The classic case for process control improvement uses a figure that shows the mean and standard deviation of a statistical distribution and the optimum of a process variable (PV). The optimum is often taken as the constraint on product quality. The case is made that if PV variability is reduced, the setpoint for the PV can be moved closer to the constraint. In my experience the bias to the setpoint due to variability is less than the bias by the operator because of measurement error, poor measurement and valve turndown, manual actions, on-off actions, abnormal operation, unknown changes in raw materials, equipment deterioration and modifications, process modifications, lack of process knowledge, operator “sweet spots” and war stories, and tactical tradeoffs between capacity and efficiency.
In this tip we will address the automation system considerations of improving setpoints, but there are bigger questions such as integration of the knowledge of the effects of equipment, piping, and process conditions and changes into the decision to move the setpoint. Simulations can provide the process knowledge. Unfortunately, maintenance and operational databases have not achieved the level of integration needed to understand how a setpoint should be changed, as noted in the Control Talk column Drowning in Data, Starving for Information – Part 4.
You may just need to move the setpoint. Better process knowledge or a simple trial run at a lower or higher setpoint might be sufficient. I have seen cases where running models provided the knowledge and confidence to try a better setpoint. In other cases, big improvements were directly made by engineers with practical process and control capability. Simply entering a better setpoint may get you a bonus, or at least a free meal.
A common example is temperature control. Distillation columns rely on temperature for an inferential measurement of composition. Temperature determines both production rate and quality in reactors. Reaction rate increases with temperature but just past the optimum, reverse reactions, side reactions, and product degradation may occur. Consequently, there is a peak in the plot of production rate versus temperature.
For bioreactors, the specific cell growth rate and product formation rate increase with temperature but cells start to shut down and die at temperatures to the right of the peak in the plot of specific growth rate versus temperature. The peak is different for cell growth and product formation, so temperature shifts are made generally after the exponential cell growth phase is fully established. However, the best size and timing of the shift are often not known. A similar situation exists in bioreactors for pH, with an even sharper peak. Measurement accuracy and control requirements of a few hundredths of a pH of setpoint are stated. The question is, do we really know the best setpoint to the same degree of precision?
Concept: Unless off-spec product is being downgraded, rejected, or recycled, the benefits of process control improvement are not realized until a setpoint is moved closer to the optimum. Greater plant knowledge offered by simulation and the integration of databases can lead to more intelligent setpoints. Better control strategies, measurements, control valves, and tuning and higher levels of control can enable operation closer to a constraint.
Details: Eliminate instrumentation and valves as the cause of a bias between the setpoint and the optimum. Coriolis meters, magnetic flowmeters, and precision throttling valves with sufficient pressure drop can eliminate limit cycles and errors in the ratios of flows and improve plant turndown. Resistance temperature detectors (RTD), integral mounted temperature transmitters, direct mounted pressure transmitters, and radar level indicators reduce measurement errors. The latest improvements in sensor, transmitter, and positioner technology can eliminate setpoints being shifted due to automation system limitations. Develop plant knowledge to find more optimum operating points. Process simulations, integrated maintenance and operation databases, and an asset management system (AMS) facilitate finding better setpoints.
Use a higher level of control to automatically find better optimums. Analyzers (Tip #63) offer a higher level of control. Feedback loops that fully exploit smart PID features (Tips #91-96 and #100) can eliminate operator actions (Tip #69), on-off actions, abnormal operation, and activation of safety instrumented systems (SIS). Valve position control (VPC) strategies described in Tip #97 and model predictive control (MPC) and a linear program (LP) can automatically optimize the setpoints of unit operations. Real-time optimization (RTO) can provide the ultimate in optimization of setpoints throughout a continuous plant if there is an accurate model and sufficient RTO expertise onsite.
Watch-outs: Operations personnel may be reluctant to believe that manual actions can be eliminated or that operator sweet spots and war stories are not valid reasons for operation further from a constraint. If there are no online process metrics (Tip #61), operators will naturally choose the setpoint that minimizes any perceived potential disruption, stress, and extra work. If the goal is business as usual rather than a more profitable business, Operations will be reluctant to make changes. If the operators are not fully trained in the higher levels of control (Tip #99) or control room support is not provided for all shifts for a sufficient duration, new control systems will be put in manual whenever something unusual happens.
Exceptions: If there are large blend tanks to attenuate product variability and if process capacity rather than process efficiency is the goal, flows may simply be set at a maximum. For conventional rather than fed-batch operations with only on-off valves (no control valves), the optimization involves batch sequences and totalized batch charge flows rather than setpoints.
Insight: Setpoints can be improved by achieving a higher level of control and knowledge.
Rule of Thumb: Use automation systems and plant knowledge to operate closer to constraints.