| The deregulation of power industry creates the independent market entities with differentobjectives, including the self-interested power producers to maximize their payoffs and theindependent system operator (ISO) to ensure the system reliability. Therefore, the traditionalsystem-centralized mechanism of unit maintenance scheduling (UMS) is not valued anymoreand appropriate decision-making models need to be built to strike a tradeoff between theproducers’ benefits and the system operation. As a part of “Research of operation andreliability of power systems in electricity market”, supported in part by the Special Fund ofthe National Priority Basic Research of China (2004CB217905) and in part by the NationalNature Science Foundation of China (50807043), a systematic study on the UMS and otherrelated aspects in the market environment are carried out in this dissertation. Major researchwork and contributions are summarized as follows.1. Generating units may experience outage caused by unexpected failures, which would posea serious risk on the producers’ benefits and the system operation. Thus, a hazard function ispresented to characterize the bathtub-shaped failure curve of generating units, and the relatedparameters can be estimated from the historical data by the least-square-fitting method.Based on this, the unit availability and the expected renewal cost for replacing the damagedcomponents can be calculated. The UMS simulation results of generating company and thepower system shows that compared with the traditional up-down dual state model, theproposed one is more helpful to evaluate and avoid the related risk.2. In power market, electricity price forecasting is an essential work for the UMS research. Aclassified forecasting model and a hybrid simulation model are presented. In the formermodel, based on the related market factors, the proximal point and support vector machinetechniques are used to classify the future prices, which are estimated respectively. Thismodel reduces the price dependence on time and provides effective information for themedium-and long-term price forecasting. On the other hand, the hybrid model is presentedfor medium-term price simulation and consists of one ARMA-ARCH model and one jumpmodel, which are used to characterize the autoregressive, heteroscedastic features and the price peaks. Simulation results show that it can simulate the electricity prices effectively.3. The UMS mode relates to the market mechanism and the system adequacy. For the systemwith insufficient adequate capacity, UMS should be determined by the ISO via minimizingthe system operation costs of the reliability costs, the production costs, the units’ renewal andmaintenance costs. Whereas, for the system with sufficient adequate capacity and highlycompetitive market, the outage planning should be scheduled through two procedures. Firstly,individual producers plan their outage periods by maximizing their benefits and submit themaintenance windows to the ISO. Subsequently, the ISO coordinates the outage requests toformulate a system-wide plan and allocates the relevant costs for its non-profit characteristic.4. In the market environment, power producers challenge some uncertain factors, includingunexpected unit failures and the fluctuating electricity prices. According to the pool-basedenergy market, the relative UMS models are proposed to optimize the producers’ payoffs,considering the two factors, respectively. They are both solved by the combination ofGenetic Algorithm and the linear programming methods. Furthermore, they are extended to ahybrid market including bilateral contracts, energy market and reserve market. Case studiesshow that they can maximize the producers’ benefits and reduce the relevant risk. Note that,the obtained outage planning is mainly analyzed from the perspective of power producersand should not be carried out without approval of the ISO.5. Based on the system operation cost analysis, an iterative coordinated UMS model isproposed, thoroughly considering the impacts of unexpected unit failures on the producersand the system. Individual producers submit their outage schedules to the ISO, and the ISOdetermines whether they can ensure the system economic operation and synthesizes thecorrective signals. If not, the obtained signals would be sent to the related producers to directthem to reschedule their outage periods. Moreover, the corresponding reschedule costs areshared pro rata among power consumers. Case studies show that through several iterations, afeasible outage planning can be obtained to ensure the producers’ benefits as well as lowerdown the system operation costs.6. According to the different functions of power producers and the ISO, a novelmarket-competitive UMS mechanism is presented. Firstly, power producers evaluate thebenefits/costs of its owned units for each available outage window and submit a set ofmaintenance bidding costs (MBCs) to the ISO. Subsequently, the ISO determines theirsatisfactory degrees (SDs) and coordinates the outage periods to attain a balance among theMBCs, SDs of generating units and the possible load curtailment. Thirdly, as a non-profitorganization, the ISO settles the final expenditure of individual producers in terms of theSD-based approach. Case studies show that the proposed scheme characterizes simplicity,fairness, efficiency and the competitive market rules, which can be adopted for a real-sizedsystem. Meanwhile, the producers obtain either the well-pleasing outage windows or the extra economic compensation, which ensure the smooth implementation of the outagescheduling.As one of the important problems in power market, outage scheduling of generating unitsaffects not only the producers’ benefits, but also the system reliability and the marketstability. This dissertation carries out a systematic research on the UMS problem andpresents the proper models for different market mechanism. Several novel points and modelson the UMS theory and application are proposed. These researches may provide animportant reference for the development of the UMS theory. |
PDF Full Text Request