Start-up optimization control method for large hydroelectric generating set based on full-view simulation and dynamic verification

Abstract

The invention belongs to the field of hydroelectric generation control engineering and fluid mechanical dynamics, and particularly discloses a starting-up optimization control method of a large hydroelectric generating set based on full-view simulation and dynamic verification. Comprising the steps of multi-dimensional real machine test database construction and benchmark evaluation, one-dimensional system simulation and macroscopic boundary definition, three-dimensional high-fidelity transient flow field simulation and microscopic mechanism insight, startup law dynamic optimization based on simulation insight and'simulation-test 'technology closed loop and iterative refining. Compared with the prior art, one-dimensional system simulation and three-dimensional high-fidelity flow field simulation are organically combined for the first time, the problem of comprehensive analysis from the macroscopic hydraulic transition process to the microcosmic complex flow mechanism is solved, the'full-transparency 'insight of the starting process is achieved, multiple performance indexes such as vibration, throw, stress and temperature can be considered at the same time, and the method is suitable for large-scale popularization and application. And the overall stability of the unit is synergistically improved.

Description

Technical Field

[0001] This invention belongs to the field of hydropower control engineering and fluid machinery dynamics, specifically involving a start-up optimization control method for large hydropower units based on full-view simulation and dynamic verification. Background Technology

[0002] With the rapid development of hydropower technology, giant hydropower units with a single unit capacity of one million kilowatts (such as the Baihetan Hydropower Station) have become a major national asset. These units are characterized by complex structures, large moments of inertia, and complex hydraulic conditions. The start-up process of the unit is a complex transient process involving the coupling of multiple physical fields such as mechanics, hydraulics, and electromagnetics. During this process, unreasonable start-up patterns (such as guide vane opening rate, acceleration control, and open-loop / closed-loop switching timing) can induce strong pressure pulsations, mechanical vibrations, and sway, impacting key components of the unit (such as top covers, bolts, bearings, and stator bases), accelerating their fatigue damage, and threatening the safe and stable operation of the power station.

[0003] Existing startup pattern optimization technologies have the following main drawbacks: Simulation methods are limited and cannot cover both macroscopic and microscopic aspects: traditional one-dimensional hydraulic transient process calculations cannot capture the complex vortices, cavitation, and pressure pulsations behind the guide vanes and inside the runner during startup; while simple three-dimensional CFD flow field simulations involve extremely large computational loads and are difficult to simulate the dynamic response of the entire process from rest to rated speed, resulting in deviations between simulation results and the actual dynamic characteristics of the real machine.

[0004] Lack of consideration for structural fatigue: Existing startup optimization methods mainly focus on speed overshoot and settling time, neglecting the transient dynamic stress and cumulative fatigue damage caused by the severe pressure pulsations during startup to critical connecting components such as the top cover and bolts. For megawatt-class units, the enormous water thrust can cause micro-deformation of structural components, leading to vibration, which traditional methods lack assessment for.

[0005] The disconnect between experiment and theory: Existing technologies often rely on trial and error to adjust PID parameters or guide vane opening modes on-site, lacking a closed-loop mechanism of "simulation prediction - real machine verification - model correction - strategy optimization", resulting in high debugging risks and long cycles.

[0006] Therefore, existing technologies lack a start-up control method for large-scale hydropower units that can integrate macroscopic system characteristics and microscopic flow mechanisms, and can perform multi-objective and multi-dimensional precise optimization under low-risk conditions. Summary of the Invention

[0007] To address the shortcomings of the aforementioned technologies, this invention provides a start-up optimization control method for large-scale hydropower units based on full-view simulation and dynamic verification.

[0008] To achieve the above objectives, the present invention provides the following technical solution: This invention provides a start-up optimization control method for large-scale hydropower units based on full-view simulation and dynamic verification, comprising the following steps: S1: For the target unit, the system executes multiple preset start-up control modes, simultaneously collects structural dynamics data, vibration data, stability data, thermodynamic data, and hydraulic data, establishes a benchmark database, and quantitatively evaluates the advantages and disadvantages of each start-up mode; S2: Based on the method of characteristics, establish a one-dimensional simulation model of the unit's hydraulic-mechanical-electric system, calculate the macroscopic dynamic response of the power generation system under different start-up conditions, define the safe operation boundary, and provide time-varying boundary conditions; S3: Based on the time-varying boundary conditions in step S2, establish a three-dimensional full-channel high-fidelity CFD model containing all flow components, conduct transient unsteady flow field simulation, and reveal the complex internal flow phenomena and excitation sources. S4: Combining the simulation results of steps S2 and S3, the start-up curve is adjusted for the strong vibration flow region. By optimizing acceleration, open-loop holding segment, and open-loop / closed-loop switching point, a multi-segment variable parameter start-up optimization curve is generated. S5: Apply the optimized startup rules from step S4 to real machine testing, compare the test results with the simulation prediction results, and solidify the rules if the consistency meets the standard; otherwise, modify the model parameters and iteratively optimize. S6: Based on the Riemannian manifold, a multi-index comprehensive scoring system is constructed to calculate the comprehensive score of each boot method and select the optimal boot method.

[0009] Furthermore, step S1 specifically includes the following sub-steps: S11: Select the start-up method, including guide vane large start, guide vane small start, variable acceleration control, and open-loop and closed-loop combined start; S12: Arrange the start-up test measurement points, including unit vibration measurement points, swing measurement points, water pressure pulsation measurement points, strain measurement points, stress measurement points, and speed measurement points; S13: Conduct a start-up test at the selected head and record the changes in guide vane opening and data at each measuring point over time; S14: Process the acquired data and calculate the peak-to-peak amplitude and stress value of the mixing frequency; S15: Compare and analyze the same data under different boot methods to quantitatively evaluate the advantages and disadvantages of each boot method.

