Flow control method for electric actuator based on multi-working-condition priori weighted fusion and related device

Abstract

The application relates to an electric actuator flow control method based on multi-working-condition priori weighting fusion and related devices, which comprises the following steps: constructing a priori database containing typical working-condition characteristic vectors and a fluid mechanics parameter set; collecting valve opening and motor output torque to generate an observation vector in real time; calculating the characteristic distance of the observation vector and the priori working condition and dividing the confidence weight; weighting and fusing the fluid mechanics parameter set based on the weight to generate actual working-condition synthetic parameters; and using the synthetic parameters to estimate the flow in real time or adjust the valve opening. The application has the advantages that the torque characteristics can be used to automatically identify and adapt to complex working condition changes, effectively compensate pressure fluctuations and mechanical wear, and realize high-precision flow soft measurement and self-adaptive control without external sensors.

Description

Technical Field

[0001] This application relates to the field of fluid control in industrial automation, and in particular to a flow control method and related apparatus for electric actuators based on multi-condition prior weighted fusion. Background Technology

[0002] In modern industrial fluid transport and process control systems, electrically operated control valves are the core actuators for regulating media flow. In standard control logic, the control system typically issues commands based on a preset mapping relationship between valve opening and flow rate. This control mode presupposes idealized boundary conditions, assuming that the pressure difference across the valve is constant and the physical properties of the medium remain stable over long-term operation. Under these conditions, each physical stroke position of the valve corresponds to a specific flow rate value, and the system only needs to control the valve to reach the designated position to complete flow regulation.

[0003] However, in practical engineering applications, due to limitations in construction costs and system complexity, complete flow meters or pressure transmitters are usually only installed on main pipelines or critical process nodes. For the numerous terminal branches or auxiliary pipelines, only electric valves are often configured without corresponding flow detection instruments. This uneven distribution of sensors results in a large number of sensing blind spots in the pipeline network. In these areas lacking feedback, the control system cannot obtain actual flow data or pipeline pressure data and is forced to rely on the static flow characteristic curve set at the factory for open-loop control.

[0004] A fluid pipeline network is essentially a dynamically coupled pressure system. Changes in the operating status of upstream pumping stations, the opening and closing of valves in adjacent pipelines, and variations in the load at the end of the pipeline all alter the pressure distribution within the network in real time. According to fluid mechanics principles, the flow rate through a valve depends not only on the valve's flow area but also on the pressure difference across the valve. When fluctuations in the pipeline pressure cause the pressure difference across the valve to deviate from its rated value, the original mapping relationship between valve opening and flow rate no longer holds. For example, when pipeline pressure increases significantly, even if the valve maintains the same opening, the actual flow rate will increase accordingly.

[0005] In scenarios lacking external sensor feedback, existing electric actuators merely function as position servo mechanisms. While they can execute position commands with high precision, they lack the ability to sense the fluid state inside the pipeline. When actual operating conditions deviate due to pressure fluctuations or changes in medium viscosity, the actuators fail to recognize these changes and rigidly execute position commands based on baseline conditions. This leads to a disconnect between control commands and actual process requirements, making it difficult for the system to maintain the expected flow control accuracy under non-steady-state conditions. Furthermore, it fails to provide effective flow metering data to the upper-level management system, limiting the improvement of overall pipeline network operating efficiency and the realization of digital management. Summary of the Invention

[0006] In order to achieve closed-loop control and real-time monitoring of fluid flow in a wide branch pipeline network lacking external sensor assistance, this application provides an electric actuator flow control method and related device based on multi-condition prior weighted fusion.

[0007] Firstly, this application provides a flow control method for electric actuators based on multi-condition prior weighted fusion, which adopts the following technical solution: A flow control method for electric actuators based on multi-condition prior weighted fusion includes the following steps: S1. Pre-acquire the operating data of the electric actuator and valve under several typical working conditions, and construct a priori database, wherein the priori database contains the feature vector and fluid dynamics parameter set corresponding to each typical working condition; S2. During the operation of the electric actuator, the valve opening data and motor output torque data of the valve are collected in real time, and a real-time observation vector reflecting the characteristics of the current working condition is generated based on a preset time window; S3. Calculate the feature distance between the real-time observation vector and the feature vector of each typical working condition in the prior database, and assign a confidence weight representing the degree of matching to each typical working condition according to the magnitude of the feature distance; S4. Based on the confidence weights, perform weighted summation on the fluid dynamics parameter sets of all the typical working conditions to generate the actual working condition composite parameters at the current moment; S5. Apply the synthesized parameters of the actual operating conditions to the flow control model to achieve real-time estimation of fluid flow or adjustment of valve opening.

