Introducing MPC and Analytics
Model predictive control (MPC) technology helps operators maintain quality while optimizing production by managing minute details and adjustments that allow you to push against process constraints. Based on a model of your process, the technology makes constant adjustments to maintain the ideal recipe. Wherever adjustments are made earlier in a process, MPC helps you make any related adjustments needed downstream.
Predictive analytics look for faults in operations using advanced machine learning and other data processing techniques. Analytics applications create models for how equipment should perform normally and alerts operators when there’s a deviation.
Together, the technologies help flag problems quickly, identifying the source and enabling an early fix. You can even catch potential problems before they occur.
How to Ferment Improvements
To get a closer look at the technology in action, let’s consider an example issue with the process of emptying a fermenter. The connectivity to multiple data sources enabled by an information platform can pull data from multiple sensors that indicate a pump is running at its normal speed, or the tank level is dropping at its usual rate, but the product flow is lower than expected based on the established models. Such a situation indicates a possible problem with the flow meter, or a leak somewhere in the system.
Putting analytics such as these alongside of MPC provide additional value from the existing instrumentation and platforms.
With further analysis, it could be decided that the MPC application might identify a needed process adjustment, such as running the pump faster, helping to close the gap and correct the model. There also could be a glitch that needs to be fixed in a different part of the process, such as another pump adding new materials to the fermenter while you’re still emptying the previous batch – an issue we’ve seen with ethanol in other industries.
Of course, the process might be working fine – but one of the sensors is failing and needs to be repaired, recalibrated or replaced. While MPC makes the best corrections possible, predictive analytics can scour the system and identify the source of a potential issue. By flagging the discrepancy, maintenance personnel can check the status of the sensor and make any needed fixes.
Ultimately, whatever the issue, MPC and analytics applications will save you precious time by adjusting the process to keep it working while an issue is identified. Acting quickly allows you to protect product quality and yield, reduce waste as well as reduce unplanned downtime and wear and tear on your equipment.
Pairing these two technologies – just one element of what an information platform can provide – can significantly enhance the day-to-day operations of your facilities. Your workforce has a clearer picture of what’s happening in the environment around them. Quality control becomes more efficient than a series of routine manual tasks unlikely to uncover problems in real time.
On the path to digital transformation, smaller steps can still have a huge impact on your business. Are you ready to get started?
Learn more about MPC from Rockwell Automation.