Hydropower, as a controllable energy source, plays crucial role in supporting essential functions such peak shaving, frequency regulation, and load reserve within modern power systems, thereby enhancing the overall economic efficiency safety of these systems. This study utilizes data from small hydropower stations advanced software algorithms to preliminarily evaluate feasibility converting conventional Zhejiang Province into pumped storage stations, with province serving focal research...

Deep neural networks (DNNs) have substantially achieved high predictive accuracy in many vision tasks. However, we find that they are poorly calibrated for crack recognition tasks, as these DNNs tend to produce both under-confident and over-confident predictions such safety-critical applications, thereby limiting their practical use real-world scenarios. To address this issue, propose a novel attack-inspired calibration loss (AICL) explicitly regularizes class probabilities be better...

To realise the accurate and real‐time compensation of harmonics a power system, this study proposes novel control strategy for shunt active filter (SAPF), adopting composite neural network proportional‐integral (PI) dual‐repetitive controller (DRC). In DRC, one repetitive is used to ensure current tracking accuracy other enhance dynamic response. The PI adopted improve response speed by turning parameters adaptively setting optimisation online. Through approach rational configuration...

A nonlinear mathematical model for hydroturbine governing system (HTGS) has been proposed. All essential components of HTGS, that is, conduit system, turbine, generator, and hydraulic servo are considered in the model. Using proposed model, existence stability Hopf bifurcation an example HTGS investigated. In addition, chaotic characteristics with different parameters studied extensively presented form diagrams, time waveforms, phase space trajectories, Lyapunov exponent, attractors,...

Small hydropower (SHP) and pumped storage (PHS) are ideal members of power systems with regard to integrating intermittent production from wind PV facilities in modern using the high penetration renewable energy. Due limited capacity SHP geographic restrictions PHS, these sources have not been adequately utilized multi-energy integration. On one hand, rapidly increasing wind/PV is mostly situated remote areas (i.e., mountain rural areas) delivered core manufacturing bases cities) for...

Load shedding processes are widespread in hydropower stations, which has great influence on the safe and stable operation of hydro–turbine governing system. In order to study dynamic characteristics system during load process, a novel nonlinear mathematical model is established considering system, generator governor particular, six transfer coefficients presented based definitions internal characteristics. After that, from viewpoint dynamics practical engineering, investigated utilizing...

Axis-orbit recognition is an essential means for the fault diagnosis of hydropower units. An axis-orbit method based on feature combination and selection proposed, aiming to solve problems low accuracy, poor robustness, efficiency existing methods. First, various contour, moment, geometric features axis orbit samples are extracted from original data combined into a multidimensional set; then, Random Forest (RF)-Fisher applied realize dimensionality reduction; finally, selected set as input...

A fractional order PID (FOPID) controller, which is suitable for control system designing being insensitive to the variation in parameter, proposed hydroturbine governing paper. The simultaneous optimization several parameters of that is,<mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M1"><mml:mrow><mml:msub><mml:mrow><mml:mi>K</mml:mi></mml:mrow><mml:mrow><mml:mi>i</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:math>,<mml:math...

Pumped‐storage unit (PSU) often acts as the peak and frequency modulation sources in modern power systems. The frequent operation condition conversions tight hydraulic couplings with conduit system make its governing control a tough work practise. To overcome difficulty proportional–integral–derivative (PID) parameter tuning during transient processes, hierarchical model predictive (MPC) scheme, which constrained generalised (CGPC) cooperates classic PID is proposed. In upper hierarchy of...

Long time field tests of a 200 MW prototype Francis turbine over its full range operation were conducted. From the experimental data, domain and frequency characteristics pressure fluctuations in at different conditions analyzed. Furthermore, reason for amplitude increase correlation between vibration fluctuation was studied by using multidimensional band energy ratio analysis method. Based on above analysis, some hydraulic stability large are found, other results also obtained.

This paper discusses a least square support vector machine (LS-SVM) approach for forecasting stability parameters of Francis turbine unit. To achieve training and testing data the models, four field tests were presented, especially vibration in<mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M1"><mml:mrow><mml:mi>Y</mml:mi></mml:mrow></mml:math>-direction lower generator bearing (LGB) pressure in draft tube (DT). A heuristic method such as neural network using Backpropagation...

As the proportion of new energy increases in modern power systems, pumped-storage plant attaches significance to balancing supply and demand. Its operational stability has attracted extensive attention. Since hydraulic characteristics with a complicated conduit system are rather complex, it is difficult for traditional continuous impedance method obtain analytical solutions frequency responses directly. Therefore, an equivalent circuit modeling-based discrete proposed address this problem....

The research is mainly based on BP-ANN neural network method, through direct and indirect data collection to predict the photovoltaic super-short term power generation, discussed influence effection about generalization of model ,which used generation. Combined with actual operation situation, analyzed compared system results. results show that measured has same trend prediction results, root mean square error absolute are small, accuracy reached 96.71% high application value.