A remote comprehensive control system for a hydroelectric generator set

Abstract

The application is a kind of remote comprehensive control system of water turbine generator unit, and relates to the technical field of water turbine generator unit control, comprising: obtaining adjustment factor, and adjusting step length tentatively according to adjustment factor, and combining with water turbine efficiency change to dynamically decide adjustment direction and step length. In the application, through multi-dimensional data acquisition, dynamic correction and self-adaptive optimization, precise adaptation and efficiency deep mining under all working conditions are realized; efficiency calculation introduces multi-dimensional correction of water temperature, mechanical loss and cavitation, which is in line with actual operating conditions, and greatly reduces the error of fixed parameters; optimization control dynamically adjusts step length and strategy according to steady-state deviation, significant change and equipment life cycle state, from strong gain large step exploration to negative gain reverse regulation, which adapts to complex scenes; steady-state determination and convergence control ensure efficient approximation of optimal solution, so that the unit can continuously mine power generation efficiency after water head fluctuation, load change, new machine running-in and aging maintenance.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of water turbine generator set control, and particularly relates to a water turbine generator set remote comprehensive control system. BACKGROUND

[0002] In the field of hydropower energy, as the core power generation equipment, the operation efficiency and control accuracy of the water turbine generator set directly determine the energy conversion efficiency and power grid power supply stability.

[0003] The current mainstream water turbine generator set control scheme mainly depends on the fixed coordination curve (corresponding relationship of water head, load and blade angle, guide vane opening based on design conditions) preset by the manufacturer, but in actual operation, the unit faces complex conditions such as water head fluctuation (such as upstream and downstream water level difference change), load dynamic adjustment (power grid dispatching demand), equipment aging and wear (cavitation of flow components, bearing loss) and seasonal hydrological change (difference in sediment content and flow between dry season and flood season), etc. The fixed coordination curve is difficult to adapt in real time, resulting in long-term operation of the unit in a non-optimal efficiency range, and energy waste is common.

[0004] In addition, the traditional control scheme lacks full life cycle adaptation capability, and cannot dynamically update control parameters in scenes such as new machine commissioning, equipment maintenance and property change. In addition, the cross-station collaborative scheduling capability is weak, and it is difficult to meet the demand of intelligent power grid for fine adjustment of hydropower resources.

[0005] Therefore, a water turbine generator set remote comprehensive control system is needed to solve the above-mentioned problems. SUMMARY

[0006] The purpose of the present application is to solve the above-mentioned problems, and a water turbine generator set remote comprehensive control system is proposed.

[0007] In order to achieve the above-mentioned purpose, the present application adopts the following technical scheme:

[0008] A water turbine generator set remote comprehensive control system, comprising:

[0009] Obtain comprehensive data of the water turbine operation state and perform corresponding processing;

[0010] Based on the obtained data, dynamically calculate the water turbine efficiency, when the start condition is met, perform optimization adjustment with a preset initial step size, and obtain the water turbine efficiency before and after adjustment;

[0011] Obtain an adjustment factor, and adjust the step size tentatively according to the adjustment factor, dynamically decide the adjustment direction and step size in combination with the water turbine efficiency change, and hibernate after converging to a local optimal efficiency point;

[0012] The optimal parameter is stored in the relational database, and the optimal parameter is provided for the initial parameter of the optimal control by self-learning iteration adaptation to the working condition change, so as to realize the whole life cycle efficiency optimization.

[0013] Preferably, the water turbine operation state comprehensive data is acquired and processed accordingly, and specifically includes:

[0014] The water turbine operation state comprehensive data includes water head data, guide vane opening data, paddle angle data, output power data and flow data, and the data is preprocessed including abnormal value filtering, missing value filling and time stamp synchronization.

[0015] Preferably, the water turbine efficiency calculation basic formula is: , wherein, is the water density, is the gravity acceleration, is the flow, is the water head, is the output power.

[0016] Based on the above formula, dynamic correction including water temperature correction, mechanical loss correction and cavitation correction is carried out to obtain the corrected formula:

[0017]

[0018] , wherein, is the mechanical efficiency, is the cavitation correction coefficient.

[0019] Preferably, the method further comprises:

[0020] When the following conditions are met simultaneously, the optimal adjustment is carried out with a preset initial step size:

[0021] Power fluctuation ≤ ± 0.5% rated power, water head fluctuation ≤ ± 0.1m, duration ≥ 5 minutes; all sensor data is fault-free and has no missing value, the efficiency calculation value is in a reasonable range; the unit vibration, shaft temperature and cavitation noise are within the safety threshold;

[0022] The preset coordinated parameter corresponding to the current water head and load is called to control the speed regulator and paddle servo mechanism to adjust the unit to the initial point, and the initial efficiency is recorded .

[0023] After the optimal adjustment with the preset initial step size, the optimal step size is dynamically adjusted according to the adjustment factor.

