Principles of Automatic Control

In industrial furnaces, if high-quality products are to be produced, several physical and chemical variables must be kept essentially constant or controlled according to a preset program. These variables include temperature, pressure, gas composition, and, indirectly, flow rate. The basic principle of automatic control is the same for all of these variables, so it is useful to briefly explain the general principles before discussing each type of control. The variable to be controlled, such as temperature or pressure, is called the controlled variable. A variable that changes by itself and can influence the controlled variable is called an independent variable. In some processes, several independent variables may exist. Automatic control works as follows: a force is derived directly or indirectly from the controlled variable. This force is balanced by gravity, mechanical spring force, gas pressure, electromagnetic force, or another corresponding force. The element that produces the balancing force may take many different forms. When the independent variable changes, the value of the controlled variable also changes. This breaks the balance between the derived force and the opposing force. The small difference between these two forces causes a mechanism to move, attempting to return the controlled variable to its original value. This original value is called the set value or control point. A simple control system can be illustrated by water-level control. The water level in a tank is the controlled variable. At the set value, the upward buoyant force of the water on the float is exactly balanced by the weight of the float. At this point, the inflow and outflow of water are equal. If the drain valve, which acts as an independent variable, is opened wider, the water level drops below the set value. The upward force on the float decreases, so the float moves downward. This increases the water inflow until the water level rises back near the set value. Two drain valves would therefore form two independent variables. The set value is usually adjustable, so the controlled variable can be changed as required. In the water-level example, the water level can be adjusted by slightly bending the float rod or by changing the length of the linkage between the float rod and the valve rod. Automatic control generally requires several units. The first unit is a measuring instrument, or primary element, which measures the controlled variable and its deviation from the set value. It also reflects the unbalanced difference between the opposing forces. The second unit is the controller, which converts this small force difference into movement of the actuating element, such as a valve or switch. In some control systems, the measuring instrument also acts as the controller. This type of system is called a self-operated control system. A pressure regulator is another example. Usually, the small force difference in the measuring instrument is not enough to overcome friction and other resistance in the actuating element, unless the deviation from the set value is very large. Therefore, a relay is often installed to convert the small force difference into a much larger force. A control device based on this principle is called a power-assisted control system. The force amplification may be mechanical, hydraulic, pneumatic, electrical, or electronic. Two-position control, such as on-off control or increase-decrease control, is the simplest form of power-assisted control. When the controlled variable deviates from the set value by a predetermined amount, the energy supplied to the process, such as fuel, fluid, or electrical energy, is either fully shut off, partially reduced, or fully opened, depending on the direction of the deviation. When the measuring instrument moves back and passes the set value, the control action reverses. As a result, the control process oscillates around the set value. If there were no friction and inertia in the instrument, the actuating element, such as a valve, would continuously open and close. If the opening and closing positions are set on opposite sides of the set value, each at a small distance from it, the switching frequency can be reduced. The distance between these two points is called the operating differential. This arrangement reduces the switching frequency, but it increases the fluctuation range of the controlled variable. If the movement speed of the actuating element is slowed, meaning the operating time from one limit position to the other is extended, the switching frequency can be reduced further. In this case, the actuating element “floats” around its correct position. However, if the actuating element moves too slowly, it may cause large fluctuations in the controlled variable. By changing from full-open/full-close control to increase-decrease control, the fluctuation range of the controlled variable can be reduced. In some special cases, this method requires manual control. The control methods discussed above cause the controlled variable to oscillate or fluctuate. In many processes, such oscillation is acceptable because precise regulation is not required. In other processes, however, accurate regulation is necessary, and continuous oscillation is undesirable. Properly designed and well-manufactured proportional control can eliminate oscillation. In proportional control, within a certain range, known as the proportional band or throttling range, the position of the actuating element is always proportional to the deviation between the controlled variable and the set value. The actuating element moves only when the indicator of the measuring instrument moves. Therefore, the measuring instrument must have almost no friction and very low inertia. For stable control without oscillation, and without integral control, each position of the actuating element must correspond to a different value of the controlled variable. From no load to full load, the value of the controlled variable decreases. This decrease is often called offset. The line representing this relationship is usually called the control characteristic curve. It does not necessarily have to be a straight line. This offset is necessary to prevent oscillation. Its value depends not only on the design of the controller, but also on the characteristics of the process it serves. The water-level control example is a self-operated proportional control system. When a large amount of water is required, that is, under full load, the water level is low so that the inlet valve remains wide open. When no water flows out of the tank, the water level is high so that the inlet valve remains closed. The term offset means that the controlled variable deviates from the set value, whether the independent variable is above or below its average value. In some processes, offset has little effect, while in others it can be serious. In those cases, corrective action is needed to compensate for the offset. This action is called reset. For example, if the load increases from the average value to a higher value and the controlled variable drops by a certain amount, this error can be eliminated by reset. Reset raises the control characteristic curve by the required correction amount. Reset, or offset correction, may be performed manually. Since most modern controllers are equipped with automatic offset correction devices, it is necessary to briefly understand this principle.