DIGITAL LOAD-FREQUENCY CONTROL OF SMALL HYDROELECTRIC SYSTEMS USING NETWORKED PROCESSORS.

摘要

Small power systems based on renewable energy sources such as low head water falls, wind, solar and ocean tides can be made economically more viable if non-conventional control techniques like direct load control are used. A real-time, networked microprocessor controller has been developed which when complemented with a conventional automatic voltage regulator provides stable voltage and frequency for a stand-alone, run-of-the-river small hydroelectric power plant with no input control, i.e. no input governors or conventional gates and guide vanes.; The control objectives include stable frequency and maximizing the generation using directly controllable loads, usually semi-or fully-automated processes producing locally usable commodities. These loads are switched based on availability of power above the base load level to maintain a high (nearly 100%) load factor on the generator. A digital control model of such a system has been presented and analyzed. The PADS (Predicting, Adjusting to one of the Dual States) control algorithm has been developed to maintain frequency within 5% of the nominal. This algorithm predicts the machine frequency for the next sampling period and if the predicted value is out of the error bounds, an immediate correction is applied else a correction is applied, if needed, after every 100 samples. Analysis on a 100kW generating station and the results of the machine behavior using PADS algorithm for various step load changes are presented.; The controller is based on a 16-bit microprocessor system as a master and up to four 8-bit microprocessor systems as remote slaves. The master provides all executive and generator control tasks while the slaves are dedicated to load control tasks. The processors are networked on a communications bus organized as a star with the master at the hub. The operating system developed has the capabilities of real-time as well as background processing in addition to the communication processing and human interface.