What is Model-Predictive Control?

CLPM Not Only Complements Your MPC Solution It Allows Your MPC Solution to Perform Better!

Model-Predictive Control (MPC) is advanced technology that optimizes the control and performance of business-critical production processes. So is Control Loop Performance Monitoring (CLPM) software. But if both help practitioners to optimize control loop performance, then what’s the difference?

MPC tools like Pavilion® and aspenONE are highly advanced, autonomous tools that actively monitor production characteristics and take actions independent of a facility’s production staff. Utilizing empirical models of each process’ dynamics in conjunction with sophisticated algorithms, MPC technologies adjust one or more dependent variables such as a control loop’s Set Point or a valve’s position to counter the potential impact of changing independent variables (i.e. disturbances). In fact, they will simultaneously adjust numerous dependent variables associated with many, many PID control loops to achieve the desired performance. That said, MPC technologies don’t identify performance issues for the purpose of correcting them. Instead, MPC technologies identify issues in order to compensate for them.

CLPM solutions like PlantESP actively monitor a facility’s control loops with the goals of isolating changes in performance and of correcting them. They utilize an extensive set of KPIs in order to identify issues ranging from PID controller tuning and mechanical (i.e. Final Control Element) to loop interaction and process architecture, all of which affect process dynamics and how a control loop performs. So just like peanut butter and jelly, MPC and CLPM are complementary technologies. By isolating control loop performance issues CLPM solutions empower production staff to make timely corrections thereby maintaining the integrity of the models by which MPC solutions function.

Certain aspects of MPC underscore the complementary nature of CLPM solutions. Consider the following:

• The Model Matters

MPC relies on accurate models in order to perform effectively. In fact, the first step when implementing an MPC solution is to calibrate each PID control loop to assure that the models are based on the performance of well-tuned controllers. However, the built-in process modeling and controller tuning feature found in MPC technologies are ill-suited for use with highly dynamic processes. In contrast select CLPM solutions have a built-in modeling and tuning function, and one in particular can generate accurate models in the midst of oscillatory and noisy process conditions, what we at Control Station refer to as the “real world.”

• Change is Constant

Virtually everything within a production facility is dynamic, especially the production processes themselves. Even the most basic wear and tear affects a process’ dynamics. Although MPC is well suited to adapt to process disturbances, the models upon which MPC operates are static and they cannot account for continuous change. CLPM solutions ferret out such changes so that optimal production can be maintained throughout a facility’s complete life-cycle.

• Proactive vs. Reactive

Changing process dynamics eventually relegate MPC’s value proposition from proactive to reactive optimization. That is particularly true as engineering staff make modifications to the means of production on a daily basis, a valve here, a damper there. Without awareness of and adjustments for these incremental changes MPC ultimately becomes a hindrance. It autonomously takes actions based on flawed models that undermine production efficiency and throughput. That’s where CLPM solutions step in.

A common misconception is that MPC and CLPM solutions provide redundant capabilities and are therefore competitive.  In truth, they are wholly complementary and amplify their respective value propositions. Equipped with real-time insight into changing process dynamics, MPC technologies gain the ability to deliver optimal production over the long-term. With its focus on the performance of each individual control loop, CLPM solutions provide just that.