[0024] Preferably, the adjustment factor is obtained as follows:

[0025] Based on the corrected formula for calculating the efficiency of the water turbine, the efficiency before and after the initial step size adjustment is obtained, and the adjusted efficiency Efficiency before adjustment Adjustment factor .

[0026] Preferably, the construction step size Relationship function with adjustment factor :

[0027]

[0028] Wherein,

[0029] is the operating condition correction coefficient; is the basic step size; is the strong gain threshold; is the medium gain threshold; is the weak gain upper limit; is the weak gain lower limit; is the minimum step size; is the maximum step size.

[0030] Preferably, the method further comprises:

[0031] After dynamic adjustment according to the adjustment factor, wait After, the power and flow fluctuation are monitored in real time during the period, and when the fluctuation and duration meet the preset requirements, it is determined as a new steady state;

[0032] Collecting , , Data under the new steady state, calculating , and recording adjustment time, step size, rotation angle and other information, and storing them in the optimization log; and based on Decision next move;

[0033] When the efficiency fluctuation after fine-tuning in both directions meets the preset requirements, and the verification results are consistent for a continuous preset number of times, it is determined as local optimal convergence, and the current rotation angle As a stable operating point.

[0034] Preferably, the relationship database stores optimal parameters, and iteratively adapts to changes in operating conditions through self-learning, to provide initial parameters for optimization control, and specifically includes the following contents:

[0035] The relationship database is divided into a basic layer, a dynamic layer and a prediction layer;

[0036] The optimal parameters obtained by new optimization need to meet the preset requirements to be stored in the database; and a confidence label is added to each group of data;

[0037] Data cleaning and model training are automatically performed every morning, and data exceeding the preset duration in the dynamic layer are excluded, and the prediction layer model is retrained.

[0038] A coordination relationship optimization report is generated monthly to compare the efficiency improvement range of the dynamic layer and the basic layer, and to provide equipment state evaluation basis for operation and maintenance.

[0039] When the optimization control is started, the optimal parameters of the dynamic layer under the same working condition are preferentially called as initial points.

[0040] Preferably, the three-level coordination architecture of local, remote and cloud is built to realize unit operation state visualization monitoring, remote control instruction issuing, cross-station data interaction, support remote intervention of operation and maintenance personnel and power grid dispatching cooperation.

[0041] As described above, the application has the following advantages:

[0042] 1. The application realizes precise adaptation and efficiency deep mining under all working conditions through multi-dimensional data acquisition, dynamic correction and self-adaptive optimization; the efficiency calculation introduces multi-dimensional correction of water temperature, mechanical loss and cavitation, which is in line with the actual operation condition, and greatly reduces the error of fixed parameters; the optimization control dynamically adjusts the step and strategy according to the steady-state deviation, significant change and equipment life cycle state, from strong gain large step exploration to negative gain reverse regulation, which adapts to complex scenarios; the steady-state determination and convergence control ensure efficient approximation of the optimal solution, so that the unit can continuously mine power generation efficiency under water head fluctuation, load change, new machine running-in and aging maintenance.

[0043] 2. The application realizes intelligent operation and maintenance of the unit throughout its life cycle by building a basic layer, a dynamic layer and a prediction layer coordination relationship library combined with a self-learning iteration mechanism; the basic layer relies on the factory curve to ensure initial control, the dynamic layer stores real-time optimal parameters and updates them according to the confidence level, and the prediction layer uses the random forest algorithm to predict the working condition, so that the new machine can quickly run-in and the aging unit can adapt to new parameters without significantly modifying the control logic, reducing operation and maintenance costs; at the same time, the three-level coordination architecture connects local, remote and cloud data interaction, visual monitoring presents the unit state in real time, encrypted communication guarantees cross-station dispatching, edge cloud fusion realizes regional power grid optimization, early fault warning and system unattended upgrade, helping operation and maintenance to shift from passive response to active prediction and from single unit control to multi-station cooperation, improving the overall operation efficiency of the hydropower plant and the adaptability of power grid dispatching, and providing key technical support for the construction of smart grid. BRIEF DESCRIPTION OF DRAWINGS

[0044] In the following description of exemplary embodiments in conjunction with the drawings, more details, features and advantages of the application are disclosed, in which:

[0045] Figure 1 The figure is a system structure diagram of the application. DETAILED DESCRIPTION

[0046] Several embodiments of the application will be described in greater detail below, with reference made to the figures. The application can take many different forms and embodiments and should not be limited to the embodiments set forth herein. These embodiments are provided so that this disclosure will be thorough and complete, and fully convey the scope of the application to those skilled in the art. The embodiments are not intended to limit the application.

[0047] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. It will be further understood that terms, such as those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the relevant art and / or the

[0048] Embodiment 1

[0049] DETAILED DESCRIPTION Figure 1 The detailed description is described with reference to the accompanying figures.