[0010] Further, in step S14, the peak-to-peak amplitude of the mixing frequency is taken with a 97% confidence level, that is, the probability of counting points in the time-domain waveform is statistically analyzed by partitioning the waveform, and 3% of unreliable data is removed before calculation; the stress is calculated based on the strain values ​​collected during the test using the following formula: in, The strain value, E The elastic modulus of the material being tested. For stress.

[0011] Furthermore, step S2 specifically includes the following sub-steps: S21: Construct a mathematical model of a hydraulic system based on the method of characteristics. The hydraulic system includes a pressure pipeline, a turbine, and a governor. S22: Conduct simulation calculations under different start-up modes, analyze the time-domain changes of unit speed, volute pressure, and tailrace pipe pressure, and propose an initial optimized start-up mode.

[0012] Furthermore, in step S21, the mathematical model of the pressure pipeline is established using the following formula: In the formula, the subscript i Let be any grid intersection point in the x-direction; Let P be the head of the water. Let P be the flow rate. , , , It is a moment The known quantities are obtained using the following formula: In the formula, constant B and R The following formula is used to obtain it: In the formula, A The cross-sectional area of ​​the pipeline is in meters. 2 ; g The acceleration due to gravity is m / s². 2 ; f This refers to the Darcy-Weisbach friction loss coefficient. D Pipe diameter, in meters (m); For a long L The pipes are divided into equal parts n The length of each subsequent segment, in meters; a The water hammer wave velocity, in m / s, is obtained using the following formula: In the formula, K Let be the bulk modulus of water; The density of water; E The elastic modulus of the pipe is expressed in N / m. 2 ;D The inner diameter of the pipe is in meters (m). e Let be the pipe wall thickness, in meters (m).

[0013] Furthermore, in step S21, the mathematical model of the water turbine is established using the following formula: In the formula, It is a flow function; It is a torque function; This represents the relative value of the guide vane opening. For guide vane opening; This is a relative value of the water head; This is a relative value of the flow rate; This is a relative value of the rotational speed; This is a relative value of the torque; This refers to the relative angular velocity deviation. This represents the relative value of the instantaneous shaft torque of the water turbine; This represents the relative value of electromagnetic resistance. Rated torque; Let be the unit's inertial time constant, in seconds; For flywheel torque, t·m 2 ; Rated power, kW; ; In the above formula, the subscript r represents the rated parameter value.

[0014] Furthermore, in step 21, the mathematical model of the speed regulator is established using the following formula: In the formula, This represents the relative value of the output signal deviation of the speed controller. This represents the relative value of the unit's speed deviation; the subscript 0 indicates the initial value. This is the permanent slip coefficient, and its value is selected in the range of 0 to 0.1; This refers to the reaction time of the guide vane relay. This is the relative value of the guide vane servo motor's stroke deviation, representing the servo motor's stroke y-deviation relative to its maximum stroke. The ratio; The relative value of the main pressure regulating valve stroke deviation represents the ratio of the main pressure regulating valve's stroke from the center position to its maximum stroke, while the maximum stroke refers to the main pressure regulating valve stroke when the input speed signal deviation is 100%. This refers to the proportional gain of the speed controller; The integral gain of the speed controller; The differential gain of the speed controller; The following formula is used to obtain it: In the formula, This refers to the transient slip coefficient of the speed governor; This is the buffer time constant of the speed controller; This is the differential time constant or acceleration time constant of the speed controller.

[0015] Furthermore, step S3 specifically includes the following sub-steps: S31: Establish a full-basin geometric model of the unit's volute, fixed guide vanes, movable guide vanes, runner, and tailrace pipe; S32: Mesh the geometric model; S33: Set boundary conditions, using SST. k-ω The turbulence model uses the closed Navier-Stokes equations. The total pressure is set at the inlet of the volute and the static pressure is set at the outlet of the tailrace pipe. The steady-state calculation of the runner adopts the rotating coordinate system method and the transient calculation adopts the sliding mesh method. S34: Arrange pressure pulsation measuring points near the impeller, volute, fixed guide vanes, movable guide vanes, bladeless area, and tailrace pipe; S35: Flow field evolution analysis for typical calculation conditions with different start-up patterns; S36: Comparative analysis of the pressure pulsation evolution of each flow component under different start-up conditions; S37: Initial screening for the optimal boot method.

[0016] Furthermore, step S31 specifically includes the following sub-steps: S311: Organize the component structural data of the turbine casing, fixed guide vanes, movable guide vanes and tailrace pipe, and use SolidWorks modeling software to create a three-dimensional geometric model. S312: After obtaining point cloud data through on-site 3D scanning, the 3D geometric model of the wheel is obtained through reverse modeling; S313: Directly establish the three-dimensional geometric structure of the watershed for the volute and tailrace pipe, obtain the corresponding watershed for the impeller and guide vanes through Boolean operations, and assemble the components to form a full watershed geometric model.

[0017] Furthermore, in step S5, if the experimental results and simulation prediction results are consistent with the preset threshold and the optimization target is achieved, the startup rule is solidified; otherwise, the experimental data is fed back to the one-dimensional and three-dimensional simulation models, the parameters are corrected, and steps S2-S5 are repeated until the optimal startup rule that meets the safety and stability requirements is obtained.

