Intelligent water turbine speed regulation system model based on digital twinning and operation method

Abstract

The invention discloses an intelligent water turbine speed regulation system model based on digital twinning and an operation method, and belongs to the technical field of intelligent water turbines. The digital twin model of the water turbine speed regulating system comprises a geometric model, a physical model, a behavior model and a rule model, and the physical model at least comprises a proportional servo valve model and a main distributing valve model and can accurately describe the dynamic characteristics of the water turbine speed regulating system. And carrying out numerical solution on the physical model through a self-adaptive variable-step implicit iterative solution algorithm to realize real-time monitoring, fault early warning and intelligent diagnosis on the water turbine speed regulating system. According to the method, the mechanism and data effective coupling technology is adopted, the problems that a traditional water turbine speed regulating system is inaccurate in fault diagnosis, high in maintenance cost, difficult in hydraulic system fault early warning and the like are solved, the safety and economical efficiency of the system are improved, and technical support is provided for intelligent operation and maintenance of the water turbine speed regulating system.

Description

Technical Field

[0001] This invention belongs to the field of intelligent water turbine technology, and specifically relates to an intelligent water turbine speed regulation system model and operation method based on digital twin. Background Technology

[0002] With the transformation of the global energy structure and the development of smart grids, hydropower, as a clean and renewable energy source, is playing an increasingly important role in the energy structure. The turbine speed control system, as the core control equipment of a hydropower station, plays a crucial role in maintaining stable operation of the unit near its rated speed, regulating the load of the power plant unit, and stabilizing the frequency. However, ① the turbine speed control system has a complex structure, encompassing numerous components from multiple fields such as mechanics, hydraulics, and electrical systems. For example, it includes high-precision mechanical transmission mechanisms, sophisticated hydraulic servo systems, and complex electrical control devices, requiring significant investment of manpower, material resources, and financial resources for maintenance, resulting in excessively high maintenance costs; ② currently, fault diagnosis of traditional turbine speed control systems mainly relies on manual experience and some simple testing equipment. A single fault phenomenon may be caused by multiple different reasons, and the same fault cause may lead to multiple different fault manifestations. For this complex structure involving multiple disciplines and components, accurate fault diagnosis is difficult.

[0003] In summary, traditional system-level modeling of hydro turbine governor systems primarily focuses on the governor's control strategy and the turbine's torque characteristics, often emphasizing the electrical and hydraulic aspects while significantly simplifying the hydraulic mechanical aspects. However, the hydraulic system is crucial for the operational safety of actual systems. Under sudden faults such as blockages, the parameters of the hydraulic system can change drastically, making traditional explicit algebraic differential equation solving algorithms prone to divergence. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a model and operation method of an intelligent water turbine speed regulation system based on digital twins. By constructing a high-precision digital twin model, real-time monitoring, fault early warning and intelligent diagnosis of the water turbine speed regulation system can be realized, thereby improving the safety and economy of the system.

[0005] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A model and operation method for an intelligent hydro turbine speed regulation system based on digital twins, comprising the following steps: Geometric model: The physical entity structure of the turbine speed regulation system is constructed using 3D modeling software. The physical entity structure of the turbine speed regulation system includes all key components of the pressure oil system, hydraulic servo system, electrical control system, and water diversion and power generation system. Physical Model: Based on the physical principles and motion laws of the key components of the turbine speed control system, mathematical equations are established to describe the intrinsic dynamic characteristics and external interactions of each key component. The physical model includes at least a proportional servo valve model and a main pressure regulating valve model; wherein, The proportional servo valve model takes a current signal as input, and the output flow rate is proportional to the valve core position. Its dynamic response is described by a first-order or higher-order transfer function. The main pressure regulating valve model considers the dead zone formed by the overlap between the main pressure regulating valve core and the main pressure regulating valve bushing, and calculates the irregular opening area of ​​the guide vane main pressure regulating valve switching chamber based on the displacement of the main pressure regulating valve core, and then calculates the flow rate through the valve. Behavioral Model: Based on the physical model, the behavioral response curves of the turbine speed regulation system are simulated under different operating conditions to characterize the dynamic characteristics of the turbine speed regulation system; Rule-based model: Based on historical operating data, standard specifications, and simulation results of the digital twin model of the turbine speed control system, the normal operating range thresholds of each key parameter in the turbine speed control system are set for fault early warning and diagnosis of the turbine speed control system.

[0006] Preferably, the physical model uses bond graphs as a modeling tool to decompose the turbine speed regulation system into the smallest unit modules with clear physical meaning. Each smallest unit module describes its internal motion laws and external interaction relationships, ensuring that the structure of the digital twin model of the turbine speed regulation system is consistent with the actual physical structure of the turbine speed regulation system.

[0007] Preferably, the key parameters in the rule model include: The static flow characteristics and dynamic response performance of the proportional servo valve are measured at the following points: guide vane main distribution displacement, guide vane proportional valve control output, and guide vane proportional valve feedback signal. These are used to warn of slow movement speed of the guide vane main distribution valve, slow response of the guide vane main distribution valve, or overshoot. The pressure characteristics of the main pressure regulating valve, the displacement stroke of the main pressure regulating valve core, and the response time of the main pressure regulating valve are measured at the corresponding points: the opening pressure of the guide vane servo, the closing pressure of the guide vane servo, the pressure of the hydraulic device, and the displacement of the guide vane main pressure regulating valve. These measurements are used to provide early warning of blockage or other abnormalities in the guide vane main pressure regulating valve system.