[0050] The detailed description is described with reference to the accompanying figures. Figure 1 A structural block diagram of a remote comprehensive control system for a hydroelectric generator set is provided, which shows the connection relationship between obtaining comprehensive data of the operating state of the hydroelectric generator set and building a local, remote, and cloud three-level collaborative architecture, and labels the main function interaction processes of each module.

[0051] In this embodiment, it includes:

[0052] Obtaining comprehensive data of the operating state of the hydroelectric generator set and performing corresponding processing;

[0053] Specifically, it includes:

[0054] The comprehensive data of the operating state of the hydroelectric generator set includes water head data, guide vane opening data, paddle angle data, output power data, and flow data; and the data is preprocessed including abnormal value filtering, missing value filling, and time stamp synchronization to filter interference and unify standards.

[0055] The detailed obtaining process of the comprehensive data of the operating state of the hydroelectric generator set is as follows:

[0056] Water head data: Collect water level data through upstream (before the dam) and downstream (tailwater) double water level meters, and calculate the net water head ( The water head loss caused by pipeline friction and local resistance); The water level data is collected by radar type water level meter (measurement range 0-50 m, accuracy ± 0.01 m, anti-water flow fluctuation interference), or differential pressure type water level meter (suitable for closed pipeline scene), avoiding the use of float type water level meter (easily blocked by sediment); The installation position of the water level meter avoids the vortex area of the water flow and the disturbance area of the water inlet, ensuring the stability of the measurement environment; Among them, the guide vane opening and the flow are collected in real time, and the water head loss and working condition characteristic curve provided by the manufacturer is called to dynamically calculate , to ensure that the corrected water head error remains within the preset range.

[0057] The guide vane opening data: the opening feedback is obtained through bus communication (such as Profibus-DP, Modbus-TCP) with the speed regulator, and the actual displacement of the guide vane is directly measured by a magnetostrictive displacement sensor, based on double verification, to obtain more accurate data; Among them, under normal working conditions, the speed regulator feedback data is the main one, and the displacement sensor data is the auxiliary one; When the deviation between the two exceeds 0.5% (such as the speed regulator showing 50% and the sensor actually measuring 49.5%), mechanical zero point calibration is triggered: control the guide vane to close to the mechanical zero position, reset the opening reference of the sensor and the speed regulator, and eliminate the mechanical drift of long-term operation.

[0058] The paddle angle data: the paddle angle is directly measured by an absolute value encoder, and the output torque of the paddle servo mechanism (through a torque sensor) is collected, to build a correlation model of the angle and the torque; If the deviation between the angle command and the actual feedback exceeds 0.2°, and the servo mechanism torque exceeds 120% of the rated value, it is determined that the mechanical is blocked, and the optimization is immediately stopped, triggering an alarm; Automatic execution of small range angle calibration (such as fine tuning from the current angle ± 0.5°) at a preset time interval (such as every hour) to verify the accuracy of the angle feedback.

[0059] The output power data: the active power, reactive power and power factor are collected by a three-phase power transmitter on the stator side of the generator, and the active power is finally used as the basis for efficiency calculation; Real-time collection of power factor, correction of measured power The rated power factor of the generator is usually 0.85, and the transmitter is calibrated by a standard power source every quarter to ensure that the long-term measurement error is less than the preset threshold.

[0060] The flow data: the time difference method ultrasonic flowmeter (range 0-200 m³ / s, accuracy ± 1.0%, installed in the water turbine inlet pipe section) is used to support full pipe and non-full pipe working conditions self-adaptation, avoiding the wear of the sensor by sediment.

[0061] The data preprocessing process specifically includes:

[0062] ​Outlier filtering: A dual verification method using the 3σ principle and physical constraints is employed. Specifically, this involves first removing data that deviates from the mean by three times the standard deviation, and then filtering values ​​that do not conform to physical logic (such as negative flow rate or head fluctuation exceeding ±5%).

[0063] Missing value imputation: Short-term missing values ​​(<5s) are imputed by linear interpolation, and long-term missing values ​​(≥5s) are imputed by using backup sensor data or soft measurement results;

[0064] Timestamp synchronization: All parameters are stamped with a unified timestamp through the GPS / BeiDou time synchronization module (time synchronization error ≤1ms) to avoid efficiency calculation deviations caused by data timing misalignment (such as matching the power at time t1 with the flow rate at time t2).

[0065] Based on the acquired data, the turbine efficiency is dynamically calculated. When the start-up conditions are met, the turbine is optimized and adjusted with a preset initial step size (the step size is the blade angle control amplitude), and the turbine efficiency before and after adjustment is obtained.

[0066] The basic formula for calculating the efficiency of a water turbine is: ,in, For water density, For gravitational acceleration, For traffic, For water head, This refers to the output power.

