Pavilion8 Control Applications are designed to meet the most demanding needs of today’s process manufacturers, whether they are making polyethylene, cement, ethanol, milk powder, paper, etc., Pavilion8 Control provides a computationally efficient multivariable model predictive control (MPC) technology for both nonlinear and linear applications. The foundation is our advanced hybrid modeling technology, combining both empirical process data with first principles knowledge of the process to create the most accurate process models possible. This sophisticated controller also incorporates variable dynamics for the tightest control during even the largest product grade transitions on complex, nonlinear processes. Integrated steady state and dynamic optimization ensures optimal control performance within process constraints and while meeting other objectives such as maximizing production or minimizing energy.
Extracting greater value from manufacturing assets is a major challenge. Companies seeking to increase profitability are shifting to customer-centric, demand-driven manufacturing environments where product quality and exemplary customer service is becoming just as essential for success as price. Moving to a more demand-driven operation offers significant benefits; but it is also more demanding on manufacturing operations. Manufacturers must be prepared to drive out inefficiencies during both steady-state operations and during transitions. According to leading analyst firm ARC Advisory Group, most companies rarely achieve more than three to four Sigmas during steady state operations and typically much lower during transitions. Successful demand-driven manufacturing requires companies to attain nearly flawless execution. Companies that approach Six Sigma during sustained operations and perform grade changes in the shortest amount of time and with the least amount of off-spec product have a distinct competitive advantage. Model predictive control (MPC) technology plays a critical role in a demand-driven manufacturing environment. MPC is a proven technology that reduces process variability and inefficiency, improves product consistency, and allows operations to push constraints to the limits. With more frequent grade changes, MPC is needed to effectively manage transitions and impart the necessary agility in accomplishing customer centric objectives, and improve overall business performance.
Controlling a major process unit effectively usually means dealing with multivariable systems. It is extremely unlikely that treating each control loop independently will provide optimal control. In most situations, the control action of one loop affects the other loops. When significant interactions among the loops exist, optimizing loops independently usually results in unstable situations. Leveraging a robust predictive modeling engine, Pavilion8 relies on model predictive control (MPC) technology to create a model of the process that accounts for multiple variables that affect the process outcome. A model will incorporate those variables you are trying to control like moisture, feed rate, melt index (control variables), the variables you can adjust to reduce variability (manipulated variables), and the variables that you cannot control but that affect the final result (temperature, energy prices). Pavilion’s MPC technology addresses these issues by taking into account the interactions among the process variables. MPC uses this model of the process to predict how the process will respond to future changes in the process or outside disturbances. This predictive capability allows the controller to determine the best way to adjust the process input variables to drive process output variables to their optimum targets while considering interactions and remaining within any imposed constraint specifications. This approach is necessary in creating a demand-driven environment, flexible enough to respond to market disturbances and demands.
Pavilion8™ provides hybrid modeling functionality that allows the user to take advantage of all of the known process information including empirical data, first principles equations, equipment specifications, and operator knowledge when building a model. Hybrid modeling results in a more robust, more accurate model across the entire range of operating parameters. This capability allows users to build robust models faster with higher accuracy than alternatives. Furthermore, because the hybrid model takes into account all availability information, the resulting control application is more representative of a wider operating range, thus increasing the sustainability of the application and the correlating value.
Pavilion8’s controller includes the ability to handle variable dynamics, explicitly accounting for the dynamic behavior of the process across a wide range of process conditions. This capability enhances product quality and improves transition management. Grade transitions can be a major source of off-spec product in most manufacturing plants due to the difficulty in controlling the process smoothly through the entire transition. In order to minimize this transition product, many manufacturers use a ‘product wheel’ to make very small transitions in a particular sequence which severely limits production scheduling flexibility. Pavilion8 accounts for the varying dynamic behavior of the process across grades. Even transitions from one end of the product spectrum to the other, can be accomplished efficiently with minimal off-spec production. Variable dynamics capability also provides flexibility to adjust production schedules to meet new orders, regardless of what is currently being produced.
Sustaining value over the lifecycle of a control application is essential to maximizing the value of your assets and achieving continuous improvement. Evaluating control performance issues with complex multivariable controls can be challenging. Via a browser-based interface, Pavilion8 provides effective controller performance metrics to help pinpoint problems by quantifying utilization, variability, deviation from target, and time at constraints for manipulated and controlled variables. By being able to correlate quality problems with specific control situations and disturbance variables, problems can be isolated and corrective action taken. By providing role-based access, different users can view controller performance anytime and anywhere to ensure controller uptime and extract greater value from manufacturing assets.