[0018] Based on the above technical solution, the embodiments of the present invention can produce at least the following technical effects: (1) Full-view insight: For the first time, one-dimensional system simulation and three-dimensional high-fidelity flow field simulation are organically combined, overcoming the comprehensive analysis problem from macroscopic hydraulic transition process to microscopic complex flow mechanism, and realizing "full transparency" insight into the start-up process.

[0019] (2) Precise optimization: Optimization based on micro-flow mechanism can be targeted and suppress vibration from the source. The optimization effect is far superior to traditional methods based on experience and macro-model.

[0020] (3) Low risk and high efficiency: Most of the optimization work is carried out in a high-fidelity digital space, which greatly reduces the number of high-risk real machine tests, shortens the optimization cycle, and reduces costs.

[0021] (4) Multi-objective coordination: It can simultaneously measure multiple performance indicators such as sway, sway, stress, and temperature, and achieve a coordinated improvement in the overall stability of the unit.

[0022] (5) Strong versatility: The closed-loop technology of “simulation-driven-experimental verification-law optimization” formed by this method can be extended to the optimization of other types of large hydropower units and various transition processes (such as shutdown and load shedding). Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0024] Figure 1 This is an overall flowchart of the optimized start-up control method for large hydropower units proposed in this invention; Figure 2 This is a schematic diagram comparing the traditional start-up rules and the optimized start-up rules of the large hydropower unit start-up optimization control method proposed in this invention. Figure 3 This is a comparison of the peak-to-peak vibration values ​​of the upper frame after the optimization of the start-up optimization control method for large hydropower units proposed in this invention. Figure 4This is a comparison of the peak-to-peak vibration values ​​of the lower frame after the optimization of the large hydropower unit start-up optimization control method proposed in this invention. Figure 5 This is a comparison of the peak-to-peak values ​​of stator frame vibration after the optimized start-up control method for large hydropower units proposed in this invention. Figure 6 This is a comparison diagram of the vibration velocity of the top cover after optimization of the start-up optimization control method for large hydropower units proposed in this invention. Figure 7 This is a feature line mesh diagram of the large-scale hydropower unit start-up optimization control method proposed in this invention; Figure 8 The proposed optimization control method for starting up large hydropower units in this invention provides the transient curve of the volute inlet pressure pulsation in start-up mode 1. Figure 9 The proposed optimization control method for starting up large hydropower units in this invention provides the transient curve of the volute inlet pressure pulsation in start-up mode 2. Figure 10 The proposed optimization control method for starting up large hydropower units in this invention provides the transient curve of the volute inlet pressure pulsation in start-up mode 3. Figure 11 The proposed optimization control method for starting up large hydropower units in this invention provides the transient curve of the volute inlet pressure pulsation in start-up mode 4. Figure 12 The evolution process of the tailrace vortex band in start-up mode 1, as revealed by the three-dimensional high-fidelity CFD simulation of the start-up optimization control method for large hydropower units proposed in this invention (sequence diagram). Figure 13 The three-dimensional high-fidelity CFD simulation of the start-up optimization control method for large hydropower units proposed in this invention reveals the evolution of pressure pulsation over time at the monitoring point in the volute domain of start-up mode 1. Figure 14 The three-dimensional high-fidelity CFD simulation of the large hydropower unit start-up optimization control method proposed in this invention reveals the evolution of pressure pulsation over time at the monitoring point in the volute domain of start-up mode 2. Figure 15 This is the optimized start-up control curve of the large hydropower unit start-up optimization control method proposed in this invention. Detailed Implementation

[0025] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. In addition, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such combination of technical solutions does not exist and is not within the scope of protection claimed by the present invention.

[0026] This embodiment uses the Baihetan Hydropower Station's 1,000 MW turbine-generator unit as a case study to develop a hydropower unit start-up control method that deeply integrates macroscopic hydraulic transition processes and microscopic internal flow field characteristics, and achieves precise, efficient, and low-risk optimization through a closed-loop "simulation-driven-experimental verification" technology. (See reference...) Figure 1 The method includes the following steps: S1: Construction of a multi-dimensional real-machine test database and benchmark evaluation; For the target unit, systematically execute multiple preset start-up control modes, and simultaneously collect the following multi-dimensional data during each start-up process: Structural dynamics data: Time-domain signals of dynamic stress in the unit's top cover and connecting bolts.

[0027] Vibration data: Vibration acceleration / velocity of the upper frame, lower frame, stator base, and top cover in the X, Y, and Z directions.

[0028] Stability data: runout (X and Y directions) of the upper guide, lower guide, and water guide bearings.

[0029] Thermodynamic data: Oil temperature rise curves for the three bearings.

[0030] Hydraulic data: Time-domain signals of pressure pulsations in specific parts of a turbine (such as the bladeless section and the tailrace).

[0031] By comprehensively comparing and analyzing the massive amount of data mentioned above, the advantages and disadvantages of each power-on method are quantitatively evaluated, and a benchmark database is established with "total vibration value of key measuring points" and "peak dynamic stress of key components" as the core evaluation indicators.

[0032] Specifically, the construction and benchmark evaluation of the multi-dimensional real-machine test database includes the following steps: S11: See also Figure 2 The start-up method is selected for the target unit, including open-loop start-up with large guide vane acceleration opening, open-loop start-up with small guide vane acceleration opening, closed-loop start-up with acceleration control, and a combination of open-loop and closed-loop start-up.

[0033] In this embodiment, the open-loop start-up of the guide vane at large acceleration opening is as follows: The first starting opening degree of the guide vane is the no-load opening limit (140% of the no-load opening degree, the rate of change of the guide vane control output). v Limited to 0.5% / s), when the unit speed exceeds 90% of the rated speed, the guide vane opening is reduced to the second starting position (125% of the no-load opening). When the unit speed exceeds 96% of the rated speed, the unit switches to no-load operation, and the guide vane opening is reduced back to the no-load opening. This starting method is currently used at the Baihetan Power Plant. Figure 2 Method 1.

