A fuzzy-power factor correction (abstract only)

ABSTRACT

Power-factor correction usually means the practice of generating reactive power as close as possible to the load which requires it, rather than supplying it from a remote power station. Most industrial loads have lagging power-factors i.e. they absorb reactive power. The load current, therefore, tends to be larger than is required to supply the real power alone and the excess load current represents a waste to the consumer. Supply tariffs to industrial customers almost always penalize low-power loads; resulting in the extensive development of power-factor correction systems for industrial plants [1].

This paper represents application of fuzzy algorithms [2] [3] to the power-factor correction of a synchronous motor. The synchronous motor is a constant speed electrical machine capable of absorbing or generating reactive power depending on the output mechanical power and the field current [4]. The objective of the fuzzy control is to operate the synchronous motor at unity power-factor, neither absorbing nor generating reactive power, regardless of the mechanical power delivered.

The model has two major components, the fuzzy controller and the synchronous motor, that interact with each other. As shown in Figure 1, the controller comprises of the control algorithm, the input interface (which maps the controller inputs into their Universe of Discourse), and output interface (which defuzzifies the output of the controller. The controller has two inputs: change in the field current (△I_f) and change in armature current (△I_a). The single output is the change in thyristor's firing angle (△α) which controls the field voltage of the synchronous motor. To achieve the desired power-factor, the controller monitors the inputs at regular intervals and accordingly issues commands to the synchronous motor, which would result in new inputs to the controller. The decision rules were generated in linguistic rather than mathematical terms according to the experimentally-obtained “V” curves of the synchronous motor [4]. The control algorithm is based on a set of 14 rules that stipulates the nature and extent of control changes for all combination of inputs. The values, the input and output variables could take, were limited to the fuzzy subsets: positive big (PB), positive medium (PM), positive small (PS), positive zero (PZ), negative zero (NZ), negative small (NS), negative medium (NM), negative big (NB). Each rule is represented as follows:

IF (△I_f is NS) AND (△I_a IS NM OR NB)

THEN △α IS PM

For implementation, the Universe of Discourse is quantized to 14 levels. The outputs, given by the decision table, are computed using the MAX-MIN composition [3] [5]. Defuzzification is accomplished by using either the centre of area (COA) algorithm or mean of maxima (MOM) algorithm [5]. Initial results show that the firing angle of the thyristors adjusts the field current so that the desired unity power-factor is achieved.

References

1. 1.T.J.E. Miller, "Reactive Power Control in Electrical Systems," New York: John Wiley & Sons, Inc., 1982, pp. 1-19.
2. 2.P.J. King and E.H. Mandani," T~e Application of Fuzzy Control Systems to Industrial Control," Automatica, Vol. 13, 1977, PP. 235-242.
3. 3.L.A. Zadeh, "Outline of a New Approach to the Analysis of Complex Systems and Decision Processes," IEEE Trans. on Sys~, Man and Cyber., SMC-3, 1973, PP. 28-44.
4. 4.G. McPherson, "An Introduction to Electrical Machines and Transformers," New York: John Wiley & Sons, INc., 1981, pp. 35-152.
5. 5.E.H. Mamdani, J.J. Ostergaard, and E. Lembessis, "Use of Fuzzy Logic for Implementing Rule-based Control of Industrial Processes," Advances in Fuzzy Sets Possibility Theory and Application edited by P.P. Wang and S.K~ Chang, 1983, New York : Plenum.

Index Terms

• Computing methodologies

  • Artificial intelligence

    • Knowledge representation and reasoning

      • Probabilistic reasoning

      • Vagueness and fuzzy logic

• Mathematics of computing

  • Discrete mathematics