[0008] By adopting the above technical solution, operational data under typical working conditions are acquired in advance, and a priori database containing feature vectors and fluid dynamics parameter sets is constructed, establishing a mapping benchmark between physical working conditions and digital features. During operation, valve opening data and motor output torque data are collected in real time and real-time observation vectors are generated, transforming the force characteristics of the current fluid load on the valve into a calculable mathematical vector. The feature distance between the real-time observation vector and the prior feature vector is calculated, quantifying the similarity between the current actual working condition and each preset typical working condition in the vector space. Confidence weights are assigned according to the feature distance, transforming the working condition identification process into a probability allocation process based on numerical matching degree, so that the typical working condition that is closer to the current state has a greater impact on the final result.

[0009] The fluid dynamics parameter set is weighted and summed based on confidence weights, and then synthesized using linear interpolation to obtain the actual operating condition parameters at the current moment. This constructs intermediate model parameters that can dynamically approximate the current real physical environment. These actual operating condition parameters are applied to the flow control model, updating the model's internal fluid dynamics coefficients in real time. This allows the calculation logic for flow estimation or opening adjustment to directly rely on the fluid state implied by the current torque characteristics. The entire process uses torque data as the core variable for operating condition perception, and through a calculation path of feature space mapping and parameter weighted fusion, it achieves dynamic reconstruction of the control model parameters as the operating environment changes.

[0010] Optionally, the feature vector in S1 and the real-time observation vector in S2 include at least one or more of the following feature dimensions: The average torque characteristics over the entire stroke or a portion of the stroke; The rate of change of torque with respect to valve opening; The component variance or spectral energy characteristics of the torque signal within a preset frequency band; The torque hysteresis characteristics of the valve in the opening and closing directions.

[0011] By employing the above technical solutions, the average torque characteristics within the entire or partial stroke are introduced to quantify the macroscopic intensity level of the fluid load. Utilizing the rate of change of torque with valve opening, the slope of the resistance curve is depicted, thereby identifying the torque growth trend under different media viscosities or valve types. Component variance or spectral energy characteristics within a preset frequency band are extracted, transforming turbulent oscillations or cavitation high-frequency noise, which are difficult to observe in the time domain signal, into a significant fingerprint in the frequency domain. Combining the torque hysteresis characteristics in the opening and closing directions, the asymmetry of fluid forces in reciprocating motion and the influence of mechanical transmission clearance are captured. This multi-dimensional feature extraction mechanism constructs a highly discriminative operating condition fingerprint, enabling the algorithm to accurately distinguish specific causal operating conditions based on differences in dynamic patterns when different physical phenomena produce similar instantaneous torque values.

[0012] Optionally, the sub-step of S3 includes: S31. The feature distance between the real-time observation vector and the feature vector of the typical working condition is obtained by using a spatial distance calculation algorithm; S32. Establish a negative correlation mapping relationship between the feature distance and the weight, so that the smaller the feature distance, the larger the original weight value obtained for the typical working condition; S33. Normalize the original weight values ​​obtained for all the typical working conditions so that the sum of all the confidence weights equals 1.

[0013] By employing the above technical solution, a spatial distance calculation algorithm is used to transform the difference between the real-time observation vector and the characteristic vector of typical working conditions into a quantifiable numerical distance. The geometric proximity between the current state and historical experience is measured in a multi-dimensional feature space. A negative correlation mapping relationship between feature distance and weights is established, inverting the distance metric into a contribution index, ensuring that the typical working condition closest to the physical characteristics of the current state dominates in model synthesis. The original weight values ​​are normalized, and the sum of the weights of all working conditions is constrained to a unit value, constructing a mathematically convex combination model. This calculation process ensures that the final synthesized parameters strictly fall within the effective envelope space composed of typical working condition parameters, avoiding parameter distortion caused by weight divergence and achieving a smooth transition in multi-model fusion.

[0014] Optionally, the sub-step of S2 includes: S21. Call up the pre-measured and stored data on the inherent mechanical friction torque of the valve as a function of valve opening under no-load and dry conditions; S22. Based on the current real-time valve opening of the valve, find the corresponding inherent mechanical friction torque value in the inherent mechanical friction torque data; S23. Subtract the inherent mechanical friction torque from the real-time collected motor output torque value to obtain pure fluid dynamic torque data; S24. Use the pure hydrodynamic torque data as the basis for constructing the real-time observation vector.

