Pete: It really comes down to using data to make decisions.  If you are going to improve efficiency it starts by measuring the current state.  Sensors at the process floor level are used to gather data and develop trend lines of how a current process or piece of equipment is running.  The first step is to identify variation or any signs that the process is moving out of control.  In most cases variation is the enemy of efficiency.  Digging into the variation and identifying the root cause is the next step.  Corrective action could be maintenance of equipment, adjusting process parameters or configuration settings, or correcting operator behavior.  What IoT brings to the improvement process is the ability to bring all the data from the different sensors to a central location so all the data is in one place and can be overlayed to understand correlation and causation.

When it comes to production, the focus is often on lost production time, what is causing production to stop or slow down.  It still is an exercise of chasing down variation and using the data to really understand the root cause.

Pete: That is an interesting question because they are often two different things.  The first place to look at for quick improvements is error logs, alerts and alarms.  Most control and SCADA systems track alerts and alarms along with process data associated with IoT.  Operators and mill managers are notorious for acknowledging alarms throughout the day because they need to make production.  The system is alarming for a reason and it is usually a maintenance issue.  A sensor is bad, a piece of equipment is not cycling properly, a limit switch is worn out.

Long-term scalability is a different animal.  This usually means reviewing the current state capacity of a given process and comparing that to what the process should be capable of in theory.  If you should be producing 40 tons per hour out of a given process and you are only producing 30, you need to start analyzing the data to categorize what faults are creating lost production or restricting the process.  Oftentimes it is not just one thing.  By trending the data and keeping track of the faults and the amount of lost production for each fault you can use the Perato process to sort the faults from highest to lowest.  Attack the highest to lowest by again finding the root cause.  As you do make changes and improvements, continue trending the data so you can see the impact of each improvement on gaining production capacity.

Pete: One of the things we will see on batching systems is a shift in the time it takes to complete a batch.  We will use the scale data to look at individual materials coming from their respective bins.  We will see months and months of consistent scale readings where the system was accurate and repeatable.  Then suddenly, we will see variation and long scale settle times because the system is constantly jogging to make the correct weight.  The system was in control and capable, then suddenly not.  A quick check of the equipment will show worn screws on the feeder that was causing variation.  Replace the screws, the system is back in control and the batch times return to normal.

Pete: The Ensch household is ruled by two Labrador retrievers, a 12-year-old chocolate named Xena, the princess warrior, and Jet, a 6-year old black lab.  They are great with the family and also good bird-dogs.