The chemical industry has some of the most important enabling technologies for modern society. Their products cover all sectors ranging from plastics, pharmaceuticals, energy and fuels, to food, agriculture and the environment. Practically all aspects of modern life are driven by products generated from the chemical industry. At the heart of the chemical industry are chemical reactors (which are vessels wherein chemical reactions take place). Chemical reactors are the key pieces in converting (less useful) raw materials into end user products.
How well these chemical reactors work is vital in determining how efficient the processes are that use them. In most cases, lower efficiency results in larger carbon dioxide emissions, which is both harmful and wasteful to society*. Given that the chemical industry (directly) consumes roughly 5% of the world's energy, the efficient design and opertation of chemical reactors is clearly an important goal for modern industrial processes, and the wellbeing of society as a whole.
Chemical reactions very rarely produce pure products. This inefficiecy results in the incorporation of separation and purification steps in a chemical process. But it is the purification of the desired products from the by-products that is the largest consumer of energy. If better reactors could be designed (that produced purer products that do not need as much separation for instance), this could have a profound impact on the energy efficiency of these plants.
Choosing the best set of reactor(s) for a given system of reactions and kinetics is often a nontrivial task. Let us consider some of the questions and challenges that might be encountered.
The problem of designing and optimizing a collection of reactors in an optimal manner—for a desired duty—is called the reactor network synthesis problem.
Consider a reactor configuration that has already been designed and optimised, such as the configuration given below:
How do we know that it is the best? If there are no other designs to compare to the current design, then there is little certainty that the current design is the best. Other reactor configurations (that we have not yet thought of) might exist that perform better.
The problem is that it is always possible to devise new designs that might do better. And we will not know if these new designs are better until each design is compared to every other design. This problem can be solved if an independent and common measure is available through which all current designs can be assessed against. In essence, we need to establish appropriate performance targets.
The AR is the collection of all possible outputs for all possible reactor configurations—even the ones we cannot imagine.
Every point in or on the AR is an output of a reactor combination. Different parts of the AR boundary correspond to distinct reactor combinations. Using the AR, we can compare the performance of different reactors in relation to each other as well as in relation to the boundary of the AR, which represents the absolute best performance of the reactor configuration.
AR theory also assists the designer in understanding how to interpret, select and ultimately optimise a network of chemical reactors, so that the flowsheet is appropriate for a desired duty—different duties may have different optimal reactor structures, and hence it is more sensible to choose the correct reactor combination for the duty at hand. When the duty changes, so should the structure.
Hence, AR theory is a method for helping to address the reactor network synthesis problem, although it also assists in understanding and setting appropriate performance targets.
Using AR theory, we can begin to understand two important problems:
The absolute limits of the reactor system, which is useful for performance targeting of new and existing designs of chemical reactors.
The physical equipment required to achieve these limits, so that we know how to approach and optimize our designs. This problem is also known as the Reactor Network Synthesis (RNS) problem.
Understanding these two aspects helps us to design efficient, robust and simpler reactor networks
Designing a network of reactors is challenging because there are often a large number of variables and design choices that must be considered before a suitable design can be established. Furthermore, even if an existing design is available, there always exists an element of uncertainty, for there may be alternate designs (which we have not thought of) that perform better. AR theory is a method for helping to address this problem. With AR theory, we can begin to systematically determine all the potentially feasible products for a given system, and in the process learn to distinguish between superior designs to inferior ones. This process will also ensure that some (potentially superior) possibilities are not left unconsidered.
Continue on to the next section , where the basics of AR theory are introduced.