[0015] By employing the above technical solution, a mechanical resistance benchmark independent of the fluid medium is established by calling upon pre-measured and stored inherent mechanical friction torque data under no-load and dry conditions. The corresponding friction value is found based on the real-time valve opening and subtracted from the total motor output torque. During data preprocessing, the mechanical impedance component of the valve body itself is separated through mathematical subtraction. The separated pure fluid dynamic torque data is used as the basis for constructing the real-time observation vector, eliminating interference from non-fluid factors such as valve assembly tightness, seal aging, or mechanical corrosion on operating condition identification. This step ensures that the torque characteristics input to the algorithm are generated solely by the interaction between the fluid and the valve core, thereby improving the system's sensitivity and accuracy in identifying changes in fluid-side physical quantities such as pipeline pressure difference and medium viscosity.

[0016] Optionally, the typical operating conditions include at least two of the following types: The reference operating condition is that the medium under the reference operating condition is room temperature clean water and the pipeline pressure difference is at the standard rated value. High viscosity conditions, wherein the viscosity of the medium under high viscosity conditions is higher than a preset viscosity threshold, and the torque change rate is within a preset smooth range; Under high pressure differential conditions, the pipeline pressure differential is higher than a preset pressure differential threshold, and the amplitude of the fluid dynamic torque is higher than a preset torque threshold. Cavitation conditions, wherein the torque data signal under cavitation conditions contains noise oscillations within a specific frequency range; In the aging and wear condition, the inherent mechanical friction torque data under the aging and wear condition deviates from the factory reference data by a preset offset threshold over the entire stroke range.

[0017] By adopting the above technical solutions, a physical reference zero point was established for comparative analysis by setting the baseline operating conditions as ambient temperature clean water and standard differential pressure environment. A high-viscosity operating condition was defined, utilizing viscosity threshold judgment and the smooth characteristics of torque change rate to identify the enhanced damping phenomenon caused by medium thickening. A high-differential-pressure operating condition was defined, capturing the increase in nonlinear load caused by pipeline pressure fluctuations by monitoring the overall rise in the amplitude of fluid dynamic torque. A cavitation operating condition was defined, achieving early warning of fluid phase change and potential cavitation damage within the valve based on noise oscillations in specific frequency bands of the torque signal. An aging and wear operating condition was defined, effectively distinguishing between the mechanical performance degradation of the valve body and fluid load changes based on the overall numerical deviation of the inherent mechanical friction torque across the entire stroke range. This classification mechanism, encompassing fluid characteristics, pipeline condition, and equipment health, constructs a multi-dimensional operating condition fingerprint database, ensuring that the system can accurately summarize the physical essence of the current operating state when facing complex industrial environments.

[0018] Optionally, the set of fluid dynamic parameters includes flow coefficients; the sub-step of S5 includes: S51. Extract the flow coefficient from the synthetic parameters of the actual operating conditions; S52. Based on the current valve opening, calculate the instantaneous flow rate in reverse using fluid dynamics principles; S53. The instantaneous flow rate is directly output through the communication interface or display module of the electric actuator.

[0019] By employing the above technical solution, the flow coefficient is extracted from the weighted and fused actual operating condition parameters, yielding a fluid conduction capacity index that includes current pressure and viscosity correction information. Combined with the current valve opening and applied fluid dynamics principles for reverse calculation, the previously invisible fluid velocity is transformed into a quantifiable instantaneous flow rate value. This value is output through a communication interface or display module, endowing the electric actuator with soft measurement capabilities. This allows it to act as a virtual metering instrument in branch pipe networks without independent flow meters, providing low-cost real-time flow data support for process monitoring.

[0020] Optionally, the sub-step of S5 further includes: S54. Receive target traffic command; S55. Update the parameters of the preset inverse mapping model of flow rate and valve opening using the current actual working condition synthesis parameters; S56. Calculate the target valve opening required to achieve the target flow command; S57. Drive the electric actuator to the target valve opening degree.

[0021] By adopting the above technical solution, the system responds to the target flow command and uses the synthesized parameters of the current actual operating conditions to perform online calibration of the preset inverse mapping model, thus reconstructing the conversion logic between the flow command and the physical valve opening. The target valve opening is calculated based on the updated model parameters, enabling the generated stroke command to automatically compensate for flow deviations caused by pipeline pressure fluctuations or changes in media characteristics. The electric actuator is driven to this corrected target position, achieving adaptive closed-loop control based on the current real operating conditions. This ensures that even in non-standard pressure differentials or non-standard media environments, the actual output flow of the actuator can still accurately follow the set target of the control system.

