Fuzzy control technology aims at a computer like a man and will enable a computer to perform analog thinking instead of conventional digital thinking. Fuji Electric quickly applied this technology to control systems with satisfactory results. This paper describes the trends of fuzzy control from the angle of comparison with conventional technology and popularization, Fuji Electric's present status and attitude, and the research and development of fuzzy theory having come in a new stage with the start of the national project. It also refers to the forecast of the technology and its problems.

As automation technology, fuzzy theory is compared with AI and control theory. The features are described from the point of efficiency and reliability in the system development.
Fuzzy theory is regarded as the intermediate between AI and control theory. It is realized by more compact instrumentations than AI and it is more easily understood than control theory.
The merits such as the ability of user-friendly expression, the generation of continuous outputs during rule processing and the possibility of the easy design of multi-variable or nonlinear systems are appreciated. The application is promoted for the automation of operator's manipulation.

Fuji Electric early started to make fuzzy control practicable. The paper introduces the features and procedures for practical use of fuzzy theory and its application, that is, fuzzy control, in our experience.
The inquiries about FRUITAX shows that industries have a keen interest in fuzzy control and proceed to make it practicable. The status of its practical use is explained citing the features of fuzzy control, purposes, application methods, and results in the actual applications in this special number.

This paper describes Fuji Electric's history in connection with fuzzy control and the whole organization of the general-purpose fuzzy control system FRUITAX, followed by an outline and the functions, users, and applications of the off-line development supporting system and on-line control system. Further, it details the general-purpose fuzzy control system FRUITAX-L which operates in conjunction with our latest AI tool AIMAX-C, based on FORTRAN/KR.

MICREX-F250 is a new-generation general-purpose programmable controller with a function of fuzzy control. The F250 realizes the function by means of a function module (FM) and makes easy to prepare programs combined with sequential control and PID control. The fuzzy FM can execute reasoning of 32 fuzzy rules in 20 ms. The setting, testing, and tuning of control rules is performed with the supporting tool FRUITAX. The FRUITAX, as a function of the ES50 which totally supports the software engineering of F250, acts together with the other supporting functions.

It is necessary to check or adjust the control rules and membership functions used for fuzzy reasoning when they are drawn up. To make these checking and adjustment easy, Fuji Electric has recently developed a system development tool for personal computers (a simulation program of fuzzy reasoning, an optimum parameter search program of fuzzy variables of premise part, and a three-dimensional display program) on the basis of a supporting tool developed with a large computer. This paper introduces an outline of the tool and its use.

Since computers became cheaper, computer-assisted instruction (CAI) has been noticed as a new method of instruction.
This paper describes the current status, methods of using, features of CAI, and comparisons with other instruction methods or books, and introduces "Fuzzy CAI" newly developed and put on the market. This tool consists of CAI course simulation, and a manual and provides information on fuzzy theory from introduction to actual systems.

Fuji Electric has developed a fuzzy control system for cohesion addition process in a water purification plant and realized automated operation. The system calculates cohesion adding rates and corrects them by fuzzy reasoning. Its effectiveness was verified by simulation and field tests; then, it was putin actual service. The results were as follows:
(1) Reduction in operater's duty
(2) Safe control corresponding to raw water maddiness
(3) Stable control against external disturbance
(4) Reduction in the quantity of chemicals used

This paper describes a method of fuzzy control application to the pump and gate of a combined sewer system and an example of application. In Yushima pumping station in Tokyo, skilled operators controlled them to meet a sudden rain water flow. To automate the control, Fuji Electric developed a fuzzy adaptive control system and put it to practical use. The result of application showed that coordination among multivariables was realized only by processing the control know-how for slight rain and for heavy rain using a few control rules and membership functions and the control became mare user-friendly.

This paper introduces a new control system applied to burning control of cement kilns which coordinates fuzzy control and optimal control. EIC-unified digital control equipment has been introduced into the control system and the kiln is controlled by a process computer A-60. Fuzzy control carries out standard value change corresponding to process state change and recovery action to a stable state from an abnormal state that cannot be met by optimal control. The application of this control method to an actual process has resulted in a reduction of about 50% in process fluctuation and a big reduction in operator's duty.

Many external-circulating vertical roller mills has been used for raw material grinding in recent cement plants considering labor saving and stability. And optimal operation by multivariable control is required for higher efficiency.
Mitsubishi Mining & Cement Co., Ltd. has completed an optimal operation system for No.5 vertical roller mill in Higashidani Plant, Kyushu Area, using Fuji Electric's general-purpose fuzzy control system.
This paper describes an outline of the system and the result of control.

In conventional supply-and-demand control in steel works, the efficient operation of converter gas generated by converter blowing depended on the experience and operation know-how of the operators.
Fuji Electric has developed an expert system by combining the accumulated know-how of the skilled operators with production rules and fuzzy rules and completed the automatic control of the converter gas supply process in Keihin Steel Works, Nippon Kohan K,K. This paper gives an outline of the system.

It is difficult to make mathematical models for cultivation process because its state and behavior is complicated and online measurable factors are often restricted to a small number. Therefore process operation has depended on the intuity and experience of operators in Kanegafuchi Chemical Industry Co., Ltd. To introduce FA into such process, we applied fuzzy theory that enables information processing of human knowledge, and has developed a fuzzy control system composed of recognition of a process state and hierarchical reasoning to determine process control amount against the state. This paper introduces mainly the method of designing fuzzy control and others.

Severer legal controls have recently been imposed on drink vendors and the sterilization of drinking water in cup-system vendors has become an indispensable function.
To cope with the situation, Fuji Electric has quickly developed a water sterilizer by chlorination and a chlorine generator. To vend sanitary, nice drinks without chlorine smell, it is an important technology to keep the concentration of chlorine constant. The application of fuzzy control technology has realized a cheap, high-performance sterilizer which can meet the quality change of drinking water, local or seasonal, to give a stable concentration of chlorine.

As a part of investigation for the upgrading of operations in nuclear power plants, Fuji Electric has developed a fuzzy control simulation program to automate control-rod operation at a reactor shutdown. A compact program was designed to include the operation know-how of the field operators by dividing control into a reasoning portion for each stage shift and a reasoning portion for control-rod operating amount to meet the different control targets in each stage of a reactor shutdown. Further, simulation with a model of an existing reactor showed that correspondence to the nonlinear characteristics of fuzzy control functioned well and the present manual operation could be automated.