[0008] Preferably, it also includes a neural network model, which is constructed through the following steps: Collect operating data of the turbine speed regulation system, the operating data including input variables and output variables, and clean and standardize the operating data; Design a multilayer perceptron neural network structure, wherein the number of nodes in the input layer of the multilayer perceptron neural network is consistent with the dimension of the input variables, the hidden layers are 1 to 5 layers with 10 to 20 nodes in each layer, and the number of nodes in the output layer is consistent with the dimension of the output variables; The multilayer perceptron neural network is trained using the mini-batch gradient descent algorithm, and L2 regularization or Dropout method is used to prevent the model from overfitting. The trained neural network model is coupled with the physical model of the turbine speed control system. The parameters of the physical model are optimized using the neural network model, thereby improving the overall simulation accuracy of the digital twin model of the turbine speed control system.

[0009] Preferably, in the proportional servo valve model, based on the response frequency The curve is fitted with a function that achieves the highest accuracy based on its correspondence with the signal. The function expression for this curve is: (1); As known from equation (1), the control currents of the proportional servo valve are 4mA, 12mA, and 20mA, representing the -1, 0, and 1 positions of the proportional valve, respectively. Assuming a first-order system model is used to describe the change in valve core position, its transfer function is expressed as: (2); In the formula: k Let τ be the system gain, and τ be the time constant. τ is calculated as follows: (3); Therefore, in the time domain, the expression corresponding to the valve core position is: (4); The proportional servo valve uses C4 control and has a rated flow rate of 40 l / min. Its frequency response is "L"-shaped. The curve shows no dead zone or obvious inflection point. It can be approximated that the flow rate of the proportional servo valve is directly proportional to the valve core position, i.e., when the pressure difference is constant: (5); In the formula: q —Proportional valve output flow; q v —Rated flow rate of the proportional valve: 40 l / min; Δ p —Proportional valve differential pressure (bar); x —Proportional valve core position.

[0010] Preferably, the key components include a pressure oil tank, a pressure oil pump group, a combination valve, an oil collection tank, a guide vane main pressure distribution valve subsystem, an emergency pressure distribution valve subsystem, a segmented shut-off valve subsystem, a guide vane actuator, a turbine main pressure distribution valve subsystem, a turbine actuator, a controller, a DI module, a DO module, an AI module, an AO module, a frequency measurement module, a feedback module, a power supply module, a water intake pipe, a turbine, and a generator.

[0011] Preferably, in the main pressure regulating valve model, the proportional servo valve receives a current signal converted from the electrical control system to drive the valve core to move, and also receives pressure from the hydraulic oil at both ends to output flow to the control chamber of the main pressure regulating valve core; based on the working principle and characteristic curve of the proportional valve, the input-output variable relationship of the proportional valve is summarized as follows: (6); (7); (8); (9); Preferably, in the main pressure regulating valve model, the input and output variable relationship of the main pressure regulating valve includes calculating the valve core displacement: (10); Based on the geometry and assembly relationship of the main pressure regulating valve, the opening area corresponding to the displacement of the main pressure regulating valve core is obtained: Based on the shape and size of the main pressure regulating valve opening, when the displacement of the main pressure regulating valve core is greater than 0, the main pressure regulating valve opening area is: (11); Because the main pressure regulating valve is designed with strict vertical symmetry, when the displacement is less than 0, the opening area is consistent with the calculation method of formula (11); Calculate the control chamber pressure: (12); Calculate flow rate: (13); An operation method for an intelligent hydro turbine speed regulation system model based on digital twins includes the following steps: S1. Construct a digital twin model of the turbine speed regulation system: Based on the design drawings and technical parameters of the turbine speed regulation system, establish the geometric model, physical model, behavioral model and rule model of the turbine speed regulation system respectively; S2. Model Solving and Simulation: An adaptive variable step size implicit iterative solving algorithm is used to numerically solve the physical model of the turbine speed control system. The solution step size is dynamically adjusted by monitoring the convergence of the system state, and the stability of the solution process is ensured by combining predictive iteration technology. It is especially suitable for the transient process simulation of pressure oil system or hydraulic servo system when switching operation or sudden failure occurs. S3. Real-time data access: Input the real-time operating data of the turbine speed control system collected by the field sensors into the digital twin model of the turbine speed control system; S4. Model Calibration and Optimization: Compare the real-time operating data of the turbine speed regulation system with the simulation output data of the digital twin model of the turbine speed regulation system, and use the difference information between the two to optimize the parameters of the physical model of the turbine speed regulation system online or offline, so as to improve the accuracy of the digital twin model of the turbine speed regulation system. S5. Condition monitoring and fault diagnosis: The real-time operating data of the turbine speed control system is compared with the normal operating range threshold set in the turbine speed control system rule model. When the real-time operating data exceeds the set threshold, a fault warning is triggered, and the fault type and cause are analyzed in combination with the turbine speed control system behavior model.

[0012] Preferably, step S5 specifically includes: Monitor the rate of change of the guide vane main displacement. If the rate of change is lower than the threshold set in the turbine speed regulation system rule model, it is determined that the proportional servo valve may have flow attenuation or blockage. Monitor the response time of the guide vane main pressure distribution valve. If the response time exceeds the response time threshold set in the turbine speed regulation system rule model, it is determined that there is a risk of blockage in the guide vane main pressure distribution valve or proportional servo valve. Monitor the changing trends of the guide vane servo opening pressure and the guide vane servo closing pressure. If the difference between them and the inlet oil pressure of the hydraulic device is abnormal or does not conform to the expected dynamic characteristics of the turbine speed regulation system behavior model, then it is determined that the working state of the guide vane main pressure distribution valve is abnormal.