[0034] In this embodiment, the open-loop start-up of the guide vane at a small acceleration opening is as follows: The guide vane starting opening is approximately 60% of the no-load opening (the rate of change of the guide vane control output). v (Limited to approximately 1.25% / s), wait approximately 8 seconds, then switch to the second starting opening of no-load Y0 plus 8% (the rate of change of the guide vane control output). v Limited to approximately 1.25% / s), when the unit speed reaches about 90% of the rated speed, the pressure returns to the third starting opening, which is 5% larger than the no-load opening (the rate of change of the guide vane control output). v Limited to approximately 1.25% / s), when the unit speed reaches about 95% of the rated speed, the speed governor engages frequency PID control. This starting method corresponds to... Figure 2 Method 2.

[0035] In this embodiment, the acceleration control closed-loop startup is as follows: From the start of unit rotation to 20% of rated speed, the frequency setpoint tracks the unit's frequency; before the unit speed reaches approximately 85% of rated speed, the unit acceleration is set to 2Hz / s; before the unit speed reaches approximately 95% of rated speed, the unit acceleration is set to 1Hz / s; after the unit speed reaches 95% of rated speed, the unit acceleration is set to 0.5Hz / s, until the frequency setpoint reaches the rated value. If the frequency setpoint is already greater than 49.7 Hz, maintain this for more than 5 seconds. If it still does not reach the rated speed of 50Hz, directly apply the rated speed of 50Hz or the grid frequency. This startup method corresponds to... Figure 2 Methods 3.1 and 3.2. The PID coefficients are different in the two methods. The PID coefficient for method 3.1 is... K p =1.0、 K i =0.1、 K d =0.1, the PID coefficient for method 3.2 is K p =2.0、 K i =0.15、 K d =0.1.

[0036] In this embodiment, the combined open-loop and closed-loop startup is as follows: After receiving the start-up command, the speed governor sets the opening control to the starting opening, with the frequency setpoint at 40Hz. The opening tracking control quickly expands to the starting opening. When the frequency rises to 40Hz, PID regulation automatically engages, and the frequency setpoint begins to automatically increase from 40Hz to 50Hz (the grid frequency in frequency tracking mode) according to a set ramp (0.4Hz / s). The unit frequency will then smoothly reach 50Hz. This start-up method corresponds to... Figure 2 Method 4.

[0037] S12: Arrange start-up test test points on the target unit.

[0038] In this embodiment, the power-on test measurement points include: S121: Unit vibration measurement points, including one measurement point each in the horizontal +X and +Y directions of the upper frame, lower frame, and top cover, and one measurement point in the vertical +X direction; S122: Unit swing measurement points, including upper guide, lower guide, water guide bearings, and one measurement point each in the +X and +Y directions; S123: Water pressure pulsation measuring points, including one measuring point each at the volute inlet, the bladeless zone, and the tailrace outlet; S124: Other measuring points, including strain measuring points, stress measuring points, rotation speed measuring points, etc.

[0039] S13: Conduct a start-up test and record the measurement data. At the selected head, conduct a start-up test and record the changes in guide vane opening over time under different start-up modes, as well as the changes in unit vibration, sway, pressure pulsation, and other data over time.

[0040] S14: Data Processing. The method for determining the mixing frequency amplitude of unit vibration, sway, and pressure pulsation uses the mixing peak-to-peak amplitude with a 97% confidence level. This involves partitioning the strain time-domain waveform acquired by the computer, counting the points in each partition, calculating the probability of each point, and removing data from the 3% unreliable region to determine the mixing peak-to-peak amplitude. Stress (MPa) Based on the strain values ​​collected during the test. (με) is calculated using the following formula: In the formula, E This represents the elastic modulus of the material being tested.

[0041] S15: Analyze the data. See also Figures 3-6The same data under different start-up methods were compared to quantitatively evaluate the advantages and disadvantages of each method. For Method 1, i.e., the start-up method with the traditional guide vane large opening, the peak-to-peak value of the water guide vane swing + X was 652 μm, the maximum peak-to-peak value of the upper guide vane swing was 214 μm, and the maximum peak-to-peak value of the lower guide vane swing was 327 μm. The peak-to-peak value of the top cover horizontal vibration + Y was 181 μm, the maximum peak-to-peak value of the upper frame vibration was 39 μm, the maximum peak-to-peak value of the lower frame vibration was 58 μm, and the maximum peak-to-peak value of the stator base vibration was 31 μm. The relative value of pressure pulsation in the bladeless zone was 24.5%, the relative value of pressure pulsation at the volute inlet was 1.3%, the relative value of pressure pulsation at the tailrace inlet was 1.8%, the relative value of pressure pulsation at the tailrace elbow was 1.5%, and the relative value of pressure pulsation at the tailrace outlet was 0.2%. The maximum standard deviation of the top cover vibration velocity was 3.43 mm / s.

[0042] S2: One-dimensional system simulation and macroscopic boundary definition; establishing a one-dimensional simulation model of the unit's hydraulic-mechanical-electrical system. This model is used to quickly calculate the macroscopic dynamic response of the entire power generation system under different start-up conditions, including: the speed increase process, torque changes, volute pressure, water hammer in the intake system, etc. Its core function is to define the safe operating boundaries during the start-up process (such as avoiding excessively high speed or excessively low pressure), and to provide accurate time-varying boundary conditions (such as flow rate and speed changing with time) for subsequent three-dimensional high-fidelity simulation.