[0022] Secondly, the computer device provided in this application adopts the following technical solution: A computer device comprising: One or more processors; Memory; One or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, the one or more applications being configured to: Perform the above-described electric actuator flow control method.

[0023] Thirdly, this application provides a computer-readable storage medium that adopts the following technical solution: A computer-readable storage medium storing a computer program that can be loaded by a processor and executed as described above.

[0024] The storage medium stores at least one instruction, at least one program, a code set, or an instruction set, wherein the at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by the processor to implement the following: The above describes a flow control method for electric actuators.

[0025] In summary, this application includes at least one of the following beneficial technical effects: 1. This application establishes a parameter synthesis mechanism based on multi-condition weighted fusion by constructing a prior database containing feature vectors of various typical operating conditions and sets of fluid dynamic parameters, and calculating the spatial distance between real-time observation vectors and prior feature vectors to allocate confidence weights. This mechanism transforms the judgment of the current unknown operating condition into a linear combination of known typical operating conditions, enabling the control model to dynamically approximate the current real physical environment. Thus, under non-standard operating conditions such as pipeline pressure fluctuations, changes in medium viscosity, or valve body mechanical wear, adaptive correction of model parameters can be achieved without adding external pressure or flow sensors, solving the problem of decreased control accuracy of a single static control model in complex environments.

[0026] 2. This application constructs a multi-dimensional operating condition fingerprint using the average torque, rate of change, frequency domain energy, and hysteresis characteristics throughout the entire stroke. Combined with a stripping technique for inherent mechanical friction data under no-load conditions, it separates the pure fluid dynamic torque from the total output torque of the motor. This signal processing method effectively eliminates interference from valve body mechanical aging, corrosion, or differences in assembly tightness on fluid operating condition identification. It ensures that the system can sensitively capture and distinguish specific physical changes caused by cavitation oscillations, high pressure differential impacts, or high viscosity damping, significantly improving the accuracy and anti-interference capability of operating condition identification.

[0027] 3. This application utilizes real-time synthesized operating parameters combined with fluid mechanics principles for back-calculation, endowing the electric actuator with dual functions of soft flow measurement and adaptive adjustment. In terms of monitoring, this method can provide visualized instantaneous flow data in branch pipe networks without independent flow meters, filling the data perception blind spot in the terminal pipe network. In terms of control, by updating the inverse mapping model of flow and opening, the system can automatically compensate for flow deviations caused by differential pressure fluctuations, ensuring that the actual flow output by the actuator accurately follows the target command, achieving low-cost and highly reliable closed-loop flow control. Attached Figure Description

[0028] Figure 1 A flowchart illustrating an embodiment of the present invention is shown. This is a flow control method for electric actuators based on multi-condition prior weighted fusion. Detailed Implementation

[0029] The present application will be further described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely illustrative of the application and are not intended to limit the scope of the application.

[0030] In the following description, numerous specific details are set forth for purposes of explanation in order to provide a thorough understanding of the inventive concept. As part of this specification, some of the accompanying drawings of this disclosure are block diagrams illustrating structures and devices to avoid complicating the disclosed principles. For clarity, not all features of the actual embodiment need to be described. Furthermore, the language used in this disclosure has been primarily chosen for readability and instructional purposes and may not have been chosen to define or limit the subject matter of the invention, thus requiring the necessary claims to determine such inventive subject matter. References to “an embodiment” or “an embodiment” in this disclosure mean that a particular feature, structure, or characteristic described in connection with that embodiment is included in at least one embodiment, and multiple references to “an embodiment” or “an embodiment” should not be construed as necessarily referring to the same embodiment.

[0031] Unless explicitly defined, the terms “a,” “an,” and “the” are not intended to refer to a singular entity, but rather to include a general category whose specific examples can be used for illustration. Therefore, the use of the terms “a” or “an” can mean any number of at least one, including “a,” “one or more,” “at least one,” and “one or more.” The term “or” means any of the options and any combination of the options, including all options unless explicitly indicated that the options are mutually exclusive. The phrase “at least one of” when combined with a list of items refers to a single item in the list or any combination of items in the list. The phrase does not require all items listed unless explicitly defined as such.

[0032] This application specifically discloses a flow control method for electric actuators based on multi-condition prior weighted fusion, referring to... Figure 1 This includes the following steps S1-S5.

