Most chemical processes are networks of different pieces of equipment. Usually, even the best pieces of equipment will give a poor overall process if linked up inappropriately in the network. This paper describes principles and procedures for better process network design. Development of the procedures began in 1972. In the years since, industrial applications have led to significant improvements in even the most modern processes. The paper begins with a fresh look at thermodynamic Second Law analysis. This classical analysis highlights inefficient parts of complex systems, drawing the engineer’s attention to excessive losses of potential. Unfortunately, the analysis is both difficult to produce and difficult to interpret. To tackle the first problem, the paper describes how Second Law information can be obtained from conventional heat and mass balances. There is no need for additional data. To tackle the second problem, the paper introduces a general distinction between ‘avoidable’ and ‘inevitable’ inefficiencies. This makes an interpretation of the analysis practically more meaningful. Next, the paper describes thermodynamic procedures and principles for specialized sub-tasks in process design. Emphasis is placed on heat recovery networks. Here, the problem is to recover as much heat as is economically justified within a process before externally supplied heat is used. The concept of ‘inevitable’ inefficiencies leads to techniques for the prediction of the ‘inevitable’ amount of external heating. This amount is called the energy target. The target either stimulates the engineer into achieving it or gives him confidence that his design is optimal. The paper continues by describing the concept of the heat recovery ‘pinch’. The pinch leads to the design of, first, heat exchanger networks, which achieve the energy targets, and, second, overall processes, which keep the targets low. Two common threads in all these procedures are the attempt to keep the engineer involved (they do
not
constitute ‘automatic’ design) and the attempt to make best practical use of inefficiencies that are ‘inevitable’ anyway. Owing to these features, the procedures usually help the engineer to find processes that are elegant in a general sense. Many designs found in practice were not only energy efficient but easily operated and maintained, safe, had relatively simple network structures and, most surprisingly, were cheap to build as well as cheap to run.
