Accurate daily runoff forecasting is of great significance for the operation control of hydropower station and power grid. Conventional methods including rainfall-runoff models and statistical techniques usually rely on a number of assumptions, leading to some deviation from the exact results. Artificial neural network (ANN) has the advantages of high fault-tolerance, strong nonlinear mapping and learning ability, which provides an effective method for the daily runoff forecasting. However, its training has certain drawbacks such as time-consuming, slow learning speed and easily falling into local optimum, which cannot be ignored in the real world application. In order to overcome the disadvantages of ANN model, the artificial neural network model based on quantum-behaved particle swarm optimization (QPSO), ANN-QPSO for short, is presented for the daily runoff forecasting in this paper, where QPSO was employed to select the synaptic weights and thresholds of ANN, while ANN was used for the prediction. The proposed model can combine the advantages of both QPSO and ANN to enhance the generalization performance of the forecasting model. The methodology is assessed by using the daily runoff data of Hongjiadu reservoir in southeast Guizhou province of China from 2006 to 2014. The results demonstrate that the proposed approach achieves much better forecast accuracy than the basic ANN model, and the QPSO algorithm is an alternative training technique for the ANN parameters selection.

Large-capacity hydropower transmission from large hydropower plants in southwestern China to load centers is an important measure for the accommodation of large-scale hydropower in China. This paper develops an optimization model for determining the short-term generation scheduling of the hydropower plants transmitting electric power to several power grids via high voltage direct current (HVDC) transmission lines, taking the Xiluodu plant as an example. In this model, minimizing the peakvalley differences of multiple receiving-end power grids is adopted as the objective to raise the grids’ enthusiasm for the absorption of trans-regional hydro energy. In addition to the hydraulic constraints, the operation constraints of individual units and HVDC power transmission limits are well considered. The study focuses mainly on modeling the stair-like power transmission curve constraint which is discrete and nonlinear. This constraint is then linearized through limiting the logical relations between multiple binary integer variables. Case studies demonstrate that the proposed model has high computational efficiency, and the peak-valley differences of the two receiving-end power grids, i.e., Zhejiang Power Grid and Guangdong Power Grid, are decreased by 15.2% and 7.33%, respectively. Moreover, frequent conversion of HVDC.

The thirst of the Earth for energy is lurching towards catastrophe in an era of increasing water shortage where most of the power plants are hydroelectric. The hydro-based power systems are facing challenges in determining day-ahead generation schedules of cascaded hydropower plants. The objective of the current study is to find a speedy and practical method for predicting and classifying the future schedules of hydropower plants in order to increase the overall efficiency of energy by utilizing the water of cascaded hydropower plants. This study is significant for water resource planners in the planning and management of reservoirs for generating energy. The proposed method consists of data mining techniques and approaches. The energy production relationship is first determined for upstream and downstream hydropower plants by using multiple linear regression. Then, a cluster analysis is used to find typical generation curves with the help of historical data. The decision tree algorithm C4.5, Iterative Dichotomiser 3-IV, improved C4.5 and Chi-Squared Automatic Interaction Detection are adopted to quickly predict generation schedules, and detailed comparison among different algorithms are made. The decision tree algorithms are solved using SIPINA software. Results show that the C4.5 algorithm is more feasible for rapidly generating the schedules of cascaded hydropower plants. This decision tree algorithm is helpful for the researchers to make fast decisions in order to enhance the energy production of cascaded hydropower plants. The major elements of this paper are challenges and solution of head sensitive hydropower plants, using the decision-making algorithms for producing the generation schedules, and comparing the generation from the proposed method with actual energy production.

There is a common lack of flexible power to quickly follow peak loads of many coal-dominated provincial power grids (PPGs) in China. A regional power grid (RPG) that covers several provinces is responsible for scheduling the power plants it manages and allocating their power to subordinate PPGs in response to different peak loads. This process is more difficult than conventional peak operations for a single PPG. A methodology for hydrothermal system generation scheduling considering multiple provincial peak-shaving demands is developed. This methodology utilizes a novel load subsection optimization model to smooth local frequent load fluctuations. The model divides the provincial load curve into several subsections to construct an expected residual load profile with the average load of each subsection in advance. Minimizing deviation between the expected and calculated load profiles is then formulated as the optimization objective. Three practical algorithms, named the multistep progressive optimality algorithm, the heuristic algorithm, and an improved load shedding method, are integrated to determine the power generation for hydropower plants, pumped-storage plants, and thermal plants, respectively. Moreover, a variable neighborhood search algorithm is proposed to coordinate the allocation of the power generation among PPGs. This methodology has been implemented for the operations of the East China Grid, which is the largest RPG in China. Comparisons with an existing optimization model and a conventional method are, respectively, given in two cases. The first case shows that the peak-shaving demands in subordinate PPGs are effectively met. The load fluctuations in local hours are smoothed, with an average reduction of 6.3% in allan variance of residual loads. The second case implies that the proposed method deals better with the load differences in multiple power grids. It should also be noted that the operating flexibility of different power plants has a significant effect on smoothing fluctuations in residual loads.