[0033] S1. Pre-acquire operating data of electric actuators and valves under several typical operating conditions, and construct a priori database, wherein the priori database contains feature vectors and fluid dynamics parameter sets corresponding to each typical operating condition.

[0034] The feature vector in S1 and the real-time observation vector in S2 contain feature dimensions that include at least one or more of the following combinations: The average torque characteristics over the entire stroke or a portion of the stroke; The rate of change of torque with respect to valve opening; The component variance or spectral energy characteristics of the torque signal within a preset frequency band; The torque hysteresis characteristics of the valve in the opening and closing directions.

[0035] Step S1 establishes a priori database, aiming to establish a quantitative reference benchmark and knowledge base for operating condition identification. This step acquires standard response data of electric actuators under different physical environments through offline laboratory calibration, computational fluid dynamics simulation, or historical data cleaning. The definition of feature vectors transforms the complex dynamic interaction between fluid and valve into a computable multidimensional mathematical fingerprint, thereby realizing the transformation from qualitative operating condition description to quantitative numerical calculation, providing a data foundation for subsequent spatial distance calculation and similarity matching.

[0036] Regarding the specific selection of feature dimensions, the average torque characteristic over the entire stroke or a partial stroke reflects the macroscopic intensity level of the fluid load acting on the valve plate. The magnitude of this characteristic value is directly related to the static pressure level of the pipeline or the density properties of the medium. The torque change rate characteristic with valve opening depicts the tangent slope shape of the resistance characteristic curve, used to identify the rheological properties of the medium; for example, the damping effect of high-viscosity media usually causes the torque change rate curve to tend to be smooth, while low-viscosity, high-flow-rate media may produce abrupt torque changes at a specific opening. The component variance or spectral energy characteristics of the torque signal within a preset frequency band transform microscopic fluid disturbances that are difficult to observe directly in the time domain signal into significant fingerprints in the frequency domain. This frequency domain characteristic is particularly sensitive to high-frequency oscillations caused by turbulent pulsation or cavitation bubble collapse. The torque hysteresis characteristics of the valve in the opening and closing directions capture the combined effects of the asymmetry of the fluid dynamic torque during reciprocating motion and the mechanical transmission chain clearance, reflecting the directional differences in fluid propulsion or resistance. For example, when operating in a high-viscosity medium, the eigenvector exhibits a higher average torque value and a relatively gentle rate of change across the entire stroke range; while when operating in a cavitation condition, the eigenvector's spectral energy value in a specific high-frequency band is higher than that under normal operating conditions.

[0037] In constructing the prior database, the selection of typical working conditions covers multi-dimensional physical scenarios, including at least two of the following types: The reference operating condition is that the medium under the reference operating condition is room temperature clean water and the pipeline pressure difference is at the standard rated value. High viscosity conditions, wherein the viscosity of the medium under high viscosity conditions is higher than a preset viscosity threshold, and the torque change rate is within a preset smooth range; Under high pressure differential conditions, the pipeline pressure differential is higher than a preset pressure differential threshold, and the amplitude of the fluid dynamic torque is higher than a preset torque threshold. Cavitation conditions, wherein the torque data signal under cavitation conditions contains noise oscillations within a specific frequency range; In the aging and wear condition, the inherent mechanical friction torque data under the aging and wear condition deviates from the factory reference data by a preset offset threshold over the entire stroke range.

[0038] The overall logic of setting up multiple typical operating conditions aims to construct a complete feature space covering three dimensions: fluid characteristics, pipeline status, and equipment health. By encompassing various typical scenarios ranging from standard environments to boundary conditions and equipment degradation cycles, the system ensures that the prior database contains the main physical states that electric actuators may encounter throughout their entire lifecycle, providing sufficient interpolation benchmarks for subsequent weighted fusion algorithms. The establishment of the benchmark operating condition provides a physical reference zero point, serving as a standard quantity for comparative analysis with other abnormal or deviating operating conditions, representing the system's standard response under ideal or rated conditions.