[0067] Based on the above formula, dynamic corrections including water temperature correction, mechanical loss correction, and cavitation correction are performed to obtain the corrected formula:

[0068]

[0069] in, For mechanical efficiency, This is the cavitation correction factor.

[0070] Water temperature correction: Real-time water temperature (T) is collected and corrected. (Density is highest at water temperature of 4℃) Dynamic adjustment Avoid fixed values ​​(1000) Errors caused by )

[0071] Mechanical loss correction: Introducing mechanical efficiency (Obtained through the manufacturer's characteristic curves, such as after bearing wear) (From 0.98 to 0.96), the corrected formula is: ;

[0072] Cavitation correction: Real-time monitoring of cavitation noise in the turbine draft tube (using acoustic sensors). When the noise exceeds a threshold (e.g., 85 dB), a cavitation correction coefficient is introduced. (When cavitation is slight) , severe ), the final formula is obtained: .

[0073] The calculated results are checked, specifically including:

[0074] If the current efficiency deviates from the historical optimal efficiency under the same conditions by more than 5%, the sensor verification process is triggered; if the efficiency is lower than the design value by 10%, the unit is determined to be abnormal.

[0075] A 10s time window is set, and the efficiency change rate in the window is calculated;

[0076] If > 0.05% / s, it is predicted that the efficiency will continue to rise, providing an accelerated adjustment reference for optimization control, and if < -0.05% / s, it is predicted that the efficiency will decrease, triggering an early reverse warning.

[0077] Not all conditions require continuous optimization (to avoid excessive adjustment and increase losses), the following scenarios are the trigger time for optimization:

[0078] I. Steady state but deviates from the design curve:

[0079] The unit power, water head, vibration and other parameters are stable (satisfy the start condition), but the operating efficiency is significantly lower than the design value (such as due to equipment aging, the actual efficiency is more than 2% lower than the corresponding efficiency of the factory curve).

[0080] The internal reason is that the fixed curve has been mismatched, and the true optimal solution under the current condition needs to be found through optimization to correct the coordination relationship.

[0081] II. Significant changes in working conditions:

[0082] Scenario 1: Water head fluctuates greatly (such as a change in water level difference between upstream and downstream of more than 5%): After the potential energy of the water flow changes, the optimal angle of the original coordination curve is no longer applicable, and the new water head needs to be re-optimized.

[0083] Scenario 2: Load changes continuously (such as grid dispatching requires the unit load to increase from 50% to 80% within 30 minutes): Load changes will change the flow characteristics of the water turbine, and optimization can quickly find the optimal blade angle under the new load to avoid low efficiency operation.

[0084] Scenario 3: Seasonal / hydrological cycle changes (such as from dry season to flood season, seasonal fluctuations in sediment content): When the long-term working condition changes, the fixed curve cannot be adapted, and the coordination relationship needs to be periodically optimized (such as monthly / quarterly full working condition scanning).

[0085] III. Changes in equipment state (full life cycle adaptation):

[0086] Scenario 1: New machine commissioning: The surface of the flow passage components is rough at the initial stage of the new machine (e.g. the optimal water film has not been formed), and the efficiency is lower than the design value. By optimization, the coordination connection can be dynamically adjusted to accelerate the process of running-in and enter the high efficiency interval in advance.

[0087] Scenario 2: After device aging / maintenance: After the wear of flow passage components (e.g. runner cavitation, guide vane seal aging) or the replacement of components (e.g. replacement of runner, repair of shafting), the original coordination connection curve is invalid, and a new adaptive relationship needs to be established by optimization.

[0088] To avoid excessive regulation and increase losses, the following situations suspend optimization and return to fixed coordination connection curve:

[0089] Unstable working condition: power and water head fluctuation exceeds threshold value (e.g. power fluctuation > ± 1% rated power), data distortion, optimization is meaningless.

[0090] Device anomaly: vibration, shaft temperature, cavitation noise exceed safety threshold, prioritize device safety, suspend optimization.

[0091] Sleep period: after optimization is completed, enter sleep (e.g. 10-30 minutes) to avoid repeated adjustment in a short time (when the working condition does not change significantly, repeated optimization will increase mechanical loss).

[0092] When the following conditions are met simultaneously, optimization adjustment is performed with a preset initial step size:

[0093] Power fluctuation ≤ ± 0.5% rated power, water head fluctuation ≤ ± 0.1m, duration ≥ 5 minutes; all sensor data is fault-free and has no missing, efficiency calculation value is within a reasonable range; unit vibration (≤ 0.1mm / s), shaft temperature (≤ 65℃), cavitation noise (≤ 80dB) are within the safety threshold;

[0094] Call the preset coordination connection parameters corresponding to the current water head and load, control the speed regulator and paddle servo mechanism to adjust the unit to the initial point, and record the initial efficiency ;

[0095] After optimization adjustment with a preset initial step size, the subsequent optimization step size is dynamically adjusted according to the adjustment factor.

