Many companies collect data from their manufacturing operations to increase productivity and improve shop operations. Others do so as part of a contractual obligation to their customers. Whatever the motivation, data collected and transformed into useful information can be valuable to any company. Unfortunately, there are too many cases where companies are collecting data because it is the 鈥渞ight thing to do鈥� 鈥� often performing little analytics on the data or having no one specifically responsible for analyzing the data. More concerning are those companies who invest in data collection and analytics but then do not act upon the information provided. In these later cases, data collection and analytics become liabilities instead of an asset.
There is no lack of software tools available for converting data into higher-level information. The challenge is finding the right tool(s) for your specific business needs. Unfortunately, there is no 鈥渙ne size fits all鈥� answer to this question. The key thing is to move forward 鈥� not necessarily holding off until the absolute best solution for your business becomes available. It is also important to have someone, preferably a team, fully engaged in the data collection process and seeking the resulting information.
There are many stories about the benefits companies have gained through their data collection efforts. The ones that I find most rewarding are those that demonstrate how different individuals perform their jobs when new information becomes available through data collection 鈥� much of which is never documented or even fully realized by the individuals involved. Here are a couple of real-life examples:
A large manufacturing facility had just implemented a centralized monitoring system for several essential pieces of equipment around the facility. This included several rooftop air handling systems. An engineer responsible for the project was demonstrating his new tool when he observed some critical information. One of the air handling units exhibited a higher vibration level than the others. It had not yet reached the point to trigger an alarm, but it was abnormal. A quick check of the motor on the unit confirmed that a mounting bolt was coming loose.
This early detection of the pending problem required tightening a single bolt and eliminated the potential premature failure of a $10,000 motor.
A data collection and monitoring system had been implemented on a complex production cell. The cell was experiencing periodic shutdowns for no obvious reasons. When inspected, all seemed well, and the system restarted normally. There was no apparent consistency between shutdowns. This occurred repeatedly over several weeks. While frustrating and costly, no one had been able to identify a reason for the failures.
I was lucky enough to be a bystander observing several engineers and maintenance personnel discussing the problem. They were in an office looking at data recorded from the cell. One individual eventually observed that a specific pressure sensor was showing an increasing value right before one of the shutdowns. With this insight, the team was able to trace data back over several days and found that this pressure started to build right before each shutdown.
The solution to the problem was the replacement of a $50 solenoid. Without the available data, the maintenance team would have had to be on the shop floor continuously observing the production line potentially for days, searching for a possible solution and then likely never identifying the specific problem while replacing parts until the problem went away. What is particularly rewarding about these two examples is seeing how individuals perform their jobs differently when data is available. In reality, they don鈥檛 typically even recognize how differently and more efficiently they perform their jobs. Some simple rules to keep in mind when considering implementing a data collection system:
Get started 鈥� the more you delay, the more potential benefit may be lost.
Select a software tool(s) that addresses specific business needs.
Make sure the information generated truly reflects your shop floor. Avoid software, or configuration of software parameters, that generate a desired result (poorly manipulated data) in lieu of an accurate representation of your shop operations (good or bad).
Have a team of individuals who are truly invested in the data collection and analysis process 鈥� they actively seek information, not having it thrust upon them.
And maybe, most importantly, document the resulting benefits. People change, management changes and priorities change. Without documented benefits, too many implementations (investments) fail when new individuals don鈥檛 recognize the value of the assets available to them.
You can learn more details about turning data into information in AMT鈥檚 DIGITAL MANUFACTURING WHITE PAPER SERIES: OVERVIEW OF DIGITAL MANUFACTURING: DATA-ENABLED DECISION MAKING.