[0039] The physical characteristics of each specific operating condition are analyzed as follows: High viscosity condition reflects changes in the rheological properties of the medium. When the medium thickens, leading to increased viscosity, the internal shear resistance of the fluid increases, enhancing the damping effect on the valve plate movement. This physical mechanism makes the rate of change of torque with opening degree tend to be flat, suppressing the agile jumps commonly seen in low viscosity media. High pressure differential condition reflects the impact of pipeline pressure fluctuations. According to fluid mechanics principles, the fluid dynamic torque is positively correlated with the pressure difference across the valve. Increased pipeline pressure causes a nonlinear increase in fluid load, directly resulting in an overall increase in the torque amplitude of the motor output throughout the entire stroke or within a specific range. Cavitation condition focuses on abnormal vibrations caused by fluid phase changes. When the local pressure is lower than the saturated vapor pressure, causing bubbles to form and collapse, the micro-jet impact induces vibrations in the valve body structure. This physical process manifests as noise oscillations within a specific frequency range in the torque signal, forming a frequency domain fingerprint distinct from normal flow. Aging and wear conditions focus on the mechanical health of the equipment itself. Wear or corrosion of mechanical parts leads to an increase in inherent frictional resistance, causing the no-load friction torque curve to drift relative to the factory reference data. This drift exists independently of the fluid load and is the basis for distinguishing between the evolution of the equipment itself and external fluid anomalies.

[0040] S2. During the operation of the electric actuator, the valve opening data and motor output torque data of the valve are collected in real time, and a real-time observation vector reflecting the characteristics of the current working condition is generated based on a preset time window.

[0041] At the hardware level, the system relies on a high-precision sensor array integrated within the electric actuator for data acquisition. Position sensors (such as precision conductive plastic potentiometers or magnetic encoders) monitor the valve stem's stroke position in real time to obtain valve opening data; current sensors (such as Hall effect sensors or shunts) or dedicated torque sensors monitor the motor's drive current or output shaft torque to obtain motor output torque data. Real-time acquisition is essential because fluid networks are typically in a dynamic state. Pulsations in upstream pumping stations or adjustments in nearby pipelines can cause transient changes and dynamic fluctuations in network pressure and flow. Only high-frequency real-time acquisition can capture these fleeting operational characteristics. A preset time window plays a crucial role in this data processing, preventing random noise from electromagnetic interference or mechanical vibrations caused by single-point sampling from misleading operational condition judgments. By accumulating and statistically analyzing data within the time window, stable trend characteristics or frequency domain characteristics of torque changes can be effectively extracted.

[0042] To improve the purity of feature extraction, the specific execution process of S2 includes sub-steps S21 to S24.

[0043] S21. Call up the pre-measured and stored data on the inherent mechanical friction torque of the valve as a function of valve opening under no-load and dry conditions.

[0044] From a physical perspective, the total torque output by the motor is actually the superposition of the mechanical friction torque between valve body components and the fluid dynamic torque that overcomes the resistance of the medium. The purpose of step S21 is to establish a mechanical impedance reference independent of the fluid medium. By performing a full-stroke scan under no-medium or no-pressure differential conditions, the mechanical differences between different valves caused by different assembly tolerances and packing tightness are eliminated.

[0045] S22. Based on the current real-time valve opening of the valve, find the corresponding inherent mechanical friction torque value in the inherent mechanical friction torque data.

[0046] The execution logic of step S22 takes into account the nonlinearity of mechanical friction characteristics. Due to the change in the geometric angle of the linkage mechanism or the local unevenness of the sealing surface, the change in mechanical friction torque with the valve stroke position is often not constant. Therefore, it is necessary to perform dynamic lookup matching based on the real-time opening to obtain accurate deduction items.

[0047] S23. Subtract the inherent mechanical friction torque from the real-time collected motor output torque value to obtain pure hydrodynamic torque data.

[0048] Step S23 achieves "mechanical-fluid decoupling" through mathematical subtraction, removing the background noise belonging to the mechanical transmission part from the total torque and separating the pure fluid dynamic torque data generated only by the interaction between the fluid and the valve plate.

[0049] S24. Use the pure hydrodynamic torque data as the basis for constructing the real-time observation vector.

[0050] Step S24 uses this pure data to construct an observation vector, which eliminates the interference of mechanical aging, corrosion or assembly differences on the identification of working conditions, and significantly improves the sensitivity of the feature vector to the identification of fluid-side physical phenomena such as changes in medium viscosity, fluctuations in pipeline pressure difference and cavitation oscillation.

[0051] S3. Calculate the feature distance between the real-time observation vector and the feature vectors of each typical working condition in the prior database, and assign a confidence weight representing the degree of matching to each typical working condition based on the magnitude of the feature distance.

