The air we breathe consists mainly of nitrogen, oxygen, and argon. These components have various applications in many industrial sectors from food/beverage to energy and healthcare. The separation of these components can be done at extreme cold temperatures, i.e. more than 100 oC lower than the coldest outside temperature ever recorded in the Canadian history. Although the raw material, air, is free and plentiful, such separation processes usually start with an energy intensive step Ė compression of air. The electricity price deregulation in 1990s has caused knock-on effects in the operation of air separation plants (ASUs). On the one hand, customers would like to take advantage of changes in the electricity price, so they adjust their production levels which varies the demand of ASUís products. On the other hand, the ASU itself may also prefer to reduce operation costs whenever possible. The consequences are more significant and frequent fluctuations in desired production rates. As business competition becomes more and more intense, the winner/survivor is more likely to be the one that can produce products that meet customersí expectations at the lowest costs. Being able to follow variations in the desired production level in a responsive manner becomes very critical in this case Ė speed matters. Like everything else in life, operations of ASUs are subject to many constraints, so one may not be able to change things as freely as desired. For example, over-purifying products costs money, but plants still adopt this practice to keep a buffer zone so that despite unexpected upsets, the plant is able to supply products that meet the quality requirements. Failure to meet the quality specifications results in product that canít be sold and inconvenienced customers. Other constraint violations leading to equipment damage could be more catastrophic: plant shut-down, equipment replacement, etc. Hence, the policy maker wants to push the plant to its limits as close as possible, for example, by smaller safety margins and faster transitions, but not to the point that could lead to negative outcomes. It is a lot like the classic game on the TV show The Price Is Right, where players try to guess the retail price of some everyday product, and the winning guess is the closest price without going over. Go just one dollar over, and you lose. Hence, to be sustainable in a competitive environment, one also needs to understand where the limitations lie before taking action.
This why MACC researchers teamed up with the chemical company Praxair to develop a systematic approach to assess the true process limitations and assist the transition of the plant from one operation point to another. It is hard enough to transition a chemical process from one state to another without breaking something, and it is even harder to complete the transition rapidly while producing products of the required quality. Due to the time-consuming model development process and challenges in solving dynamic optimization studies with a detailed model, previous studies on operation and design tend to avoid the use of detailed dynamic models. The necessity of considering process dynamics and the effort in deriving such a model were not deemed to be fully justified. The full potential of using optimization studies to identify true plant bottlenecks (i.e. during normal operation and transition) had not been explored.
The research of Yanan Cao and Prof. Chris Swartz at McMaster, in collaboration with Praxair, offers meaningful insight to the problem. They developed what is called a two-tier optimization framework which can be used to factor in the dynamics and find better design and operation practice. Their optimization framework is essentially an advanced computer algorithm that uses a detailed model of the chemical plant that considers not just how the plant works while it is running normally (tier one), but what is going on in the plant while it is transitioning between operation points (tier two). The computer algorithm then provides information to suggest ways that the existing plant can be improved with some modifications that may not be obvious to see/justify using standard engineering methods. Process dynamics in the second tier had seldom been considered before, in part because the resources required and challenges faced were significant. However, the team showed that it is essential to consider the actual process dynamics through the use of dynamic models: certain constraints only become true limiting factors while the plant is in the transition mode. With this new information, engineers will be able to design chemical plants that operate with a much smaller safety margin without increasing the risk of violating operational constraints, saving lots of money. The team is now working the next step: modifying the algorithm to determine the optimal design directly by considering investment as well as operating costs.
The results of their study will be published in the September issue of the Journal of Process Control. Journal subscribers can access the article at the link below:
Abhinav  Garg, Prashant Mhaskar. Model predictive control of uni-axial rotational molding process [more]
Nor Farida Harun. Fuel Composition Transients in Solid Oxide Fuel Cell Gas Turbine Hybrid Systems for Polygeneration Applications [more]
Manufacturers are making decisions... [more]
Prashant Mhaskar, Abhinav  Garg, Brandon Corbett. Modeling and Control of Batch Processes: Theory and Applications [more]