Better control through predictive behaviour

Paul Boughton

In order to optimise plant operations at its chemical production facility in Spain, Solvay looked for a partner to help increase production and achieve better control through predictive behaviour.

Solvay is an international chemical and pharmaceutical group with headquarters in Brussels and operations in more than 50 countries. At its Torrelavega production centre in Spain, it produces a wide variety of chemical products. With various control solutions already in place, the company looked to further optimise its plant and incorporate the best elements of traditional PID algorithms and model-based control and optimisation technologies.

Solvay's plant in Torrelavega, Spain is a production centre for chemical products such as soda ash, sodium bicarbonate (BICAR), chlorine, sodium hypochlorite, hydrochloric acid and caustic soda. In order to optimise plant operations at its chemical production facility, Solvay looked for a partner.

With Honeywell's Experion PKS and TPS control systems already installed at the Torrelavega site, Solvay had extensive knowledge as to how these solutions worked. This assisted the company in implementing Honeywell's Profit Loop algorithm for most of its existing control loops.

Profit Loop is Honeywell's algorithm that represents a single input/single output (SISO) model-predictive controller specifically designed with the operating simplicity and computational efficiency of a standard PID controller. By using a simple model of the process to predict the effect of past, present and future control moves on the process variable, the controller knows how much to move the process in order to meet a company's desired control objectives.

Having already implemented Honeywell's Experion Process Knowledge System and TotalPlant Solution control systems, Solvay had extensive knowledge of Honeywell's control solutions. The company subsequently selected Honeywell's Profit Loop, a patented algorithm representing SISO model-predictive controller to help predict the effect of past, present and future control moves on the process variable.

By using the Profit Loop algorithm over standard PID, Solvay identifed several benefits:

Easy Loop Tuning - With Profit Loop tuning, it was no longer necessary to adjust the gain, reset and derivative times. With only one parameter, via the performance ratio, the 'one-knob' tuning enabled single-variable adjustment depending on the loop response.

Robust Range Control Algorithm (RCA) - The use of Honeywell's RCA allows Profit Loop to control the process variable within a dead band without controlling to a particular setpoint. This is especially valuable for tank levels in which the process variable is maintained within those levels while significantly reducing control valve wear. The solution minimizes valve movements as well as valve maintenance.

User-friendly Profit Loop Assistant - When other PID equations, operators had used the assistant for loop tuning. This acts in a way that performs the steps on the output values and stores the results. When a result is reached, there are two options: either a model with the transfer equation or suggested parameters dezpending on the type of loop and on the required response. This provides signficant time savings when implementing PID gain, reset and derivative parameters.

"We found that Profit Loop incorporates the best elements of traditional PID algorithms and model-based control and optimization techologies," said Rafael Truan Cacho, Control and Instrumentation Manager, Solvay Torrelavega.

For more information, visit www.honeywell.com/ps