Temperature control is a prerequisite for essentially every chemical reaction in which people are interested. Temperature affects the rate of reaction and often the completeness of the reaction. The human body incorporates a biological temperature control system to maintain a narrow range of body temperature. Processes designed to produce various materials also require temperature control. The engineer has a choice between an analog and a digital temperature controller.
Some analog home thermostats consist of a copper strip spiral. As the strip expands with heat, the spiral expands, moving a mechanical lever. The furnace or air conditioner responds accordingly. Analog controllers only react to the current environment.
The microprocessor in a digital temperature controller receives numeric input from the environment and manipulates it to enable a greater degree of control. If a system heats up quickly, the analog system will only react when the controller reaches its desired temperature, called the setpoint (SP). The source of heat may be turned off, yet the system will overshoot the SP because it is absorbing energy from the warm radiating surfaces surrounding the system. A digital temperature controller calculates the rate at which the temperature is rising and triggers the appliance to respond before the SP is reached. The controller used past data to predict and change the future results.
There are many algorithms or calculation schemes that a digital temperature controller might employ. One of the most common is the proportional-integral-derivative or PID controller. It uses three separate calculations to maintain a constant temperature.
The error (e) is the difference between the actual temperature (T) and the setpoint temperature (SP). The proportional calculation changes an input stream to a process based on the magnitude of E. An E of 2 would require an input of energy twice that of an E of 1.
The proportional control keeps the system from overshooting the SP, but the response may be sluggish. The integral method anticipates that future data trends will endure. In the above example, if T increases by an E of 2 and then an E of 4, the system might anticipate that the next E will be 8, so instead of doubling the response, it might triple the response and not wait for the next measurement.
A proportional and integral (PI) controller may oscillate around the SP, bouncing between too warm and too cool. A derivative control method will dampen the oscillation. The rate of change of E is used in the calculation.
The PID controller uses a weighted average of the three calculations to determine what action should be taken at any moment. This digital temperature controller is the most common and effective, as it uses current, historical, and anticipated data. Other control schemes require information about the nature of the system. Such knowledge boosts the ability of the controller to anticipate the future response of the system.