Several studies show a large potential for improvement of the energy efficiency in industry through process integration. So far, however, few of the identified measures have been implemented. One reason is that the profitability of making these investments is dependent on the values of parameters such as investment costs, operating availability of new technologies, and future energy prices and policy instruments – all of which are highly uncertain. Another reason to the low implementation rate is probably that it is difficult to understand how the potential for different process integration measures interplay in a long-term perspective.
While researches usually identify the most radical, visionary measures, industry legitimately decide on the safe investments with a short payback time but relatively small energy savings. For industry itself it would probably be more cost-effective to make the more long-term, strategic investments if these can be shown to be robust enough both technically and economically. It is therefore important to increase the knowledge of the consequences of decisions, the order in which decisions are made, and the implications for later decisions.
Decisions must be strategically right, robust enough (that is, profitable under different future scenarios) and well analyzed with respect to uncertain parameters. Traditionally, the effects of uncertainties have been studied through sensitivity analysis or scenario analysis. To improve the understanding of risks, opportunities, and robustness related to major decisions under uncertainty, it is more advantageous to use methods from stochastic programming. These methods can also be applied in process integration studies. With a method that is based on stochastic programming, the information to base decisions on can be substantially improved, and thereby new systems solutions can be implemented.
The aim of the project is to develop a method for optimization of process integration investments under uncertainty. The method should be based on existing methods and tools used in stochastic programming and process integration. Uncertain parameters to consider are, for example, future energy prices, policy instruments, and operating availability. Furthermore, the methodology should be possible to use for studies of the effects of lock-in situations, that is, how decisions made with a short-term perspective can make it impossible or substantially more expensive to make strategic and radical changes in the future. One part of the project is also to study how optimization can be made with respect to several, partly conflicting objectives, such as the net present value versus CO2 emissions reductions.
If all possible measures that are identified in a process integration study could be implemented, the energy saving would be substantial. That is, however, usually not the case. Some measures can be combined, others can only be combined under certain restrictions, and others not at all. A deep understanding of the process integration is demanded to identify and quantify these opportunities. The methodology developed in the project should make it possible to optimize these complex combinations of different process integration measures.
The overall aim of the project is to use the developed methodology to show how to identify and quantify the large potential for cost-effective energy savings that exist in an industrial energy system, the risks that are associated with the investments required to reach these energy savings, and the importance of making the right series of decisions in a long-term, strategic perspective.
The project is carried out within the Process Integration research programme, and is financed by the Swedish Energy Agency and the Södra Foundation for Research, Development and Education.