[0096] The requirement "power fluctuation ≤ ± 0.5% rated power, water head fluctuation ≤ ± 0.1m and duration ≥ 5 minutes" ensures that optimization is started when the working condition is stable. If the working condition fluctuates frequently (e.g. power fluctuates high and low, water head fluctuates greatly), adjusting the paddle angle at this time not only makes it difficult to find the true optimal solution, but also may exacerbate unit vibration and damage the device. By strict stable condition, let optimization have a clear goal and avoid doing useless work.

[0097] For example, if the water head fluctuates greatly, the potential energy of the water flow is unstable, and forcibly adjusting the blade angle may cause the flow characteristics of the hydraulic turbine to be disordered, thereby reducing the efficiency and damaging the equipment.

[0098] "all sensor data without failure, no missing, efficiency calculation value reasonable", to ensure the reliability of data. Sensor failure (such as water level meter blockage, displacement sensor drift) or data missing, efficiency calculation, adjustment direction may be wrong, based on such error data for optimization, the result must deviate from the optimal, even cause unit abnormal.

[0099] Combined with outlier filtering and missing value filling, the logic of data verification, repair and activation is formed to ensure the accuracy of optimization regulation from the source;

[0100] The vibration of the unit is limited to ≤0.1mm / s, the shaft temperature is ≤65℃, and the cavitation noise is ≤80dB to ensure the adjustment in a safe state of the equipment. If the vibration is too large and the shaft temperature is too high, it indicates that the unit may have mechanical failure (such as bearing wear) or abnormal operation (such as severe cavitation), at which time the optimization regulation may exacerbate the damage to the equipment. By setting safety thresholds, the optimization control is limited to a safe range, giving priority to the service life and safe operation of the unit.

[0101] All start-up conditions (power fluctuation, water head fluctuation, sensor status, equipment status) are clear quantitative indicators (such as 0.5% rated power, 0.1m water head), which can be directly read and judged by the operation and maintenance personnel or the control system without complex analysis. In engineering applications, conditions that can be seen, measured and judged are easier to implement and reduce the technical threshold;

[0102] When a new machine is put into operation, it can quickly enter high-efficiency grinding based on pre-set coordination parameters; after the equipment is aged, as long as the steady-state and safety conditions are met, the optimization can still be started to adapt to the new optimal parameters. This design allows the system to run throughout the life cycle of the unit, without the need to significantly modify the control logic due to changes in equipment status (such as wear and tear after maintenance), reducing the operation and maintenance costs.

[0103] Get the adjustment factor and adjust the step size based on the adjustment factor. Combine the water turbine efficiency change to dynamically decide the adjustment direction and step size, and sleep after converging to the local optimal efficiency point;

[0104] The process of obtaining the adjustment factor is as follows:

[0105] Based on the corrected formula for calculating the efficiency of the water turbine, the efficiency before and after the initial step size adjustment is obtained, and the efficiency after adjustment is divided by the efficiency before adjustment to get the adjustment factor .

[0106] Construct the step size Step size (actual step size is the control amplitude of the blade angle, including direction, where positive is the original adjustment direction, and negative is the reverse) and adjustment factor Relationship function:

[0107]

[0108] Wherein,

[0109] is the working condition correction coefficient (adaptation head , load difference, such as high head working condition K sensitivity is higher, the step size response needs to be enlarged);

[0110] is the basic step size (initial reference step size, such as 0.5°, preset according to the mechanical accuracy of the unit);

[0111] is the strong gain threshold; Indicates that the adjustment efficiency is significantly improved (such as ≥3%), which needs to be accelerated to explore with large step size;

[0112] is the medium gain threshold; Indicates that the efficiency is steadily improved (1%~3%), and the step size is kept medium;

[0113] is the weak gain upper limit; Indicates that the efficiency is improved gently (≤1%), which needs to be reduced;

[0114] is the weak gain lower limit (negative gain threshold); Indicates that the efficiency is decreased (≤-1%), which needs to be adjusted in the opposite direction;

[0115] is the minimum step size; used to avoid small step size leading to adjustment shock or response lag;

[0116] is the maximum step size; used to avoid large step size damaging the unit stability (such as exceeding the water flow response capability);

[0117] Wherein, is the key to adapt to the whole working condition, which needs to be dynamically calculated according to the influence of head and load on efficiency sensitivity, and the calculation method is as follows:

[0118] Based on the water turbine characteristic curve provided by the manufacturer, the empirical formula: ;

[0119] Wherein, , These are the rated head and rated power of the water turbine, respectively.

[0120] , , These are the preset coefficients.

