Identifying Key Factors of Calculated Inflow-Outflow Imbalance of Three Gorges-Gezhouba Dams

[Objective] Numerous influencing factors contribute to the imbalance between the calculated inflow and outflow discharges in the reach between the Three Gorges and Gezhouba Dams. Analyzing the significance of individual influencing factors and identifying the key drivers underlying this imbalance are of considerable importance for formulating power generation plans and conducting hydrological analysis of the Three Gorges-Gezhouba cascade hydropower stations. [Methods] Historical records of discharge and power output in the Three Gorges Reservoir Area were collected from 2018 to 2023. Based on the calculation logic of inflow and outflow discharges, 19 potential influencing factors that may affect the inflow-outflow imbalance in the reach between the two dams were selected. Grey relational analysis (GRA) and random forest (RF) model were employed to identify the key factors contributing to the inflow-outflow discharge imbalance between the Three Gorges and Gezhouba Dams. [Results] The results of GRA showed that the grey relational grade of the total power generation discharge of the Gezhouba Dam reached 0.696, ranking first. The grey correlation degree of the power generation flow of the Three Gorges Dam was 0.695, ranking second. The grey correlation degrees of the total active power of Gezhouba Dam, the total power generation flow of the Three Gorges Dam, and the total active power of the Three Gorges Dam were 0.661, 0.651, and 0.636, respectively, ranking third, fourth, and fifth. For RF model, two methods—rank assignment summation and normalization summation—were adopted to integrate the two indicators, namely %IncMSE (percentage increase in mean squared error) and IncNodePurity (increase in node purity). The results indicated that in both methods, the total power generation discharge of the Gezhouba Dam and the storage-release discharge of the Gezhouba Dam ranked first and second, respectively, in terms of importance. Specifically, the normalization summation method not only reflected the importance ranking of different influencing factors, but also demonstrated that the relative importance of each factor through specific indicator values. Among these factors, the total power generation discharge of the Gezhouba Dam scored the highest (1.25), followed by its storage-release discharge (1.22). In contrast, the total power generation discharge of the Three Gorges Hydropower Station, which ranked third, had a significantly lower score (0.61) than the storage-release discharge of the Gezhouba Dam, which ranked second. [Conclusion] The total power generation discharge of Gezhouba Dam is the key influencing factor causing the inflow-outflow discharge imbalance. This is mainly attributed to the following reasons: 1) the Gezhouba Dam launched its capacity expansion and renovation project in 2013, with the total installed capacity of its 19 generating units increasing by 475 000 kilowatts. However, the NHQ curve adopted for calculating the power generation discharge of the units still remains the original factory curve without any updates, which leads to calculation errors in power generation discharge. 2) The head loss of the Gezhouba Dam is derived from the calculation based on the inflow discharge, instead of being accurately determined for each individual unit, and this calculation method will also induce certain errors. 3) Due to the complex flow conditions in front of the Gezhouba Dam, there are often differences in water head between the left and right banks. Nevertheless, water-head data from a single monitoring station are applied uniformly in power generation discharge calculations, which may further contribute to calculation errors.