The Study Of Power System For Mid-Frequency Quencher Based On Fuzzy Control

The quench processes of thermal treatments can be classified as middlefrequency induction heating, flame heating, electric contact heating, electrolyteheating and laser heating in accordance to heating sources on workpiecesurfaces, in which the quenching technique of middle frequency inductionheating is applied very widely. This technique utilizes the vortex resulted fromthe treated workpiece surface for heating and causes the rapid raise of workpiecesurface temperature for the purpose of surface heat treatments. The heatingmethod of middle frequency induction has precise heating depth and surfaceareas without any external hot source so heat loss is low and the workingenvironment is clean. The heated workpiece need not contact to the inductor toeasily accomplish the power denseness, which results in short heating time andeasily enforcing the process automatization. Therefore, the middle frequencyinduction heating is widely used in the quenching technique but the middlefrequency quench system is the representative non-linear inertia tache and hastime delay. Non-linear and delay system is widely used in industries, and controlquantity can not timely reflect on the controlled object due to its delay for thecontrol output, which results in that control performances of the overshoot,measuring time and stability get worse. With the constant progress of industrialautomatization technique, the demand for the quality of industry control isimproving, and intelligent control has provided an effective approach for thisproblem.Fuzzy logic is the basic way of intelligent control. It need not establish anaccurate mathematical model of the controlled object, has stronger robustnessand is not sensitive to the change of the controlled object parameter. As for thefuzzy control system, the good effect of fuzzy control is base on the completecontrol rules and the appropriate fuzzy process. But fuzzy control has defects,for example, simple fuzzy processing may result in the control precision fallingand the quality degradation. To improve the precision, fuzzy quantification levelnumber must be increased, then the rule complicated degree will also beimproved, the decision-speed will be slowed down, even the control can not beoperated. So how to automatically generate and adjust the subject function andfuzzy rules, based on control experience of operators and knowledge of experts,is the key question to design the good performance fuzzy controller. How to usethe fuzzy theory in control system to improve various performances of thewhole control system is one of the most attractive subjects in the field ofintelligent control at present. The fuzzy controller in this paper is a typical dual-input and single-outputstructure. The whole fuzzy system is composed of input-output languagevariable, subject function, fuzzy rules and illation, and anti-fuzzy theory. Thetemperature deviation, E, and deviation variability, Ec, are control input, and Uis the output to control the middle frequency heater. Fuzzy control rules andoperations are the core of the fuzzy controller, concluded from artificial controlexperience and based on control experience of operators and knowledge ofexperts, and fuzzy language is applied to improve the real-time control speed.Fuzzy controller adopts discrete discussion region, and after the fuzzy assembleand subject function is off-line calculated on some illation rules, a control formis made, in which we can look up during our practical control. In real time control work, as long as this form is saved in correspondingEPROM and the quantification level of the actual error, e, and the error variety,â–³e, is calculated, we can look up this form, then obtain the output controlvariable of the system. The basic fuzzy controller can be easily operated by thesoftware, of which control rules have no any restriction. We can makediscussion and modification together with experienced operators andqualitatively adopt all kinds of excellent control ideas. However, the ability of fuzzy to control integrated quantitativeinformation is not very good, a piece of desirable FUZZY control form can beput into practice after careful amendment over and over again and thequantification factor and the ratio factor also affect the capability of the wholesystem, of which all the above will result in following defects existing in simplelookup form of fuzzy controller. (1) Self-adaptation ability is limited. Because quantification factor and ratiofactor can not be modified after the simple fuzzy lookup form is amended, whenthe object parameter changes, it can not properly adjust its own control rule andthe good performance can not be brought into operation. (2) If the structure of the lookup form is not rational, or quantificationfactor and ratio factor is not appropriate, it will lead to the system oscillation,especially when the system error approaches zero, high frequency oscillationwill be generally produced. (3) Precision is not high. This is because the level of fuzzy control ruleform is limited. The system precision can be improved by increasing the numberof quantification level, but this will result that the lookup form is very bulky andcovers much space in the EPROM to increase the operation time. This loopsystem requires higher precision, and, apparently, this method that only adoptsbasic fuzzy controller is far from requirements of the system. Aiming at thisquestion, if we adopt adjustment factor fuzzy control, the system performancewill be greatly improved. PID control is one of the traditional correction strategies and widelyapplied in industrial process control for its simple operation, good robustnessand high reliability. Today, most control systems still adopt PID structure. Butthe process and the controlled object of actual industry manufacture arenon-linear and uncertain, so it is very difficult in establishing precisemathematic models and the conventional PID controller is difficult to obtainideal control purpose. Because of the effect of complicated parameteradjustment method, the parameter of conventional PID controller has badadjustment, poor performance and worse adaptability to operationalcircumstance in practical manufacture processes. When fuzzy control is appliedin the non-linear system, various features of fuzzy control system are morepreferable than common PID control system, such as easy accomplishment,good control effect and rapid respond speeds, and it also has the fine ability to fitthe change of controlled object parameters. The commutating section in the main loop of the middle frequencyinduction quenching electric source applies a three-phase bridge rectifyingcircuit, which has features of simple structure and high power factors. Thecontravariant section uses a series resonant all-bridge DC/AC contravariantcircuit, acts IGBT as the main switch element and makes the most ofcharacteristics of loading series resonant to cause IGBT to be in states ofcut-down at zero current and open-on at zero voltage and to reduce the switchloss of IGBT. The dc-to-ac converter has the output power factor approaching to1 and high heating efficiency, which will have a good applied prospect.