[0121] By constructing step size With regulatory factors The relationship function, based on efficiency changes (e.g.) (Reflecting the efficiency improvement) precisely adjusts the blade angle control range. Under high head conditions, the operating condition correction coefficient is utilized. Increasing the step size response can quickly capture the optimal efficiency point; the efficiency improvement is gradual. By reducing the step size at the same time, we can avoid missing the extreme value and enable efficient optimization under different water head and load conditions, thereby improving the overall power generation efficiency of the unit.

[0122] From strong gain to negative gain, the segmented step size strategy adapts to all scenarios such as steady-state deviation and sudden changes in operating conditions (head fluctuation, load change, etc.), solves the fixed curve mismatch problem, and ensures that the unit continues to approach the optimal efficiency under complex operating conditions.

[0123] Set minimum step size (Avoid adjusting oscillations and response lag) and maximum step size (To prevent disruption of unit steady state), the adjustment range is constrained within a safe and effective range to reduce mechanical wear and extend equipment life; when efficiency declines ( Timely reverse adjustments can prevent inefficient or even damaging operation.

[0124] By combining efficiency verification (such as triggering sensor verification when the deviation from the historical best exceeds 5%) and operating condition stability judgment (optimization is only sought when power and head fluctuations meet the conditions), we can avoid data distortion and blind adjustments when the operating conditions are unstable, thus ensuring the safe operation of the unit and reducing the risk of failure.

[0125] After dynamic adjustment based on the regulating factor, wait back( Based on the unit's inertia setting (40s for mixed-flow turbines and 60s for axial-flow turbines), power and flow fluctuations are monitored in real time. When both the fluctuation and duration meet the preset requirements (fluctuation ≤ ±0.2% and lasts for 10s), it is determined to be a new steady state.

[0126] Acquiring new steady-state data , , Data, calculated Simultaneously, it records information such as adjustment time, step size, and turning angle, and stores it in the optimization log;

[0127] And based on Decide the next move, including:

[0128] If (Efficiency significantly improved): Confirm the correct direction, update the reference point ( ), continue to fine-tune in the same direction with the current step;

[0129] If (Efficiency has no significant change): Step size contraction 50% (such as 0.5° to 0.25°), continue to explore in the same direction, avoid missing the optimal area;

[0130] If (Efficiency decreased): Immediately reverse the adjustment ( ), repeat the steady-state waiting and efficiency evaluation, if the efficiency still decreases after reversing, determine the current point as the local optimal point;

[0131] Substitute into the relationship function of step size and adjustment factor , dynamically calculate the adjustment step size of next time, balance the optimization speed and accuracy;

[0132] When the forward and reverse fine-tuning (minimum step size 0.1°) is completed, the efficiency fluctuation meets the preset requirements (for example, ≤0.05%), and the consecutive preset number of verification results (3 times) is consistent, it is determined as the local optimal convergence, and the current turning angle as the stable running point;

[0133] The dormancy period is 15-30 minutes by default, which can be remotely configured, and only monitors the working condition changes during the period without executing optimization;

[0134] When the water head fluctuation exceeds ±5%, the load change exceeds ±10% of the rated power, or the dormancy period ends, the optimization process is automatically restarted;

[0135] The blade turning angle is limited in the safety range given by the manufacturer (such as -15°~+30°), if the adjustment instruction exceeds the range, it is automatically truncated and an alarm is given;

[0136] If vibration exceeds limit, shaft temperature rises suddenly, cavitation alarm and other abnormalities occur during optimization, stop optimization immediately, and restore the blade turning angle to the last stable running point.

[0137] Build a relationship database to store optimal parameters, adapt to working condition changes through self-learning iteration, provide initial parameters for optimization control, and realize whole life cycle efficiency optimization;

[0138] Specifically includes the following content:

[0139] The relationship database construction is divided into basic layer, dynamic layer and prediction layer;

[0140] The basic layer (factory curve): stores the three-dimensional association table of water head, load, guide vane opening and paddle angle provided by the manufacturer as the initial control basis (for example, when H=50 m, P=80% rated power, G=75%, B=12°);

[0141] The dynamic layer (real-time optimal data): stores the optimal parameters obtained by optimization according to water head intervals (every 5 m) and load intervals (every 10% rated power), and at least stores 3 sets of effective data in each interval to ensure statistical reliability;

[0142] The prediction layer (working condition prediction data): based on the random forest algorithm, the historical data of the basic layer and the real-time data of the dynamic layer are fused to train a working condition and optimal parameter prediction model, and the optimal parameters for future possible working conditions (such as H=70 m in flood period and H=30 m in dry period) are predicted in advance to accelerate the optimization convergence.

