(63b) Cleaner Chemical Industry Through HEN Advanced Retrofit Approach

The economics of industrial production, the limitations of global energy supply, and the realities of environmental conservation are an enduring concern for all industries. Wherever you turn, there’s another entreaty to save energy, reduce carbon emissions and protect the environment for posterity.

Industrial plant’s heat exchangers network (HEN) shall always be ready to face a high level of operational changes along its lifetime. These changes are short-term, such as process disturbances; uncertainty in feed stock conditions and product demand, and long-term such as the need to process more raw materials that warrants the debottlenecking of the facility including the HEN to increase its capacity. Nowadays, and since late seventies of the last century, another important frequent change is happening due to the continuous escalation in energy prices in a rate that is higher than plant equipment cost, also warrants continual modification of the facility’s HEN to enhance the plant energy efficiency along its lifetime (sometimes reaching up to 30 years). In such cases, the HEN retrofit objective/task is to produce a practically implementable cost effective HEN design modification that satisfies the new process objective and its new operating constraints. There are many possible modifications for an existing HEN to retrofit the original design to the new objective. It can include a combination of all possible process operating and design conditions modifications, existing HEN topological/structural modifications, and existing HEN unit design modifications and parametric modifications (such as heat transfer enhancement to improve U) as well. In huge industrial facilities such problem can be solved using pinch technology or MINLP , where we have an objective to minimize total cost of the existing HEN modifications and the rest of the plant’s process units design and operation modifications minus the realized energy cost saving. There are some known drawbacks in the existing methods. In one hand the most widely used to date pinch method and its modifications is non-systematic and involves heuristics that produce solutions whose quality depends entirely on the designer experience in applying the heuristics. For instance, determining the location of the pinch point for an assumed best global DT_min, (which global DT_min should be used) is a little arbitrary. As the value of assumed best global DT_min changes, the location of the pinch point changes and therefore the assessment of which unit to transfer heat across the pinch also changes. The same also applies upon process changes that result in same output. The retrofit is a one package solution that has no consideration for yearly step-by-step modifications and practicality for future retrofit projects. The mathematical programming-based methods on the other hand have the drawback of providing little or almost no scope for user interaction. Nowadays, it is possible to solve larger problems than before but still not on the scale needed in many industrial mega sites. The problem of the HEN retrofit is NP-hard, and including all possible process design modifications, different types and configurations of heat exchanger units and so on in the superstructure is impractical and will make the problem harder, intractable and impossible to solve. In addition, the MINLP model cost objective function needs comprehensive data base for the model solution to be realistic with information/data that is normally not available during conceptual phase of the retrofit project, especially in de-centralized engineering environment where projects cost control depends entirely on rigorous cost calculation that always comes after the retrofit project basic engineering and is carried out at different departments in most of the companies. Besides neither practical nor logical to assume with high fidelity all possible retrofit solutions that can be implemented for a given plot plan of a facility and rigorously develop cost correlation to all of it before running the mathematical program especially when there is always more than one option to carry out a designated modification such as the increase of the network surface area. Neither one of these methods (pinch technology, network pinch and mathematical programming) can rigorously assess the existing network; rigorously target for the solution before the go ahead of the retrofit project for practically attainable waste heat recovery and/or minimum number of units that need to be added to reach desired energy consumption targets or part of it, and to systematically renders solutions that can be implemented on phases with a guarantee that the solution implemented today in phase one will not become an obstacle, or need to be demolished, for a tomorrow retrofit solution to be implemented at later time in the future on phases two, three and so on along the plant’s lifetime. The paper exhibits new elements, in the representation and assessment of the existing HEN studied for retrofit and in the systematic finding of the retrofit solutions that is evolutionary in implementation and revolutionary in consideration. It presents the concept of plant life-time retrofit-ability that addresses the identification of any possible contradiction between now and future retrofits to render opportunities of continuing retrofit projects to achieve ultimate target. The method is illustrated with two industrial case studies for NGL and oil refining plants.