17.2 Process Engineering in Practice

There is a limit to the amount of material that can be covered in a single textbook with reasonable length and depth of coverage. In this text we have chosen topics that we consider to form a firm basis for further study in automation and control. Hopefully you now have a firm foundation for further study on automation and control-related topics.

Topics Not Covered

Important topics not addressed in this textbook to a significant extent include the following:

• Statistical process control (SPC): SPC is used by process engineers and operators to determine if a process is producing a product that is meeting specified quality targets. SPC techniques attempt to differentiate between normal statistical deviations in product quality and deviations due to operating problems (these may be control related)

• Advanced digital control techniques: An introduction to the most important advanced control technique, model predictive control (MPC), was presented in Chapter 16. There are a large number of important topics in MPC that were not presented. These include the use of different types of models (linear step response models were covered), different model updating techniques (output additive disturbance was covered), and different objective functions (least squares over a finite prediction horizon was covered). Nonlinear models and adaptive linear models can also be used.

• Inferential control: Very often the most important process variable (product quality or property) cannot be measured on-line. Other measurements must be used to "infer" the quality variable; one example is to use tray temperature in distillation to infer the distillate purity.

• Noise filtering: In practice, measurements are corrupted by noise. Often "filters" are used to "average" recent measurements to obtain a better estimate of the actual value.

• Specific control system hardware and software: We have chosen not to cover specific hardware or software for automation projects in detail. Although important in the actual implementation of a control project, we feel that the role of the process engineer is to specify control objectives, determine the proper measured variables, etc. Instrumentation specialists and other project engineers will be involved with many of the detailed specifications necessary to complete an automation project.

Art and Philosophy of Process Engineering

The majority of students majoring in chemical engineering become process engineers in their first industrial position. A typical responsibility is to provide technical support for an entire operating unit (termed a "plant" in Chapter 15). This includes troubleshooting (determining why a product is suddenly off-specification, for example), and capital project justification (designing new equipment to reduce energy consumption, minimize waste, maximize profitability, etc.), among numerous other things. One of the first things that you will do is review process flow diagrams (to understand the basic process flows) and process and instrumentation diagrams (P&IDs, to understand the basic control strategy). You will spend a lot of time tracing pipes through piperacks, to understand physically the layout of the process unit. You will also spend a lot of time talking to process operators, shift foremen, and the unit manager to better understand basic operating procedures. An important part of process operation, naturally, is the automation and control strategy. Each process plant will have an engineer whose primary responsibility is the basic maintenance of the computer control system. Most often a distributed control system (DCS) is used. As a process engineer for a particular unit, you will be expected to achieve some level of knowledge of the control system.

The point of the previous paragraph is that you, as a process engineer, do not operate in isolation. There are many people involved in the operation of a chemical process, no matter which industry you are working in. There is an entire history to the operation of a particular process, and you will not be expected to be an expert overnight. You need to be willing, however, to listen and learn from the experiences of others. At the same time, you need to realize that not everything you will be told will be technically correct and that some information is biased (due to management and labor relation issues, etc.). You will need to achieve a delicate balance between believing everything you are told and being so skeptical that you believe nothing.

So, will you be deriving equations, manipulating block diagrams, and using the Routh stability criterion on a daily basis? In all probability, no. But hopefully this textbook has given you some basic knowledge of dynamics and control that you can use in a broad, qualitative sense. Without explicitly trying, you will probably begin to think in terms of inputs, outputs, and objectives. You will begin to develop a feel for the order of magnitude of an effect of an input change on an output. You will naturally perform "back of the envelope" calculations based on material and energy balances to understand these relationships (for example, if the steam flow to an exchanger changes by a certain amount, how much do you expect the process fluid temperature to change?).

Eventually, as you learn more about the process, you may conceive ideas to improve the control of the process. There is no better way to develop better control strategies than to achieve a deeper knowledge of the process. The best "loop-tuner" in the world will not achieve good control loop performance if a bad input-output pairing is made to begin with.

Rules of Process Operations

We have all heard of Murphy's law—"if something can go wrong, it will go wrong." Similarly with process operations, it is not a question of if a control element will fail, but rather, when will the control element fail? Safe processes are designed with the knowledge that most moving equipment will eventually fail. Also, sensors will fail or need to be recalibrated. As we noted in Chapter 1, all control valves are designed with a bypass line. When the control valve fails, it can be taken out of service while the flow is regulated manually through the bypass valve. Similarly, virtually all pumps have spare pumps placed "in parallel" with them. When one pump fails (detected as a loss of pressure), the other pump automatically "kicks on" before the process is affected. This also allows preventive maintenance to be performed on all pieces of moving equipment.

It should also be noted that the probability of a failure or technical problem occurring between 8 am and 5 pm on Monday–Friday is roughly 27% (you do the math), so 73% of failures and technical problems occur outside the normal working hours of a process engineer. If the technical problem is critical to continued process operation (remember, many of these operating units are producing several $ million/day of product), and you are the process engineer "on-call," you will likely be called in to help solve the problem. This is usually exciting (sometimes you may wish for less excitement in your life), sometimes with almost an "emergency room" sense of urgency as you and many others try and troubleshoot/solve the problem.