[0143] The new optimal parameters obtained by optimization need to meet the preset requirements (for example, the efficiency is higher than the basic layer data by more than 1% under the same working condition, and the deviation of continuous 2 optimization results is less than or equal to 0.2°) to be stored; and a confidence label is added to each set of data, the more the number of optimization times and the higher the data consistency, the higher the confidence (for example, if the same working condition is optimized 5 times, the confidence is 95%);

[0144] When the new data of the dynamic layer conflicts with the data of the basic layer, if the confidence of the dynamic layer data is greater than or equal to 90%, the parameters of the corresponding working condition of the basic layer are automatically replaced;

[0145] If the confidence is 80%-90%, it is marked as to be verified, and the next optimization of the working condition is verified, and after 2 consecutive verifications, it is replaced;

[0146] If the confidence is less than 80%, it is determined as abnormal data and is not stored and triggers manual review.

[0147] Data cleaning and model training are automatically performed every morning, and old data (device aging leading to parameter failure) exceeding the preset time (for example, 1 year) in the dynamic layer is removed, and the prediction layer model is retrained;

[0148] An association relationship optimization report is generated every month to compare the efficiency improvement amplitude of the dynamic layer and the basic layer (for example, an average improvement of 1.8%), which provides a device state evaluation basis for operation and maintenance;

[0149] When the optimization control is started, the optimal parameters of the same working condition in the dynamic layer (confidence greater than or equal to 90%) are preferentially called as the initial point; if not, the prediction layer data is called; if the prediction layer has no data, the basic layer data is called;

[0150] The working condition matching rule adopts the nearest neighbor matching. When the current working condition (H=52 m, P=75%) has no completely matched interval, the average parameters of the H=50-55 m, P=70%-80% interval are called to ensure that the initial point is close to the optimal point.

[0151] A three-level collaborative architecture of local, remote, and cloud is built to realize the visualization monitoring of unit operation state, the issuance of remote control instructions, and the cross-station data interaction, supporting remote intervention of operation and maintenance personnel and coordination of power grid dispatching;

[0152] Specifically includes:

[0153] 20+ key parameters such as head, guide vane opening, and efficiency are displayed in the form of instrument panel, trend chart, and flow chart, and automatic color change alarm is supported when the parameters exceed the limit (such as shaft temperature exceeding 65°C);

[0154] The optimization phase (initialization / exploration / convergence / sleep), current step size, and efficiency change curve are dynamically displayed, and manual pause / restart of optimization is supported;

[0155] Operation and maintenance personnel can manually adjust the guide vane opening and paddle angle (permission classification, password verification required) through the operation panel, and can trigger the local manual priority mode to cut off remote control in emergency situations;

[0156] An encrypted connection with the unit LCU (local control unit) is established through a 5G private network / VPN (virtual private network), with a transmission rate of ≥100 Mbps and a delay of ≤200 ms;

[0157] The running state of 10+ units is monitored at the same time, and hierarchical queries by station, unit, parameter, and multi-unit efficiency comparison report are supported; optimization parameter configuration (such as step size, sleep period) and start / stop optimization instructions are issued, and one-key batch start / stop of units in the same region is supported;

[0158] A hybrid architecture of edge computing combined with cloud computing is adopted: real-time data preprocessing (such as outlier filtering, efficiency calculation) is performed on the edge side (local unit), only key data (such as optimal efficiency, alarm information) is uploaded to the cloud, and the bandwidth pressure is reduced; to achieve:

[0159] Regional power grid optimization: gather multi-station unit data, based on efficiency and load model, provide optimal power distribution scheme (such as preferentially scheduling units with efficiency ≥90% to full load) for power grid dispatching;

[0160] Fault warning and diagnosis: based on big data analysis, identify potential unit faults (such as bearing wear causing slow efficiency decline), and push early warning information 7-14 days in advance;

[0161] System upgrade and maintenance: support remote push of algorithm patches and coordination relationship database update packages to realize unattended system upgrade.

[0162] The above formulas are all dimensionless values calculated, the formulas are obtained by collecting a large amount of data to simulate a formula of the most recent real situation, and the preset parameters in the formulas are set by a person skilled in the art according to actual conditions.

[0163] The above only describes some exemplary embodiments of the present application by way of illustration, and it is needless to say that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present application. Therefore, the above drawings and descriptions are illustrative in nature and should not be understood as limiting the scope of protection of the claims of the present application.

[0164] It should be noted that, in this document, relational terms such as first and second and the like can be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without limitation, an element preceded by "comprises... a" does not, without more constraints, foreclose the existence of additional identical elements in the process, method, article, or apparatus that comprises the element.

[0165] It should be understood that, in various embodiments of the present application, the size of the sequence number of each process described above does not mean the order of execution, and the execution order of each process should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0166] Those skilled in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. A person skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0167] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the system, device and unit described above can refer to the corresponding process in the foregoing method embodiments, which will not be described here.

[0168] The units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, that is, may be located in one place, or may be distributed to multiple network units. Part or all of the units can be selected to achieve the purpose of the embodiment scheme according to actual needs.

[0169] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit.