[0043] Specifically, one-dimensional system simulation and macroscopic boundary definition include the following steps: S21: See also Figure 7 The mathematical model of the hydraulic system is constructed based on the method of characteristics, mainly including pressure pipelines, turbines, and governors.

[0044] In this embodiment, the mathematical model of the pressure pipeline is established using the following formula: In the formula, the subscript i represents any grid intersection point in the x-direction; Let P be the head of the water. Let P be the flow rate. , , , It is a moment The known quantities can be obtained using the following formula: In the formula, constant B and R The following formula is used to obtain it: In the formula,A The cross-sectional area of ​​the pipeline is in meters. 2 ; g The acceleration due to gravity is m / s². 2 ; f This refers to the Darcy-Weisbach friction loss coefficient. D Pipe diameter, in meters (m); For a long L The pipes are divided into equal parts n The length of each subsequent segment, in meters; a The water hammer wave velocity, in m / s, is obtained using the following formula: In the formula, K Let be the bulk modulus of elasticity of water, which is generally taken as . ; The density of water is approximately 1000 kg / m³. 3 ; E The elastic modulus of the pipe is expressed in N / m. 2 D is the inner diameter of the pipe, in meters. e Let be the pipe wall thickness, in meters (m).

[0045] In this embodiment, the mathematical model of the water turbine is established using the following formula: In the formula, It is a flow function; It is a torque function; This represents the relative value of the guide vane opening. For guide vane opening; This is a relative value of the water head; This is a relative value of the flow rate; This is a relative value of the rotational speed; This is a relative value of the torque; This refers to the relative angular velocity deviation. This represents the relative value of the instantaneous shaft torque of the water turbine; This represents the relative value of electromagnetic resistance. Rated torque; Let be the unit's inertial time constant, in seconds; For flywheel torque, t·m 2 ; Rated power, kW; ; In the above formula, the subscript r represents the rated parameter value.

[0046] In this embodiment, the mathematical model of the speed governor is established using the following formula: In the formula, This represents the relative value of the output signal deviation of the speed controller. This represents the relative value of the unit's speed deviation, with the subscript "0" indicating the initial value. This is the permanent slip coefficient, and its value is selected in the range of 0 to 0.1; This refers to the reaction time of the guide vane relay. This is the relative value of the guide vane servo travel deviation, representing the travel deviation (guide vane opening) y-value relative to the maximum travel. The ratio; The relative value of the main pressure regulating valve stroke deviation represents the ratio of the main pressure regulating valve's stroke from the center position to its maximum stroke. The maximum stroke refers to the main pressure regulating valve stroke when the input speed signal deviation is 100%. This refers to the proportional gain of the speed controller; The integral gain of the speed controller; The differential gain of the speed controller; The following formula is used to obtain it: In the formula, This refers to the transient slip coefficient of the speed governor; This is the buffer time constant of the speed controller; This is the differential time constant or acceleration time constant of the speed controller.

[0047] S22: See also Figures 8-11Simulation calculations were conducted under different start-up methods to analyze the time-domain changes in unit speed, inlet and outlet water pressure of the volute, and outlet water pressure of the tailrace pipe. Based on this, an optimized start-up method was proposed. When the target unit adopted method 2 (i.e., open-loop start-up with guide vane small acceleration), the start-up time of the unit increased by nearly 12 seconds compared to method 1, reaching 84.39 seconds. However, the peak-to-peak values ​​of the inlet and outlet pressures of the two units decreased by 31.2% to 45.6% compared to method 1. When adopting method 3 (i.e., closed-loop start-up with unit acceleration control), the start-up time of the unit increased by nearly 20 seconds compared to method 1, reaching 92.76 seconds. However, the peak-to-peak values ​​of the inlet and outlet pressures of the two units decreased by 27.4% to 44.2% compared to method 1. Both methods 2 and 3 can ensure the operational stability of the unit well, but the start-up time of method 2 is shorter than that of method 3. Therefore, considering both start-up time and unit stability, method 2 is the optimal start-up method.

[0048] S3: High-fidelity 3D transient flow field simulation and microscopic mechanism insight; Based on the high-precision time-varying boundary conditions provided in Step 2, a high-fidelity 3D full-channel CFD model is established, encompassing all flow components (volute, fixed guide vanes, movable guide vanes, runner, and draft tube). Transient unsteady flow field simulation is performed to accurately reproduce the complex flow phenomena inside the turbine during startup, including the generation and evolution of gap vortices at the guide vane tips, flow separation and vortices in the runner's internal flow channels, and the morphology, frequency, and pressure pulsation characteristics of the draft tube vortex bands. Through this step, the macroscopic unit response is directly correlated with the microscopic flow field excitation source, fundamentally explaining the physical causes of the various vibration and pressure pulsation phenomena observed in Step 1.

[0049] Specifically, the three-dimensional high-fidelity transient flow field simulation and microscopic mechanism insight include the following steps: S31: Establish a full-basin geometric model of components such as the turbine casing, fixed guide vanes, movable guide vanes, and tailrace pipe.

[0050] In this embodiment, establishing a full-flow-domain model of components such as the turbine casing, fixed guide vanes, movable guide vanes, and tailrace includes the following steps: S311: Organize the structural data of components such as the turbine casing, fixed guide vanes, movable guide vanes, and tailrace pipe, and use SolidWorks modeling software to create a three-dimensional geometric model.

[0051] S312: After obtaining point clouds through on-site 3D scanning, the 3D geometric model of the wheel is obtained through reverse modeling.

