Doctoral Dissertations

Date of Award

8-1995

Degree Type

Dissertation

Degree Name

Doctor of Philosophy

Major

Electrical Engineering

Major Professor

B. K. Bose

Committee Members

W. L. Green, F. W. Symonds, B. R. Upadhyaya

Abstract

Traditionally, AC drive systems have been designed by the expert engineer using paper, pencil and calculator, and then tested through laboratory prototype. This procedure is expensive and time-consuming. Besides, the design is not necessarily optimum. An expert system aided automated design of AC drive systems is presented here. The comprehensive design package for AC drive systems first configures the AC drive system based on the user's application specifications, and designs the nameplate specifications of the AC drive system. Next, it designs and optimizes the power converter system and then designs and fme-tunes the controllers for the AC drive system. Gnce the complete design is over, it downloads the optimized DSP controller code to a target experimental drive system, for its validation. The expert system for the AC drive configuration selection and motor design was developed embedding the expert's know-how on matching the application specs with the characteristics of AC drive configurations available. The voltage-fed type of configuration was selected for the 1-200 kW power range of the AC drive system. A wide range of application classes, such as pumps, fans, winders, crushers, conveyors, machine tool drives, robotic arms, elevators, electric vehicles and traction, were embedded for the design and configuration of AC drive systems, and the expert system knowledge base was made flexible enough to tackle even the custom-designed application specified by the user. The design criteria for the motor nameplate specs were incorporated in the expert system in terms of the rule base. Intelligent defaults were implemented to meet any unresolved or conflicting choices. The user interface of the expert system was in terms of the natural language dialog and menu-driven consultation. The expert s\ stem based automated design of the power converter system involved development of the rule-base for designing the ratings of the power devices, heat sink, snubbers and filter capacitors. Rules were also included to permit expert system's access to databases and graphics files. The databases were developed to store the data sheets of power semiconductor devices, such as the diodes and transistors (BJT/IGBTs), which would store their ratings and switching times. Graphical data were implemented for the devices using the best-fitting curves obtained through TableCurve software. For the power converter optimization, the models of the power converter and machine, as well as the half-bridge for snubber performance optimization were developed in SIMNON program. In order for the expert system to communicate with SIMNON simulation, C language based interface routines achieving the data and control format conversion were developed. The expert system rulebase was developed to control the simulation of the power converter system in a manner that would obtain optimum values of snubber components. The results showed successful applications of the expert system for power converter optimization. The expert system was then developed to perform the initial design of the control gains of the AC drive system. This was demonstrated with the example of indirect vector controlled induction motor drive with synchronous current control. The next task was to achieve a hybrid simulation software for the AC drive system. The power converter and machine simulation model was implemented in SIMNON, whereas the controllers were implemented in DSP 320C25 assembly and C language, being executed in the simulator environment. Then, the communication protocol was developed between the expert system and DSP simulator, in terms of C-language based data format and control conversion routines. Next, expert system rule-base was developed which would control the simulation of the AC drive controllers so as to optimize the controllers for the AC drive system. The user interface in this case involved graphical display of intermediate results and natural language dialogues for status information. The hybrid simulation software was developed flexibly, so as to be usable in stand alone operation. Next, the complete software was tested for its tuning performance in all the modes of tuning. The results obtained demonstrated the expert system's ability and usefulness to perform off-line optimization of the AC drive controllers. Moreover, the optimized DSP-based controller code was now available, and was downloaded to the experimental drive system. The next task involved the experimental validation of the expert system based control timing of the indirect vector controlled induction motor drive system. The hardware and software for this IGBT inverter fed 5 hp induction servo drive were integrated in a DSP 320C25 based system, and the tests were performed on the AC drive, to obtain the results of the tuning. These were found to match well with the tuned gain values obtained through the expert system based control tuning.

Download
Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Share

COinS