[0013] Preferably, the method further includes step S6: intelligent decision support, which generates operation optimization suggestions or equipment maintenance strategies for the hydro-generator unit based on the simulation results and fault diagnosis conclusions of the digital twin model of the turbine speed regulation system, and pushes the suggestions or strategies to the operator or the turbine speed regulation system control system.

[0014] A computer device includes one or more processors, said processors storing one or more executable programs, which, when executed by said one or more processors, are used to implement the operation method of a digital twin-based intelligent water turbine speed regulation system model.

[0015] A storage medium storing one or more executable programs, wherein when the one or more executable programs are executed, they are used to implement the operation method of the aforementioned intelligent water turbine speed regulation system model based on digital twin.

[0016] The present invention can achieve the following beneficial effects: 1. By employing the effective coupling technology of mechanism and data, a digital twin model of the speed control system is constructed, which includes the geometric model, physical model, behavioral model and rule model of the speed control system. This enables the simulation of the behavior of the normal and faulty systems under various operating conditions, and can support advanced applications such as intelligent early warning, fault diagnosis and performance evaluation of the speed control system.

[0017] 2. This invention first developed a series of underlying supporting technologies for digital twin research, such as mathematical-physical modeling and data-driven model optimization. Then, it constructed a digital twin model of the speed control system for a hydro turbine, including geometric, physical, behavioral, and rule models. Aiming at real-time data analysis based on the twin, the twin model is unified into function calls to support advanced applications such as early warning and diagnosis. This breaks down the barriers between simulation and on-site monitoring and analysis. Utilizing the twin model to analyze real-time data provides enhanced perception data, enabling intelligent operation guidance and optimization suggestions. Attached Figure Description

[0018] The present invention will be further described below with reference to the accompanying drawings and embodiments: Figure 1 Digital Twin Key Technology Architecture Diagram; Figure 2 Overall framework diagram of the digital twin model of the water turbine speed regulation system; Figure 3 : Relationship between the signal variation span and response frequency of a proportional servo valve and the fitting curve of the exponential function; Figure 4 Schematic diagram of the proportional servo valve control system for the guide vane unit of the hydraulic follow-up system; Figure 5 : Response curves of the spool position of a proportional servo valve under different step currents; Figure 6 : Geometric model diagram of the main pressure regulating valve core; Figure 7 Schematic diagram of the closed-loop control system of main pressure regulating valve and proportional valve in series; Figure 8 Comparison curves of simulation and experimental data of guide vane main displacement under load opening adjustment condition 1; Figure 9 Simulation curves of the dynamic response of the main pressure regulating valve with multiple parameters (given displacement, current, flow rate, valve core displacement, displacement change rate, acceleration, and control pressure). Detailed Implementation

[0019] In this embodiment, physical entity analysis is conducted on the speed control system. The system to be included in the modeling is extracted, and the equipment of each part is divided to form a modeling list. The design drawings, technical parameters, and operating data of the turbine speed control system are collected, and a digital twin model is constructed and refined. This includes key components such as the governor, pressure oil system, hydraulic servo system, electrical control system, and water diversion and power generation system.

[0020] The simulation modeling of the speed governor system utilizes the construction of a digital twin model through minimum units. This is achieved using a distributed physical graph modeling approach, which decomposes the complex physical system into relatively independent yet interconnected modules (i.e., minimum units). This allows for in-depth analysis of the physical system's structure and motion mechanisms, thereby supporting the construction of the digital twin model. Specifically, the complex physical system (such as mechanical, hydraulic, and electrical systems) is decomposed into unit modules with clearly defined physical meanings. Each module describes its inherent motion laws (such as dynamics and thermodynamics) and external interaction relationships (such as signal flow and energy flow). Bond graphs are used as the modeling tool, employing unified symbols to represent energy transfer and conversion, avoiding the abstractness of traditional mathematical modeling and ensuring that the model structure remains consistent with the actual system structure.

[0021] The equipment list for modeling the speed governor system is shown in Table 1: Table 1

[0022] Key technologies for digital twins are attached. Figure 1 .

[0023] The present invention provides a model of an intelligent water turbine speed regulation system based on digital twins, which is specifically constructed as follows: Taking the proportional valve and main pressure regulating valve of the guide vane unit in a hydraulic servo system as examples, the steps for constructing a twin model are as follows: Proportional valve: (1) Proportional valve principle A proportional servo valve (or proportional valve for short) can convert the input electrical control signal proportionally and output a flow control signal. When the proportional servo valve is in automatic operation, it is the main component that outputs the pressure oil signal, controls the piston movement of the main pressure regulating valve to output pressure energy, and then operates the relay.

[0024] After the proportional servo valve is powered on normally, the displacement sensor of the electromagnet feeds back the actual displacement of the valve core to the controller. The controller compares the actual displacement value with the command value. If the actual value deviates from the command value, the electromagnetic thrust of the electromagnet will change until the new electromagnetic thrust is equal to the thrust of the feedback spring, thereby ensuring that the actual displacement value is equal to the command value.

[0025] The proportional servo valve is a direct-acting spool valve with a valve core and sleeve structure, and is driven by a high-frequency response proportional electromagnet with position feedback.

