Research On Unit Commitment Considering The Uncertainty Of Wind Power Forecast Error

Recently,as the structure of global energy changes and renewable energy develops,exploiting wind energy has become a significant way to realizes replace of clean energy.However,the increase of wind power uncertain along with wind power capacity becomes a challenge for safe and economic operation of the power system.Unit commitment is one of the core issues of safe and economic operation.Operators make decisions by predicted wind power,but wind power forecast cannot ensure 100%accuracy.Although the prediction accuracy is constantly improving,the absolute error of the prediction is still large because of the increase of wind power penetration,so the impact of the errors on power system dispatching and operation cannot be ignored.Therefore,modeling of prediction errors,improving the utilization of prediction information,rational allocation of reserve on the wind farm side and coordinating the resources in the system are of great significance for improving the security of power system operation.Therefore,this paper first studies the uncertainty of forecast error and according to description method.And then a multi-correlation model of forecast errors is established considering the temporal and spatial correlation of forecast errors caused by geographical location and climate factors and the conditional correlation of forecast errors given wind power output level.Secondly,a chance-constrained programming based unit commitment model is built to accurately describe the correlation of prediction errors.And the model achieves an effective link between forecast error correlation and traditional unit commitment.Based on the study above,an interval unit commitment model and corresponding reserve decision-making method are proposed considering the flexibility of reserve and the participation in the regulation of wind farm side.The main contributions of this paper are as follows.1)Based on the research of the distribution characteristics and the spatial-temporal correlation of forecast errors,the conditional correlation between forecast errors and wind power output level is studied.Then the analytical expression of multiple correlations describing spatial-temporal-conditional correlation simultaneously is derived,and according to model of multiple correlations is established.Therefore,random variables of wind power forecast error can be generated directly by sampling.Case study based on practically wind farm data verifies that the proposed error model can effectively express the spatial-temporal correlation,and the wind power forecast error can change with the predicted wind power.2)A chance-constrained programming based unit commitment model considering multiple correlations of forecast errors is established.By using chance-constrained programming,the multiple correlations of forecast errors and the unit commitment with wind power integrated are effectively connected.In the model,the theory of conditional value at risk is introduced,and reserve and risk considering multiple correlations are well balanced by chance-constrained programming.At the same time,the impact of wind power uncertainty on scheduling reduced because of the refined expression of characteristics of wind power forecast error.Finally,the calculation and analysis of the proposed model are conducted on New England 10-unit system.By comparison of the results with multiple correlations and with no correlation.the economy and utility of the proposed method are verified.3)A two-stage interval unit commitment model is proposed.In the model,considering the correlation of wind power forecast errors and the constraints of system operation,the error interval of wind power is determined based on chance-constrained programming,thus a two-stage decision method from the wind power output interval determination to the optimal unit commitment is exploited.Through the iterative solution of the two stages based on C&CG.the optimal interval and the corresponding unit commitment scheme for this interval are finally given.In the model.the optimal reserve capacity is determined by the tradeoff between the risk of wind curtailment and loss of load and reserve cost.The model realizes the effective connection between power grid and wind power uncertainty,reduces the impact of wind power fluctuation and the pressure of wind power connection,and improves the efficiency of allocation of energy resources.The performance of the proposed method is verified by New England 10-unit system and IEEE 118-bus system.