Self-Commissioning of AC Motor Drives

Abstract

In modern motion control and power conversion applications, the use of inverter-fed electrical machines is fast growing with continuous development in the field of power electronics and drives. The Variable Voltage Variable Frequency (VVVF) supply for electrical machines gives superior performance in terms of speed control, efficiency and dynamics compared to the machines operated directly from the mains. In one of the most basic configurations, a drive system consists of a closed loop speed control that has a current controller inside the loop. For effective and stable current control, the controller gains need to be set according to the parameters of the machine at hand. Besides, accurate parameter information is helpful in ensuring better machine exploitation as well as maintaining higher efficiency in various operating modes and conditions. The traditional methods of determining machine parameters consist of extensive machine testing under prescribed supply and ambient conditions. These methods become impracticable when the machine cannot be isolated from its load or the test equipment cannot be made available. Under such conditions, the alternatives are needed that use only the available hardware included in a standard drive to completely define the machine parameters. Self-commissioning thus comes into play in such situations. The automatic determination of machine electrical parameters before the drive is put in continuous operation is called self-commissioning of the drive system. In this thesis, self-commissioning of AC electric motors is studied, analyzed and results are presented for the implementation of different self-commissioning methods either proposed in the literature or developed in the course of this research. By far the commonest control strategy of AC machines is the vector control that allows dc machine like decoupled control of machine flux and torque. The separation of flux and torque producing current components depends heavily on the parameters of the machine at hand. In case the parameters fed to the controller do not match the actual machine parameters, the control performance deteriorates both in terms of accuracy and efficiency. For synchronous machines using permanent magnets, the magnetic model of the machine is important both for flux estimation accuracy at low speeds and for deriving maximum torque out of machine per ampere of input stator current. The identification of the magnetic model of permanent magnet synchronous machines requires special tests in a laboratory environment by loading the machine. A number of machine parameter identification methods have been studied in the past and proposed in the literature. As the power amplifier implied is almost always an inverter, the estimation of machine parameters at start-up by generating special test signals through the inverter have been researched in depth and are investigated in this thesis. These techniques are termed as offline parameter identification strategies. Other methods that focus on parameter updating during routine machine operation are called online parameter estimation methods. In this thesis, only the offline identification schemes are studied and explored further. With continuous improvements in power semiconductor devices' switching speeds and more powerful microprocessors being used for the control of electric drives, generating a host of test signals has been made possible. Analysing the machine response to the injected test signals using enhanced computational power onboard is relatively easier. These conditions favour the use of even more complex test strategies and algorithms for self-commissioning and to reduce the time required for conducting these tests. Moreover, the universal design of electric drives renders the self commissioning algorithms easily adaptable for different machine types used in industry. Among a number of AC machines available on the market, the most widely used in industrial drives are considered for study here. These include AC induction and permanent magnet synchronous machines. Induction machines still play a major part in industrial processes due, largely, to their ruggedness and maintenance-freeness; however, the permanent magnet machines are fast replacing them as competitive alternatives because of their low volume-to-power, weight-to-power ratios and higher efficiency. Their relatively light weight makes these machines a preferred choice in traction and propeller applications over their asynchronous counterparts