[0026] Frequency response curves are important tools for analyzing the dynamic characteristics of a system, typically used to describe the input-output relationship of a linear time-invariant (LTI) system at different frequencies. The frequency response curve of a proportional servo valve generally includes two parts: amplitude-frequency response and phase-frequency response. In this study, the amplitude-frequency curve is selected as the object of study, represented by the four curves shown above, representing the frequency response under four signal values. The vertical axis of the amplitude-frequency response curve represents the system gain, in decibels (dB). It is calculated as follows: where Iout is the output signal and Iin is the input signal. In this proportional servo valve, when the absolute value of the gain is less than 3, it can be considered that there is no attenuation or amplification of the input and output signals. The response frequency f of this proportional servo valve... c The correspondence between the signals is shown in Table 2: Table 2

[0027] Based on the data from the four points mentioned above, a function with the highest accuracy is fitted, as shown in the attached figure. Figure 3 As shown, the function expression for this curve is: (1) Furthermore, as known above, the control current of the proportional servo valve is 4-(12)-20mA, representing the -1, 0, and 1 positions of the proportional valve, respectively. Assuming that the change in valve core position can be described by a first-order system model, its transfer function can be expressed as: (2) In the formula: k is the system gain, τ is the time constant, and τ is calculated as follows: (3) Therefore, in the time domain, the expression corresponding to the valve core position is: (4) The proportional servo valve uses C4 control and has a rated flow rate of 40 l / min. Its frequency response is "L"-shaped. The curve shows no dead zone or obvious inflection point. It can be approximated that the flow rate of the proportional servo valve is directly proportional to the valve core position, i.e., when the pressure difference is constant: (5) In the formula: q—Proportional valve output flow rate; qv — Rated flow rate of proportional valve: 40 l / min; Δp — proportional valve pressure difference (bar); x — Proportional valve core position.

[0028] (2) Geometric model of proportional valve The geometric model of the proportional valve was created using SolidWorks. The proportional valve is a four-way direct-acting proportional servo directional control valve, employing a valve core and sleeve structure with servo valve performance. It has a nominal diameter of 6mm. The control and drive components of the proportional valve mainly include a high-frequency response proportional electromagnet with electrical displacement feedback, an integrated digital controller, and a seven-pin connector. (3) Proportional valve physical model Based on the input-output relationship of the proportional valve, a physical model is established: The proportional servo valve receives a current signal converted from the electrical control system, driving the valve core to move. It also receives pressure from hydraulic oil at both ends, outputting flow to the control chamber of the main valve core. Based on the proportional valve's working principle and characteristic curves, the input-output variable relationship of the proportional valve is summarized as follows: (6) (7) (8) (9) (4) Proportional valve behavior model When the current changes abruptly from 12mA to different target values ​​(14mA, 16mA, and 20mA), the valve spool position response curve of the proportional valve is as follows: Figure 5 As shown. During this process, the main configuration output pressure received by the system remains at 10 bar, while the input pressure provided by the pressure tank remains at 40 bar. Faced with such a significant change in current within a short period of time, the proportional valve spool response will exhibit a certain degree of hysteresis.

[0029] Analysis using a first-order linear model shows that the response time of the valve core to changes in current amplitude exhibits a roughly linear relationship with frequency. This means that as the rate of current change increases, the valve core's position adjustment will show a more significant delay. To improve simulation accuracy, future work could consider using higher-order mathematical models or piecewise function models to more accurately describe and predict the valve core's behavior, thereby better reflecting the dynamic characteristics of the actual system.

[0030] (5) Proportional valve rule model The rule model of a proportional valve is not limited to static flow characteristics but also includes dynamic response performance. In a hydraulic system, the proportional valve, as a control element, regulates the pressure, flow rate, and direction of the fluid according to the proportion of the input signal. The most important indicators for evaluating the correctness of a proportional valve model are its static flow characteristics and dynamic response performance. As a pilot control element, the proportional valve indirectly controls the position of the main pressure regulating valve in a hydraulic system by adjusting the pilot pressure, thereby adjusting the flow rate and pressure of the entire system. Therefore, the rule model of a proportional valve must fully consider its impact on the main pressure regulating valve. The correspondence between the measurement points and fault warning types for the proportional valve rule model is shown in Table 3.

[0031] Table 3

[0032] Static traffic characteristics: The static flow characteristic of a proportional valve refers to the relationship between the flow rate through the valve and the input signal (current) under steady-state conditions. The error between the static characteristics of a normal proportional valve and the characteristic curve in the proportional valve's manufacturer's manual should be within a certain range.

[0033] This characteristic is most directly reflected in the relationship between the flow rate through the proportional valve and the position of its valve core, indirectly affecting the operating speed and dynamic performance of the main valve core. By analyzing the displacement data of the main valve core obtained from actual measurement points, the operating speed of the main valve core can be calculated. Comparing these measured data with the expected results of the proportional valve rule model can effectively assess whether the current proportional valve's performance is satisfactory. Dynamic response performance: The dynamic response performance of a proportional servo valve refers to the time and process required for it to adjust the opening to the set value based on feedback adjustment signals of the target value and the current value. This rule is also mentioned in the DL / T 563-2016 standard, which sets the threshold for the proportional valve response time range based on historical data, twin data, and standard data.