[0170] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited to this. Any skilled person in the art can easily think of changes or replacements within the technical range disclosed in the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0171] The above only describes some exemplary embodiments of the present application by way of illustration, and it is needless to say that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present application. Therefore, the above figures and descriptions are illustrative in nature and should not be understood as limiting the scope of protection of the claims of the present application.

Claims

1. A remote integrated control system for a hydro-generator unit, characterized in that, include: Acquire various types of data and process them accordingly; The turbine efficiency is dynamically calculated based on the acquired data. When the start-up conditions are met, the turbine is optimized and adjusted with a preset initial step size, and the turbine efficiency before and after adjustment is obtained. These include: When the following conditions are met simultaneously, optimization adjustment is performed using a preset initial step size: Power fluctuation ≤ ±0.5% of rated power, head fluctuation ≤ ±0.1m, duration ≥ 5 minutes; all sensor data are fault-free and complete, and the calculated efficiency value is within a reasonable range; unit vibration, shaft temperature, and cavitation noise are all within safe thresholds. By calling the preset coordination parameters corresponding to the current head and load, the governor and blade servo mechanism are controlled to adjust the unit to the initial point and record the initial efficiency. ; After optimization adjustment with a preset initial step size, the optimization step size is dynamically adjusted according to the adjustment factor. The process for obtaining the regulatory factor is as follows: Based on the corrected formula for calculating turbine efficiency, the efficiency before and after the initial step size adjustment is obtained, and the adjusted efficiency is... Divide by efficiency before adjustment Obtain the regulating factor ; Build step size With regulatory factors Relational functions: ； in, This is the working condition correction factor; Base step size; Strong gain threshold; The gain threshold is used for medium gain. This is the upper limit of the weak gain. This represents the lower limit of weak gain. Minimum step size; This is the maximum step size; The adjustment factor is obtained, and the step size is adjusted tentatively based on the adjustment factor. The adjustment direction and step size are dynamically determined by combining the changes in turbine efficiency. After converging to a local optimal efficiency point, the system goes into hibernation. A relational database is built to store optimal parameters. Through self-learning and iterative adaptation to changes in operating conditions, initial parameters are provided for optimization control, thereby achieving efficiency optimization throughout the entire life cycle.

2. The remote integrated control system for a hydro-generator unit according to claim 1, characterized in that, Acquire various types of data and perform corresponding processing, specifically including: Various types of data include head data, guide vane opening data, blade angle data, output power data, and flow rate data; and the data undergoes preprocessing including outlier filtering, missing value imputation, and timestamp synchronization.

3. The remote integrated control system for a hydro-generator unit according to claim 1, characterized in that, The basic formula for calculating the efficiency of a water turbine is: ,in, For water density, For gravitational acceleration, For traffic, For water head, This refers to the output power. Based on the above formula, dynamic corrections including water temperature correction, mechanical loss correction, and cavitation correction are performed to obtain the corrected formula: ； in, For mechanical efficiency, This is the cavitation correction factor.

4. The remote integrated control system for a hydro-generator unit according to claim 1, characterized in that, Also includes: After dynamic adjustment based on the regulating factor, wait Subsequently, power and flow fluctuations are monitored in real time. When both the fluctuation and duration meet the preset requirements, a new steady state is determined. Acquiring new steady-state data , , Data, calculated Simultaneously, it records the adjustment time, step size, and turning angle information, and stores them in the optimization log; and based on Decide on the next move; After fine-tuning in both the forward and reverse directions, if the efficiency fluctuations meet the preset requirements and the verification results are consistent for a preset number of consecutive times, it is determined to be a local optimal convergence, and the current turning angle is adjusted accordingly. As a stable operating point.

5. The remote integrated control system for a hydro-generator unit according to claim 1, characterized in that, A relational database is built to store optimal parameters. Through self-learning and iterative adaptation to changes in operating conditions, initial parameters are provided for optimal control. Specifically, this includes the following: The construction of the collaborative relational database is divided into a basic layer, a dynamic layer, and a prediction layer; The newly obtained optimal parameters must meet preset requirements before being stored in the database; and a confidence label must be added to each set of data. Every day at midnight, data cleaning and model training are automatically performed, removing data from the dynamic layer that exceeds the preset time and retraining the prediction layer model. A monthly report on the optimization of the linkage relationship is generated, comparing the efficiency improvement between the dynamic layer and the basic layer, and providing a basis for equipment status assessment for operation and maintenance. When the optimization control starts, it prioritizes calling the optimal parameters of the dynamic layer under the same working condition as the initial point.

6. The remote integrated control system for a hydro-generator unit according to claim 1, characterized in that, A three-tiered collaborative architecture encompassing local, remote, and cloud-based systems is established to enable visualized monitoring of unit operating status, remote control command issuance, and cross-plant data interaction, supporting remote intervention by maintenance personnel and collaboration with grid dispatch.

Images