[0052] Feature distance mathematically represents the geometric proximity between the current operating point and a known reference point in a multi-dimensional feature space, and physically quantifies the similarity between the current unknown state and historical experience states. The purpose of allocating confidence weights is to transform the traditional black-and-white discrete condition determination into a probabilistic fuzzy matching process. This mechanism allows the system to effectively adapt to intermediate transitional states or composite conditions that do not strictly belong to a single category. Weight allocation follows the inverse logic of distance and weight; that is, the closer the feature distance, the higher the similarity of physical characteristics, and thus a larger confidence weight is assigned. For example, assuming the current real-time observation vector is V1, if its geometric position in the feature space is closer to the feature vector of the high-pressure differential condition compared to the feature vector of the baseline condition, it indicates that the current fluid's physical behavior is more inclined towards the high-pressure differential scenario.

[0053] Specifically, the sub-steps of S3 include S31-S33.

[0054] S31. The feature distance between the real-time observation vector and the feature vector of the typical working condition is obtained by using a spatial distance calculation algorithm.

[0055] In step S31, the selection of the spatial distance calculation algorithm is based on the statistical characteristics of the feature space. For simple spaces that are homogeneous in all dimensions, Euclidean distance can be used, while Mahalanobis distance can be used to eliminate the interference of correlation between variables and scale differences.

[0056] S32. Establish a negative correlation mapping relationship between the feature distance and the weight, so that the smaller the feature distance, the larger the original weight value obtained by the typical working condition.

[0057] Step S32 uses an exponential decay function or a reciprocal function to construct a negative correlation mapping logic, ensuring that as the feature distance increases, the weight contribution rate of the corresponding working condition decreases rapidly, thereby highlighting the dominant role of highly correlated working conditions.

[0058] S33. Normalize the original weight values ​​obtained for all the typical working conditions so that the sum of all the confidence weights equals 1.

[0059] The normalization process in step S33 aims to construct a mathematically convex combination model to prevent the physical parameters synthesized later from exceeding reasonable physical limits due to weight divergence. Continuing the previous example, if the calculated distance d1 between the real-time observation vector V1 and the high-pressure differential condition is 2, and the distance d2 between it and the baseline condition is 8, after inverse mapping transformation, their original weights are 0.5 and 0.125, respectively. After normalization, the final determined confidence weights w1 and w2 are 0.8 and 0.2, respectively, and their sum is strictly equal to 1. This indicates that the algorithm determines the current condition mainly exhibits the characteristics of the high-pressure differential condition, while also possessing a small amount of baseline condition characteristics.

[0060] S4. Based on the confidence weights, perform weighted summation on the fluid dynamics parameter sets of all the typical working conditions to generate the actual working condition composite parameters at the current moment.

[0061] The physical significance of parameter weighted summation lies in constructing an intermediate-state model that dynamically approximates the current real physical environment using the principle of linear interpolation. The fluid dynamics parameter set specifically includes key indicators such as the flow coefficient Kv, flow resistance factor, and pressure correction coefficient. Compared to the traditional discrete lookup table method, this synthesis method achieves a continuous and smooth transition of parameters throughout the entire working domain, avoiding abrupt changes in control signals due to operating condition switching. Continuing with the aforementioned example of high pressure differential operating condition identification, assuming the flow coefficient Kv1 under the baseline operating condition is 100 and the corrected flow coefficient Kv2 under the high pressure differential operating condition is 80. If the current algorithm identifies the current state as having 20% ​​of the baseline operating condition characteristics (weight 0.2) and 80% of the high pressure differential operating condition characteristics (weight 0.8), the system calculates the synthesized flow coefficient Kv as 100 multiplied by 0.2 plus 80 multiplied by 0.8, resulting in 84. This synthesized value reflects the drift in the valve's effective flow capacity parameter under the current high pressure impact.

[0062] S5. The actual working condition synthesis parameters are applied to the flow control model to realize real-time estimation of fluid flow or adjustment of valve opening, wherein the set of fluid dynamic parameters includes flow coefficients.

[0063] The dynamic update mechanism of the flow control model no longer relies on fixed factory parameters, but instead loads the synthetic parameters calculated in step S4 in real time. This enables the system to have two functional modes: a soft measurement mode for flow monitoring and an adaptive adjustment mode for precise control. At the same valve opening, because the synthetic parameter Kv changes from the baseline value of 100 to 84, reflecting the specific impact of the high pressure differential environment, the real-time flow value calculated by the control model, or the target position that needs to be adjusted to achieve the target flow, will also change accordingly, thus avoiding calculation errors caused by using rigid static formulas.

[0064] Specifically, the sub-steps of S5 include S51-S57.

[0065] S51. Extract the flow coefficient from the actual operating condition synthesis parameters.

[0066] S52. Based on the current valve opening, calculate the instantaneous flow rate in reverse using fluid dynamics principles.