[0034] The dynamic response performance of a proportional valve directly affects the response speed and overshoot of the main pressure regulating valve spool, i.e., the time it takes for the spool to move and the degree to which it exceeds the target position. Good dynamic response ensures that the setpoint is reached quickly and smoothly, reducing overshoot and oscillation, thereby improving the stability and accuracy of the system.

[0035] Main pressure regulating valve: (1) Principle of main pressure regulating valve The main pressure regulating valve consists of a valve body, a main piston, a main bushing, etc. It is a hydraulically controlled main pressure regulating valve with an auxiliary relay, and its interface is a proportional servo or pilot valve.

[0036] The auxiliary servo of the main pressure regulating valve is a differential pressure type, with the control chamber (large chamber) area approximately twice the area of ​​the constant pressure chamber (small chamber). The working oil of the main pressure regulating valve is supplied to the constant pressure chamber of the auxiliary servo, while the control oil of the proportional servo valve (or pilot valve) is sent to the control chamber of the auxiliary servo of the main pressure regulating valve via the emergency stop solenoid valve. When the proportional servo valve (or pilot valve) is in the intermediate equilibrium position, the piston of the main pressure regulating valve is stationary. A signal from the microcomputer controller causes the proportional servo valve (or servo motor or manual handle) to move, increasing the pressure in the control chamber of the auxiliary servo of the main pressure regulating valve, causing the piston of the main pressure regulating valve to move downwards, and opening the servo. Conversely, a signal from the microcomputer controller causes the proportional servo valve (or servo motor or manual handle) to move in the closing direction, decreasing the pressure in the control chamber of the auxiliary servo of the main pressure regulating valve, causing the piston of the main pressure regulating valve to move upwards, and closing the servo.

[0037] (2) The parameters of the main pressure regulating valve are shown in Table 4. Table 4

[0038] (3) Geometric model of main pressure regulating valve The geometric model of the main pressure regulating valve was created in SolidWorks by precisely defining its key components and functional characteristics. In addition to establishing the external model of the main pressure regulating valve, its internal structure was also meticulously modeled, including major components such as the main valve piston (valve core), main valve bushing, small piston screw (connecting bolt), small piston, adjusting support plate, valve cover, and large support plate. This ensures that the model not only geometrically accurately represents the external design of the main pressure regulating valve but also reflects its valve core position and complex flow regulation function. Based on the actual product, the geometric dimensions and positions of each component were defined, ensuring seamless connection of the hydraulic components. Detailed mating relationships, such as concentricity, parallelism, and coincidence, were also set in the geometric model to ensure correct docking and motion characteristics between components. This provides a solid foundation for subsequent assembly modeling and dynamic simulation analysis, ensuring the reliable operation of the main pressure regulating valve in practical applications.

[0039] The geometric model of the main pressure regulating valve core is attached. Figure 6 .

[0040] (4) Physical model of the main pressure regulating valve A physical model is built based on the input-output relationship of the main pressure regulating valve.

[0041] Based on the above working principle of the main pressure regulating valve, the input and output variable relationships of the main pressure regulating valve are summarized, including the calculation of valve core displacement: (10) Based on its geometric dimensions and assembly relationship, the opening area corresponding to the displacement of the main pressure regulating valve core is obtained: The switching ports of the main pressure regulating valve are irregularly shaped, with rounded top and bottom edges and a chamfered rectangle in the middle. Furthermore, the piston distribution and switching port positions of the main regulating valve core strictly adhere to a top-to-bottom symmetrical design. The main regulating piston height is 30.8mm, while the height of the main regulating switching ports (A and B ports) is 30mm, resulting in a 0.4mm overlap (dead zone) for the main regulating valve core. Although the maximum opening of the main regulating switching port is 30mm, the main regulating valve core is also constrained by the internal assembly, limiting its displacement to between -15.5mm and 15.5mm.

[0042] Based on the shape and size of the main valve opening, when the displacement of the main valve core is greater than 0, the area of ​​the main valve opening is: (11) Because the main pressure regulating valve is designed with strict vertical symmetry, when the displacement is less than 0, the opening area is consistent with the above calculation method. The displacement change curve of the main pressure regulating valve is calculated according to the above method, where positive and negative represent the opening direction.

[0043] Calculate the control chamber pressure: (12) Calculate flow rate: (13) (4) Behavioral model of main pressure regulating valve In a closed-loop control system consisting of a proportional valve and a main pressure regulating valve connected in series, the system input is a given main pressure regulating valve spool position signal, mdv_x_given. This given signal is compared with the actual measured main pressure regulating valve spool position, x_mdv, and converted into a current signal I_pv acting on the proportional valve through feedback gain, thereby adjusting the flow rate in the main pressure regulating valve control chamber and the position of the main pressure regulating valve spool. When the system receives a step change position command from -0.5 to 0.5, this command is converted into a corresponding current signal after feedback and gain adjustment, thus affecting the operating state of the proportional valve. As the proportional valve responds to the change in current, it changes the flow rate Q_pv through the main pressure regulating valve control chamber, causing the main pressure regulating valve spool to displace, and simultaneously causing a change in the control chamber pressure p_kong. When the given signal is -0.5, the flow rate Q_pv generated by the system is negative, meaning that the hydraulic oil flow direction is towards the direction of closing the control chamber. As the main pressure regulating valve spool approaches its maximum closing stroke limit, the flow rate gradually decreases until it reaches zero, at which point the pressure in the control chamber also drops to zero. Conversely, when the signal step reaches 0.5, the flow rate Q_pv turns positive, indicating that the hydraulic oil begins to flow towards the opening control chamber. The main valve spool then moves to its maximum opening stroke, at which point the flow rate drops to zero again, while the pressure in the control chamber rises to equal the inlet pressure.

