Control of polyolefin production plants: practical examples from nonlinear model predictive control to monitoring instrumentation

Polyolefin has revolutionised life and also saved numerous ones. Radar would not have existed, electric cables would still use paper and gutta-percha insulation, and supermarkets with their pre-packed see-through wrapping would not exist had it not been for polyolefin. The main raw materials, ethylene and propylene, typically come from ethane or naphtha crackers. In such crackers, advanced process control (APC), such as model predictive control (MPC), and also real-time optimization (RTO) have been in common use for a long time. The situation is different in the production plants for polyolefin: APC technology has not been in such wide use as in the olefin production plants. Various factors can contribute; amongst these is the need for coping with the nonlinearities observed in producing polyolefin. Examples of applying APC for the polyolefin industry are drawn from proprietary-developed controllers. Amongst these are the world's first nonlinear MPC based on physical models (implemented at an industrial-scale polypropylene plant in Norway back in 1993) as well as a model-based reactor controller (also based on physical models) for the high-pressure polyethylene production that has been operational from around 1990, also in Norway. However, there is more to the control of industrial polyolefin production processes than selecting a control algorithm. Other parts of the control system need to behave as well. Examples of adapting and implementing simple methods to monitor the state of the various control components are discussed. An example of such monitoring is the stability of pressure transmitters used for the control of high-pressure polymerisation reactors. Monitoring the injection of initiators is also worthwhile, as uncontrolled variations can affect the reaction adversely, even resulting in a loss of production. Valve hysteresis in valves that are extra strained will affect the stability of the reactions as well as the properties of the polymer produced. Ways of investigating the effect and of rectifying deficiencies are addressed. The observation of the variation in process measurements is commonly used as a way to assess the goodness of control. Utilising this, and recognising that polyolefins are soft materials that need to be be handled with care, enables the development of methods to further improve the quality of polyolefin delivered to the customer, and such methods are discussed.