring devices?
Root Cause Analysis for an Improved ROI
Connecting sensors to your facility’s equipment can provide you with valuable insight into a variety of data points on your machines or process. Using baseline measurements, you can compare relevant data from those sensors to a rules engine, which can then suggest changes to the process or provide warnings when critical thresholds are reached. By developing a rules engine on the basis of historical data, you gain a complete picture of your process that can deliver actionable, prescriptive analytics.
How can that benefit you?
Assume a vibration sensor on a pump is sending a signal that the motor will likely fail in the near future. The operator’s initial assumption is that it will need to be replaced, causing an interruption in the production line.
However, if your production equipment is all part of an integrated process, with a supervisory control system consolidating inputs from sensors across the whole production environment, you may find that the root cause of the pump motor’s impending failure is actually low pressure in an upstream line causing cavitation in the pump. A holistic view that provides root cause analysis lets you see the true issues in your facility; rather than replacing a pump only to have the same line pressure issue also prematurely wear out the replacement, you can see where the real failure points are and address them before they create significant and repeated negative impacts.
The need then is to balance the implementation of monitoring systems and collection of big data within the automated industrial engineering environment with a specific eye toward the value you can expect back in return. Where will downtime hit you with the biggest impact on production? What equipment is critical in your processes to the point that you absolutely have to know, in advance, if something is going to fail?
Starting with the assets that are essential to the success of your process, evaluate where sensors and monitoring devices will provide the most benefit and work down the line from there. If you have sensors on every control valve in your facility, you might want to stop and ask yourself if the data those sensors provide is necessary and if you’re getting enough value out of them to justify their cost.
Problem Solving in an Automated Environment
As much as you automate processes in a facility or implement monitoring systems, things don’t always go according to plan. Regardless of how much you apply big data analytics to your processes, equipment fails and provides you with upset. You can’t fully automate industrial engineering; you’ll always have line operators, control room operators and lead operators. Automated process control drives efficiency, but in a manufacturing facility where you would have needed 40 people to produce your daily output, you still need a few people to provide control and oversight over your data-driven industrial automation engineering processes.
Sometimes you already possess the needed knowledge base internally and you can run your facility without any outside help. But at other times you may need to bring in a fresh set of eyes and have someone with broad experience across industries come in to consult and help get you past some unexpected bumps in the road.
When Percipio comes on board in an industrial automation engineering consultancy, we evaluate the entire system and process. That can take on many different forms, depending on the need at hand. Are you looking to build out a new plant? Or are you considering an expansion? Have there been equipment failures that are forcing the question of whether or not to implement some higher level of data collection? Quite often there is a set of operational issues that the on-site team just can’t seem to figure out.
Percipio’s consultation process for automated industrial engineering probes on whether your installed automation process control solution is providing you the benefits it should. Is it giving you satisfactory results for your production facility? We have experience assessing what can be done to potentially lower costs, apply helpful and value-adding technologies, and optimize production.
In investing or consulting, our approach centers on value, and we don’t recommend anything without the expectation it will drive a positive ROI.
In automated industrial engineering consulting, we see a significant amount of demand but not a lot of expertise to meet that demand. As technologies continue to evolve and automation drives more processes, experts are needed who can help on-site personnel and plant managers effectively understand how to join the control systems and technology to drive value-based solutions to clients’ needs. If you need a partner to come alongside you and lend their expertise to your situation, reach out to us today. We’re always ready to apply our experience and capabilities to produce data-driven, value-add solutions to complex automated industrial engineering challenges.