[0044] (4) Main pressure regulating valve regular model The rule model of the main pressure regulating valve revolves around pressure characteristics, displacement stroke, and response time, which are crucial for ensuring system stability and efficiency. The following is a detailed description of each aspect. Table 5 shows the correspondence between the measurement points and fault warning types for the main pressure regulating valve rule model: Table 5

[0045] Pressure characteristics: The pressure characteristics primarily reflect the pressure difference (ΔP) between the inlet and outlet of the main pressure regulating valve, which determines the pressure drop as fluid passes through the valve. Precise control of ΔP is a key factor in ensuring the normal operation of the system. A range threshold for the pressure in the guide vane relay switching chamber is constructed based on historical and twin data. Furthermore, to adapt to different operating conditions, the rule-based model must also possess the ability to self-correct based on real-time data, ensuring that the temperature range is reasonably defined under various operating conditions. By comparing the measured data with the range in the rule-based model, it is possible to effectively assess whether the current operating status of the main pressure regulating valve is abnormal.

[0046] Valve core stroke: Displacement stroke refers to the distance the main distribution valve core moves from position 0 to its maximum opening. This parameter directly affects the valve opening and flow rate, and consequently, the maximum movement speed of the relay. When more precise control of the guide vane opening is required, a smaller displacement stroke is often chosen to ensure that the guide vane opens at a speed that is not too fast. The main distribution valve core stroke is set based on historical data for various operating conditions or environments, serving as a rule model. When actual operating conditions change, the system is reminded to promptly switch to the corresponding main distribution valve core stroke.

[0047] Response time: The response time of the main pressure regulating valve specifically refers to the time required from when the proportional valve receives a command and begins to act until the main pressure regulating valve is fully open. Shortening this time interval can improve the overall system's reaction speed and dynamic response capability. Similarly, a range of main pressure regulating valve core response time is set as a rule model based on historical data and twin data. The guide vane response time is calculated based on the actual measured guide vane displacement. If the response time exceeds the threshold range of the rule model, the system is alerted that the main pressure regulating valve core response speed has decreased, potentially indicating a blockage fault in the main pressure regulating valve or proportional valve.

[0048] After simulating and modeling the water turbine, the model is constructed based on data-driven and neural network algorithms. The specific steps are as follows: (1) Data preparation and preprocessing Collect turbine operating data, including input variables (such as guide vane opening, head, and rotational speed) and output variables (such as power output, flow rate, and efficiency). Data sources include on-site monitoring systems or historical operating records, ensuring data coverage across different operating conditions. The data is then cleaned and standardized to remove abnormal data caused by sensor malfunctions or operational errors. Input and output data are scaled to the [0,1] or [-1,1] range to avoid network convergence difficulties due to dimensional differences.

[0049] (2) Neural network model design Network architecture selection: Input layer: The number of nodes is consistent with the dimension of the input variables (such as guide vane opening, head, and rotational speed).

[0050] Hidden layers: Multilayer perceptron (MLP) is used, and the number of hidden layers and nodes needs to be determined experimentally. For example, 3 hidden layers (10-20 nodes per layer) can better balance complexity and training efficiency.

[0051] Output layer: The number of nodes is consistent with the dimension of the output variables (such as output power, flow rate).

[0052] Activation functions: ReLU is used in the hidden layers (to avoid gradient vanishing), and a linear function is used in the output layer (for regression tasks).

[0053] (3) Model training Employ mini-batch gradient descent (e.g., batch size=32) to balance memory usage and convergence speed. Stop training when the validation set error fails to decrease for several consecutive rounds to avoid overfitting. Add L2 regularization or Dropout layers to suppress overfitting.

[0054] Calculate the MSE, MAE, and R² scores on the test set to evaluate the model's prediction accuracy, plot the comparison curve between predicted and actual values, and analyze the error distribution.

[0055] (4) Model optimization Use grid search or Bayesian optimization to adjust parameters such as the number of hidden layers, number of nodes, and learning rate to further improve performance.

[0056] Real-time prediction and deployment: The trained model is deployed to the turbine control system, receiving sensor data in real time and outputting prediction results. Combined with a PID controller, optimized scheduling of the turbine is achieved.

[0057] This invention employs a data-physical integrated digital twin modeling technique to achieve a holistic twin model of the system. The overall approach is to determine the input and output of each device model by theoretically analyzing the physical entity and comprehensively considering the on-site sensor data. Based on the physical relationships between parameters and combined with the characteristics of the devices, the physical model of the system is determined. Then, the parameters of the physical model are optimized by utilizing the behavioral differences between the actual data and the physical model, thereby completing the model establishment.

[0058] This invention constructs high-precision digital twins of axial-flow propeller turbine units and mixed-flow turbine units. The models cover all key components of the governor system, including the electrical control system, hydraulic servo system, pressure oil system, and hydroelectric power generation system. The constructed digital twins comprehensively cover the electrical, hydraulic mechanical, and hydraulic aspects. In particular, for the hydraulic system, every component in the governor's hydraulic system is accurately reproduced. Through the series connection of models, the hydraulic subsystem is constructed in detail, making up for the shortcomings of traditional models in the details of the hydraulic mechanical side. It also provides a solid foundation for a deeper understanding of the dynamic behavior of the governor under normal and fault conditions and its impact on the performance of the entire system.