[0052] S313: The three-dimensional geometric structure of the watershed is directly established for the volute and tailrace pipe, while for the impeller and guide vanes, the corresponding watershed needs to be obtained through Boolean operations, and the components are assembled to form a full watershed geometric model.

[0053] S32: Mesh the geometric model.

[0054] S33: Set appropriate boundary conditions; use SST for calculation. k-ω The turbulence model uses the closed Navier-Stokes equations. The boundary conditions are the total pressure at the volute inlet and the static pressure at the tailrace outlet. The pressure magnitudes are calculated based on Bernoulli's equation and known upstream and downstream water levels. The runner uses the rotating coordinate system method in steady-state calculations and the sliding mesh method in transient calculations. All runner walls, including the walls on the runner side of the upper crown gap and lower ring gap, are set as rotating walls.

[0055] S34: Set up pressure pulsation measurement points; arrange pressure monitoring points near the impeller, volute, fixed guide vane, movable guide vane, bladeless area, and tailrace pipe to compare and analyze flow characteristics.

[0056] S35: See also Figure 12 We conduct flow field evolution analysis on typical calculation conditions with different start-up patterns, namely, flow field distribution analysis corresponding to different times and guide vane openings.

[0057] In this embodiment, the flow field evolution analysis for typical computational conditions with different start-up patterns includes: S351: Analysis of the external characteristics of a water turbine at typical calculation operating points, including runner speed, mass flow rate, torque, power, etc.

[0058] S352: Analysis of the variation law of the force characteristics of the turbine runner under typical calculation conditions.

[0059] S353: Full-channel streamline distribution analysis at typical calculation points.

[0060] S354: Pressure distribution analysis of the intermediate section of the volute and fixed guide vane at typical calculation operating conditions.

[0061] S355: Analysis of pressure and velocity distribution at the mid-section of the active guide vane under typical calculation conditions.

[0062] S356: Analysis of vortex distribution of active guide vanes at typical calculation operating points.

[0063] S357: Analysis of the pressure and velocity distribution on the impeller cross section and impeller wall at typical calculation conditions.

[0064] S358: Analysis of turbine vortex distribution at typical calculation points.

[0065] S359: Analysis of pressure and velocity distribution across tailrace section at typical calculation points.

[0066] S3510: Vortex distribution analysis of tailrace pipe at typical calculation conditions.

[0067] S36: See also Figure 13 and Figure 14 A comparative analysis was conducted on the pressure pulsation evolution of the volute, fixed guide vanes, movable guide vanes, bladeless zone, and tailrace pipe under different start-up conditions.

[0068] S37: Compare and select the optimal start-up method. Within 6.5s-30s, the average pressure changes of both Method 1 and Method 2 are very stable, but the amplitude of the pulsating pressure change in Method 1 is greater than that in Method 2. Within 30s-42s, although the amplitude of the pulsating pressure change in Method 2 is greater than that in Method 1, the average pressure fluctuation of Method 1 is greater than that in Method 2. From the perspective of unit stability, Method 2 is better.

[0069] S4: Dynamic optimization of start-up patterns based on simulation insights; combining the simulation results of steps S2 and S3, the start-up patterns are reverse-engineered and optimized. For the strong vibration flow regime range revealed by the 3D simulation (such as severe vortex bands in the tailrace pipe at specific speeds / openings), the start-up curve is adjusted in the 1D simulation to actively avoid or quickly traverse this unfavorable range. By adjusting acceleration, setting open-loop holding segments, and optimizing open-loop / closed-loop switching points, the rise curves of power and torque are smoothed to reduce the impact on the mechanical structure. Finally, a multi-segment variable parameter start-up optimization curve is generated, which macroscopically satisfies the system stability requirements and microscopically effectively suppresses the generation of unfavorable flow regimes.

[0070] See Figure 15 For the target unit, the final optimized start-up method is as follows: the guide vane start-up opening is approximately 60% of the no-load opening (the rate of change of the guide vane control output). v (Limited to approximately 0.5% / s), wait approximately 15 seconds, then switch to the second starting opening of 160% of the no-load opening (the rate of change of the guide vane control output). v Limited to approximately 0.5% / s), when the unit speed reaches about 90% of the rated speed, the pressure returns to the no-load opening limit, which is 5% larger than the no-load opening (the rate of change of the guide vane control output). v (Limited to about 0.5% / s), when the unit speed reaches about 95% of the rated speed, the speed governor engages frequency PID regulation.

[0071] S5: Closed-loop and iterative refinement of the "simulation-experiment" technology; Apply the optimized startup rules from step S4 to the actual machine test, repeating the data acquisition process of step S1. Compare and verify the test results with the simulation prediction results. If the consistency is high and the optimization goal is achieved, the rules are solidified; if there are deviations, the test data are fed back to the one-dimensional and three-dimensional simulation models to correct the model parameters, initiating a new round of "simulation optimization-experiment verification" iterations until the optimal startup rules that meet all safety and stability requirements are obtained. The optimized startup rules were applied to the target unit, increasing the overall startup time by 3.3 seconds. The vibration and peak-to-peak swing values ​​of this startup method changed, with the upper guide swing decreasing by 33%, the lower guide swing decreasing by 29%, the water guide swing decreasing by 19%, and the pressure pulsation in the bladeless zone decreasing by 19%. Stability was improved while the startup time was extended by a relatively short amount.

[0072] S6: Booting method evaluation; Based on the Riemannian manifold, a multi-index comprehensive scoring system is constructed to calculate the comprehensive score of each booting method and select the optimal booting method.

