Intelligent Control in Industrial Applications

The number of industrial applications that use IC systems is rapidly increasing, where one can find IC systems in both large and small industrial applications. Another growing area of IC applications is developing household appliances, which are small but complex control systems. Many systems that use fuzzy logic or neural networks for control apply these techniques to solve problems that fall outside the domain of conventional feedback control, e. g., in the case of a washer machine it is easier to control the duty cycle by a FLC than a PID controller. When we view fuzzy or neural control as only a non-linear counterpart of conventional feedback control techniques, the possibilities of using IC are reduced. Thus, a narrow conceptual view of IC system application leads to designers not appreciating or recognizing new areas of opportunities. If you use only the IC systems as a conventional controller the difference is quite small. For instance, using a FLC as a PID4 1 Intelligent Control for LabVIEW controller with the error and the change in error as inputs, the fuzzy controllers look similar to the conventional PID controller except that fuzzy control provides a non-linear control law. Another case is the use of a neural network applied to the set-point regulation problem, usually by replacing a conventional controller’s law and/or plant model with an artificial neural network. However, if we apply IC systems to standard and non-standard techniques we could handle high-level control systems. Let us imagine the control system of the train developed in Sendai, Japan by Hitachi. Here fuzzy logic was used to select the notch position that will best satisfy the multiple, often-conflicting objectives. An additional example is that many Japanese companies such as Matsushita, Sanyo, Hitachi, and Sharp, have incorporated neural network technology into a product known as the kerosene fan heater. In Sanyo’s heater, a neural network learns the daily usage pattern of the consumer, thus allowing the heater to automatically start to preheat in advance . For many industrial applications one could complement the conventional controllers by an intelligent controller generating a new one, rather than using IC alone. The industrial challenge is focused on developing control systems that are capable of adapting to rapidly changing environments and on improving their performance based on their experience. In other words, modern control systems are being developed that are capable of learning to improve their performance over time (to learn) much like humans