[0059] Firstly, in terms of model design, this invention employs appropriate equation transformations to design a more targeted physical model structure based on the physical process, reducing instability caused by equation distortion. Furthermore, the model can be directly compared with actual measurement points, supporting twin applications. Secondly, in terms of the solution algorithm, an adaptive variable-step-size implicit iterative solution algorithm is introduced. This algorithm monitors the convergence of the state, promptly reducing the solution step size, and combines predictive iteration techniques to ensure solution accuracy. The solver exhibits better numerical stability under instantaneous dynamic changes in hydraulic systems, such as switching operations or fault occurrences, providing a foundation for subsequent fault behavior simulation research.

[0060] The above embodiments are merely preferred technical solutions of the present invention and should not be considered as limitations on the present invention. The scope of protection of the present invention should be limited to the technical solutions described in the claims, including equivalent substitutions of the technical features described in the claims. That is, equivalent substitutions and improvements within this scope are also within the scope of protection of the present invention.

Claims

1. A model of an intelligent water turbine speed regulation system based on digital twin, characterized in that, include: Geometric model: The physical entity structure of the turbine speed regulation system is constructed using 3D modeling software. The physical entity structure of the turbine speed regulation system includes all key components of the pressure oil system, hydraulic servo system, electrical control system, and water diversion and power generation system. Physical Model: Based on the physical principles and motion laws of the key components of the turbine speed control system, mathematical equations are established to describe the intrinsic dynamic characteristics and external interactions of each key component. The physical model includes at least a proportional servo valve model and a main pressure regulating valve model; wherein, The proportional servo valve model takes a current signal as input, and the output flow rate is proportional to the valve core position. Its dynamic response is described by a first-order or higher-order transfer function. The main pressure regulating valve model considers the dead zone formed by the overlap between the main pressure regulating valve core and the main pressure regulating valve bushing, and calculates the irregular opening area of ​​the guide vane main pressure regulating valve switching chamber based on the displacement of the main pressure regulating valve core, and then calculates the flow rate through the valve. Behavioral Model: Based on the physical model, the behavioral response curves of the turbine speed regulation system are simulated under different operating conditions to characterize the dynamic characteristics of the turbine speed regulation system; Rule-based model: Based on historical operating data, standard specifications, and simulation results of the digital twin model of the turbine speed control system, the normal operating range thresholds of each key parameter in the turbine speed control system are set for fault early warning and diagnosis of the turbine speed control system.

2. The intelligent water turbine speed regulation system model based on digital twin as described in claim 1, characterized in that, The physical model uses bond graphs as a modeling tool to decompose the turbine speed control system into the smallest unit modules with clear physical meaning. Each smallest unit module describes its internal motion laws and external interaction relationships, ensuring that the structure of the digital twin model of the turbine speed control system is consistent with the actual physical structure of the turbine speed control system.

3. The intelligent turbine speed control system model based on digital twin as described in claim 1, characterized in that, The key parameters in the rule model include: The static flow characteristics and dynamic response performance of the proportional servo valve are measured at the following points: guide vane main distribution displacement, guide vane proportional valve control output, and guide vane proportional valve feedback signal. These are used to warn of slow movement speed of the guide vane main distribution valve, slow response of the guide vane main distribution valve, or overshoot. The pressure characteristics of the main pressure regulating valve, the displacement stroke of the main pressure regulating valve core, and the response time of the main pressure regulating valve are measured at the corresponding points: the opening pressure of the guide vane servo, the closing pressure of the guide vane servo, the pressure of the hydraulic device, and the displacement of the guide vane main pressure regulating valve. These measurements are used to provide early warning of blockage or other abnormalities in the guide vane main pressure regulating valve system.

4. The intelligent water turbine speed regulation system model based on digital twin as described in claim 1, characterized in that, It also includes a neural network model, which is constructed through the following steps: Collect operating data of the turbine speed regulation system, the operating data including input variables and output variables, and clean and standardize the operating data; Design a multilayer perceptron neural network structure, wherein the number of nodes in the input layer of the multilayer perceptron neural network is consistent with the dimension of the input variables, the hidden layers are 1 to 5 layers with 10 to 20 nodes in each layer, and the number of nodes in the output layer is consistent with the dimension of the output variables; The multilayer perceptron neural network is trained using the mini-batch gradient descent algorithm, and L2 regularization or Dropout method is used to prevent the model from overfitting. The trained neural network model is coupled with the physical model of the turbine speed control system. The parameters of the physical model are optimized using the neural network model, thereby improving the overall simulation accuracy of the digital twin model of the turbine speed control system.

5. The intelligent turbine speed control system model based on digital twin as described in claim 1, characterized in that, In the proportional servo valve model, based on the response frequency The curve is fitted with a function that achieves the highest accuracy based on its correspondence with the signal. The function expression for this curve is: （1）； As is known from equation (1), To control the current, the proportional servo valve has control currents of 4mA, 12mA, and 20mA, representing the -1, 0, and 1 positions of the proportional valve, respectively. Assuming a first-order system model is used to describe the change in valve spool position, its transfer function is expressed as: （2）； In the formula: k Let τ be the system gain, and τ be the time constant. τ is calculated as follows: （3）； Therefore, in the time domain, the expression corresponding to the valve core position is: （4）； The proportional servo valve uses C4 control and has a rated flow rate of 40 l / min. Its frequency response is "L"-shaped. The curve shows no dead zone or obvious inflection point. It can be approximated that the flow rate of the proportional servo valve is directly proportional to the valve core position, i.e., when the pressure difference is constant: （5）； In the formula: q —Proportional valve output flow; q v —Rated flow rate of the proportional valve: 40 l / min; Δ p —Proportional valve pressure differential; x —Proportional valve core position.