[0073] Specifically, this invention achieves comprehensive scoring through 6 core calculations, each step of which is designed around "strengthening the contribution of small indicators". The formula is rigorous and operable.

[0074] Let N be the number of schemes to be evaluated, M be the number of evaluation indicators, and X = [x ij ]∈R N×M} (x ij Let be the j-th indicator value of the i-th scheme (the smaller the indicator value, the better). All calculations are designed around the logic of "small indicator → big contribution → high score".

[0075] S61: Inverse Min-Max Normalization Traditional Min-Max normalization results in a situation where "the larger the index value, the larger the normalized value," which contradicts the core requirement of this invention. Therefore, an inverse transform is designed: Where, max j (X(:, j )) is the first j The maximum value of each indicator, min j (X(:, j )) is the first j The minimum value of each index. This transformation ensures that: x ij The smaller X norm ( i , j The larger the value, the more it lays the foundation for "small indicators with high contributions" in subsequent weighted calculations. ε It is a very small value (usually 10⁻⁸) used to avoid division by zero errors.

[0076] S62: Calculation of Riemannian manifold metric tensors A Gaussian kernel function is used to construct a metric tensor to characterize the nonlinear correlation between metrics—the stronger the correlation, the larger the tensor value. Among them, g jk (X) is the first j The and the first k The metric tensor value of each indicator; range (0,1], the larger the value, the stronger the non-linear correlation between the two indicators; For the first j , k Euclidean distance of each index; For kernel width (empirical formula: This ensures that the kernel function can effectively capture the correlation scale between indicators.

[0077] S63: Riemann Distance Calculation The essence of the distance between two points in a Riemannian manifold is the line integral of the tensor, which is approximated in discrete scenarios as: Where, d M (x ij x, 0) is the Riemann distance from the j-th index of the i-th scheme to the zero reference point; the range is (0,+∞), x ij The smaller the value (the better the indicator), the smaller the distance; g jj (X) is the autocorrelation metric tensor value of the j-th index (the result of S62 calculation).

[0078] S64: Inverse Distance Weight Calculation The weights are designed based on Riemann distance, giving higher weights to smaller indicators, and the weights satisfy the normalization condition: Among them, w ij It is the inverse distance weight of the j-th index of the i-th scheme; the range is (0,1), satisfying ; It is the reciprocal of the Riemann distance; its function is to quantify the logic of "small distance → high weight", directly strengthening the weight contribution of small indicators.

[0079] S65: Calculation of manifold regularization term To avoid abrupt changes in weights on the Riemannian manifold, a regularization term is introduced to constrain the geometric smoothness of the score. The smaller the regularization term, the more stable the score. Where R is the manifold regularization term; its range is [0, +∞), and a smaller value indicates a more stable score; λ is the manifold regularization coefficient, which controls the strength of the geometric constraints (empirical value 0.01~0.1). The j-th indicator is the normalized mean. It is the squared difference between the normalized value and the mean; its function is to quantify the degree of deviation of the indicator value. The greater the deviation, the heavier the regularization penalty. This regularization term improves the robustness of the scoring results by penalizing extreme values ​​that deviate from the mean.

[0080] S66: Calculate the overall score By combining the normalized value and inverse distance weights, and subtracting the regularization term (to eliminate unstable factors), the final score is obtained: Score S i The physical meaning: Small indicators achieve high contributions through "large normalized values ​​+ high weights", while the regularization term ensures that the score is not distorted by local outliers, ultimately resulting in S. i The larger the value, the better the solution.

[0081] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A start-up optimization control method for large-scale hydropower units based on full-view simulation and dynamic verification, characterized in that, Includes the following steps: S1: For the target unit, the system executes multiple preset start-up control modes, simultaneously collects structural dynamics data, vibration data, stability data, thermodynamic data, and hydraulic data, establishes a benchmark database, and quantitatively evaluates the advantages and disadvantages of each start-up mode; S2: Based on the method of characteristics, establish a one-dimensional simulation model of the unit's hydraulic-mechanical-electric system, calculate the macroscopic dynamic response of the power generation system under different start-up conditions, define the safe operation boundary, and provide time-varying boundary conditions; S3: Based on the time-varying boundary conditions in step S2, establish a three-dimensional full-channel high-fidelity CFD model containing all flow components, conduct transient unsteady flow field simulation, and reveal the complex internal flow phenomena and excitation sources. S4: Combining the simulation results of steps S2 and S3, the start-up curve is adjusted for the strong vibration flow region. By optimizing acceleration, open-loop holding segment, and open-loop / closed-loop switching point, a multi-segment variable parameter start-up optimization curve is generated. S5: Apply the optimized startup rules from step S4 to real machine testing, compare the test results with the simulation prediction results, and solidify the rules if the consistency meets the standard; otherwise, modify the model parameters and iteratively optimize. S6: Based on the Riemannian manifold, a multi-index comprehensive scoring system is constructed to calculate the comprehensive score of each boot method and select the optimal boot method.

2. The method according to claim 1, characterized in that, Step S1 specifically includes the following sub-steps: S11: Select the start-up method, including guide vane large start, guide vane small start, variable acceleration control, and open-loop and closed-loop combined start; S12: Arrange the start-up test measurement points, including unit vibration measurement points, swing measurement points, water pressure pulsation measurement points, strain measurement points, stress measurement points, and speed measurement points; S13: Conduct a start-up test at the selected head and record the changes in guide vane opening and data at each measuring point over time; S14: Process the acquired data and calculate the peak-to-peak amplitude and stress value of the mixing frequency; S15: Compare and analyze the same data under different boot methods to quantitatively evaluate the advantages and disadvantages of each boot method.

