Turning Data Into Profits


Most modern control systems collect data—lots of data. Some control systems store every analog input every second for 30 or 90 days. Maybe your system collects data in CSV files or perhaps in a SQL database. When the control system was installed, the integrator might have asked what data you wanted to collect. But how are you to know which data you do or don’t need to collect? So we collect it all.

If we are lucky, we might see trends that help us interpret how the process is performing. We might use logged data to determine what went wrong after a process upset. We might use data to refine the tuning on a loop. These are all good ways to use data, and likely have added some incremental value to your facility.

But does your plant data still have untapped value? What if you could remove bottlenecks and increase your facility’s production by 5 percent? What if you could reduce steam usage and reduce the amount of energy needed to produce each unit of product by 5 percent?

Could you leverage your facility’s data, analytics and model predictive control to produce tangible value? Perhaps a 50 percent reduction in variability, 5 percent reduction in energy input per production unit and/or 5 percent increase in product produced.

Model predictive control (MPC) is an advanced process control technique that solves complete process control scenarios to determine the optimal settings for control variables. MPC is uniquely capable of optimizing processes that have significant lag time, such as temperature control processes like heating, evaporation and drying processes.

A successful MPC project begins with data. Remember all that data you’ve been collecting? You can put it to use on an MPC project that saves you money and has a calculated and proven return on investment.

Start with the data and use analytics to determine which process variables are interacting with other process variables. Combine analytics with process expertise, and you can build complex models, matrices of manipulated variables, disturbance variables, controlled variables, hard constraints, soft limits and desired setpoints. These models configured in an MPC solver and integrated into a modern control system can often provide significant improvements in productivity and efficiency and reduced variability of finished products.

Consider using MPC and analytics to turn your data into profit.

______

Monte Vander Velde, Interstates Strategic Projects Manager