6. The intelligent water turbine speed regulation system model based on digital twin according to claim 1, characterized in that, The key components include a pressure oil tank, a pressure oil pump unit, a combination valve, an oil collection tank, a guide vane main pressure distribution valve subsystem, an emergency pressure distribution valve subsystem, a segmented shut-off valve subsystem, a guide vane actuator, a turbine main pressure distribution valve subsystem, a turbine actuator, a controller, a DI module, a DO module, an AI module, an AO module, a frequency measurement module, a feedback module, a power supply module, a water intake pipeline, a water turbine, and a generator.

7. The intelligent turbine speed regulation system model based on digital twin according to claim 5, characterized in that, In the main pressure regulating valve model, the proportional servo valve receives a current signal converted from the electrical control system, driving the valve core to move. It also receives pressure from hydraulic oil at both ends, outputting flow to the control chamber of the main pressure regulating valve core. Based on the working principle and characteristic curve of the proportional valve, the input-output variable relationship of the proportional valve is summarized as follows: （6）； （7）； （8）； （9）。 8. The intelligent water turbine speed regulation system model based on digital twin according to claim 1, characterized in that, In the main pressure regulating valve model, the input and output variables of the main pressure regulating valve are related, including the calculation of the valve core displacement: （10）； Based on the geometry and assembly relationship of the main pressure regulating valve, the opening area corresponding to the displacement of the main pressure regulating valve core is obtained: Based on the shape and size of the main pressure regulating valve opening, when the displacement of the main pressure regulating valve core is greater than 0, the main pressure regulating valve opening area is: （11）； Because the main pressure regulating valve is designed with strict vertical symmetry, when the displacement is less than 0, the opening area is consistent with the calculation method of formula (11); Calculate the control chamber pressure: （12）； Calculate flow rate: （13）。 9. A method for operating a digital twin-based intelligent turbine speed regulation system model as described in any one of claims 1 to 8, characterized in that, Includes the following steps: S1. Construct a digital twin model of the turbine speed regulation system: Based on the design drawings and technical parameters of the turbine speed regulation system, establish the geometric model, physical model, behavioral model and rule model of the turbine speed regulation system respectively; S2. Model Solving and Simulation: An adaptive variable step size implicit iterative solving algorithm is used to numerically solve the physical model of the turbine speed control system. The solution step size is dynamically adjusted by monitoring the convergence of the system state, and the stability of the solution process is ensured by combining predictive iteration technology. It is especially suitable for the transient process simulation of pressure oil system or hydraulic servo system when switching operation or sudden failure occurs. S3. Real-time data access: Input the real-time operating data of the turbine speed control system collected by the field sensors into the digital twin model of the turbine speed control system; S4. Model Calibration and Optimization: Compare the real-time operating data of the turbine speed regulation system with the simulation output data of the digital twin model of the turbine speed regulation system, and use the difference information between the two to optimize the parameters of the physical model of the turbine speed regulation system online or offline, so as to improve the accuracy of the digital twin model of the turbine speed regulation system. S5. Condition monitoring and fault diagnosis: The real-time operating data of the turbine speed control system is compared with the normal operating range threshold set in the turbine speed control system rule model. When the real-time operating data exceeds the set threshold, a fault warning is triggered, and the fault type and cause are analyzed in combination with the turbine speed control system behavior model.

10. The operation method of a digital twin-based intelligent turbine speed regulation system model according to claim 9, characterized in that, Step S5 specifically includes: Monitor the rate of change of the guide vane main displacement. If the rate of change is lower than the threshold set in the turbine speed regulation system rule model, it is determined that the proportional servo valve may have flow attenuation or blockage. Monitor the response time of the guide vane main pressure distribution valve. If the response time exceeds the response time threshold set in the turbine speed regulation system rule model, it is determined that there is a risk of blockage in the guide vane main pressure distribution valve or proportional servo valve. Monitor the changing trends of the guide vane servo opening pressure and the guide vane servo closing pressure. If the difference between them and the inlet oil pressure of the hydraulic device is abnormal or does not conform to the expected dynamic characteristics of the turbine speed regulation system behavior model, then it is determined that the working state of the guide vane main pressure distribution valve is abnormal.

11. The operation method of a digital twin-based intelligent turbine speed regulation system model according to claim 10, characterized in that, It also includes step S6: intelligent decision support, which generates operation optimization suggestions or equipment maintenance strategies for the hydro-generator unit based on the simulation results and fault diagnosis conclusions of the digital twin model of the turbine speed regulation system, and pushes the suggestions or strategies to the operators or the turbine speed regulation system control system.

12. A computer device, characterized in that, It includes one or more processors, each processor storing one or more executable programs, which, when executed by the one or more processors, are used to implement the operation method of a digital twin-based intelligent water turbine speed regulation system model according to claim 9 or 10.

13. A storage medium, characterized in that, It stores one or more executable programs, which, when executed, are used to implement the operation method of a digital twin-based intelligent water turbine speed regulation system model as described in claim 9 or 10.

Images