3. The method according to claim 2, characterized in that, In step S14, the peak-to-peak amplitude of the mixing frequency is taken with a 97% confidence level, that is, the probability of counting points in the time-domain waveform is statistically analyzed by partitioning the waveform, and the data in the unreliable region is removed before calculation; the stress is calculated based on the strain values ​​collected during the test using the following formula: ; in, The strain value, E The elastic modulus of the material being tested. For stress.

4. The method according to claim 1, characterized in that, Step S2 specifically includes the following sub-steps: S21: Construct a mathematical model of a hydraulic system based on the method of characteristics. The hydraulic system includes a pressure pipeline, a turbine, and a governor. S22: Conduct simulation calculations under different start-up modes, analyze the time-domain changes of unit speed, volute pressure, and tailrace pipe pressure, and propose an initial optimized start-up mode.

5. The method according to claim 4, characterized in that, In step S21, the mathematical model of the pressure pipeline is established using the following formula: ; ; In the formula, the subscript i Let be any grid intersection point in the x-direction; Let P be the head of the water. Let P be the flow rate. , , , It is a moment The known quantities are obtained using the following formula: ; In the formula, constant B and R The following formula is used to obtain it: ; ; In the formula, A The cross-sectional area of ​​the pipeline is in meters. 2 ; g The acceleration due to gravity is m / s². 2 ; f This refers to the Darcy-Weisbach coefficient for hydraulic losses along the friction path. D Pipe diameter, in meters (m); For a long L The pipes are divided into equal parts n The length of each subsequent segment, in meters; a The water hammer wave velocity, in m / s, is obtained using the following formula: ; In the formula, K Let be the bulk modulus of water; The density of water; E The elastic modulus of the pipe is expressed in N / m. 2 ; D The inner diameter of the pipe is in meters (m). e Let be the pipe wall thickness, in meters (m).

6. The method according to claim 4, characterized in that, In step S21, the mathematical model of the water turbine is established using the following formula: ; ; ; ; ; In the formula, It is a flow function; It is a torque function; This represents the relative value of the guide vane opening. For guide vane opening; This is a relative value of the water head; This is a relative value of the flow rate; This is a relative value of the rotational speed; This is a relative value of the torque; This refers to the relative angular velocity deviation. This represents the relative value of the instantaneous shaft torque of the water turbine; This represents the relative value of electromagnetic resistance. This is the rated torque; Let be the unit's inertial time constant, in seconds; For flywheel torque, t·m 2 ; Rated power, kW; ; In the above formula, the subscript r represents the rated parameter value.

7. The method according to claim 4, characterized in that, In step 21, the mathematical model of the speed regulator is established using the following formula: ; ; ; In the formula, This represents the relative value of the output signal deviation of the speed controller. This represents the relative value of the unit's speed deviation; the subscript 0 indicates the initial value. This is the permanent slip coefficient, and its value is selected in the range of 0 to 0.1; This refers to the reaction time of the guide vane relay. This is the relative value of the guide vane servo motor's stroke deviation, representing the servo motor's stroke y-deviation relative to its maximum stroke. The ratio; The relative value of the main pressure regulating valve stroke deviation represents the ratio of the main pressure regulating valve's stroke from the center position to its maximum stroke, while the maximum stroke refers to the main pressure regulating valve stroke when the input speed signal deviation is 100%. This refers to the proportional gain of the speed controller; The integral gain of the speed controller; The differential gain of the speed controller; The following formula is used to obtain it: ; ; ; In the formula, This refers to the transient slip coefficient of the speed governor; This is the buffer time constant of the speed controller; This is the differential time constant or acceleration time constant of the speed controller.

8. The method according to claim 1, characterized in that, Step S3 specifically includes the following sub-steps: S31: Establish a full-basin geometric model of the unit's volute, fixed guide vanes, movable guide vanes, runner, and tailrace pipe; S32: Mesh the geometric model; S33: Set boundary conditions, using SST. k-ω The turbulence model uses the closed Navier-Stokes equations. The total pressure is set at the inlet of the volute and the static pressure is set at the outlet of the tailrace pipe. The steady-state calculation of the runner adopts the rotating coordinate system method and the transient calculation adopts the sliding mesh method. S34: Arrange pressure pulsation measuring points near the impeller, volute, fixed guide vanes, movable guide vanes, bladeless area, and tailrace pipe; S35: Flow field evolution analysis for typical calculation conditions with different start-up patterns; S36: Comparative analysis of the pressure pulsation evolution of each flow component under different start-up conditions; S37: Initial screening for the optimal boot method.

9. The method according to claim 1, characterized in that, Step S31 specifically includes the following sub-steps: S311: Organize the component structural data of the turbine casing, fixed guide vanes, movable guide vanes and tailrace pipe, and use SolidWorks modeling software to create a three-dimensional geometric model; S312: After obtaining point cloud data through on-site 3D scanning, the 3D geometric model of the wheel is obtained through reverse modeling; S313: Directly establish the three-dimensional geometric structure of the watershed for the volute and tailrace pipe, obtain the corresponding watershed for the impeller and guide vanes through Boolean operations, and assemble the components to form a full watershed geometric model.

10. The method according to claim 1, characterized in that, In step S5, if the experimental results match the simulation prediction results by a factor greater than or equal to the preset threshold and the optimization objective is achieved, the startup rule is solidified; otherwise, the experimental data is fed back to the one-dimensional and three-dimensional simulation models, the parameters are corrected, and steps S2-S5 are repeated until the optimal startup rule that meets the safety and stability requirements is obtained.

Images