[0067] S53. The instantaneous flow rate is directly output through the communication interface or display module of the electric actuator.

[0068] S54. Receive target traffic command.

[0069] S55. Update the parameters of the preset inverse mapping model of flow rate and valve opening using the current actual operating condition synthesis parameters.

[0070] S56. Calculate the target valve opening required to achieve the target flow command.

[0071] S57. Drive the electric actuator to the target valve opening degree.

[0072] Steps S51 to S53 implement the virtual flow meter mode. The system uses the synthesized flow coefficient Kv and the current valve position, substitutes them into the fluid dynamics formula for calculation, thereby filling the data sensing blind spot of the branch pipeline network. Steps S54 to S57 implement the closed-loop control mode. This mode reconstructs the mapping relationship between flow command and physical opening, realizing automatic compensation for pipeline network pressure difference fluctuations or changes in medium characteristics. The continuous updating of the inverse mapping model ensures that the valve opening provided by the actuator matches the flow rate required by the system under the current operating conditions. For example, in the monitoring scenario, assuming the current valve opening is 50%, the theoretical flow rate shown in the static table should be 50 cubic meters per hour. However, because the system identifies the high pressure difference condition, according to the corrected parameter logic, the system directly outputs the corrected actual flow rate as 65 cubic meters per hour. In the control scenario, assuming the user sets the target flow rate to 50 cubic meters per hour, under the baseline operating conditions, the valve only needs to be opened to 50%. However, the system senses that it is currently in a high viscosity condition, and the increased fluid resistance leads to a decrease in flow rate. The algorithm automatically calculates that the valve needs to be opened to 60% to achieve a flow rate of 50 cubic meters per hour, and the controller then drives the motor to the 60% position.

[0073] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0074] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements an electric actuator flow control method based on multi-condition prior weighted fusion as described in the above embodiment.

[0075] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, it implements an electric actuator flow control method based on multi-condition prior weighted fusion as described in the above embodiment.

[0076] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments of this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0077] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0078] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A flow control method for an electric actuator based on multi-condition priori weighted fusion, characterized in that, Includes the following steps: S1. Pre-acquire the operating data of the electric actuator and valve under several typical working conditions, and construct a priori database, wherein the priori database contains the feature vector and fluid dynamics parameter set corresponding to each typical working condition; S2. During the operation of the electric actuator, the valve opening data and motor output torque data of the valve are collected in real time, and a real-time observation vector reflecting the characteristics of the current working condition is generated based on a preset time window; S3. Calculate the feature distance between the real-time observation vector and the feature vector of each typical working condition in the prior database, and assign a confidence weight representing the degree of matching to each typical working condition according to the magnitude of the feature distance; S4. Based on the confidence weights, perform weighted summation on the fluid dynamics parameter sets of all the typical working conditions to generate the actual working condition composite parameters at the current moment; S5. Apply the synthesized parameters of the actual operating conditions to the flow control model to achieve real-time estimation of fluid flow or adjustment of valve opening.

2. The electric actuator flow control method according to claim 1, characterized in that, The feature vector in S1 and the real-time observation vector in S2 contain feature dimensions that include at least one or more of the following combinations: The average torque characteristics over the entire stroke or a portion of the stroke; The rate of change of torque with respect to valve opening; The component variance or spectral energy characteristics of the torque signal within a preset frequency band; The torque hysteresis characteristics of the valve in the opening and closing directions.

3. The electric actuator flow control method according to claim 2, characterized in that, The sub-step of S3 includes: S31. The feature distance between the real-time observation vector and the feature vector of the typical working condition is obtained by using a spatial distance calculation algorithm; S32. Establish a negative correlation mapping relationship between the feature distance and the weight, so that the smaller the feature distance, the larger the original weight value obtained for the typical working condition; S33. Normalize the original weight values ​​obtained for all the typical working conditions so that the sum of all the confidence weights equals 1.

4. The electric actuator flow control method according to claim 3, characterized in that, The sub-step of S2 includes: S21. Call up the pre-measured and stored data on the inherent mechanical friction torque of the valve as a function of valve opening under no-load and dry conditions; S22. Based on the current real-time valve opening of the valve, find the corresponding inherent mechanical friction torque value in the inherent mechanical friction torque data; S23. Subtract the inherent mechanical friction torque from the real-time collected motor output torque value to obtain pure fluid dynamic torque data; S24. Use the pure hydrodynamic torque data as the basis for constructing the real-time observation vector.

5. The electric actuator flow control method according to claim 4, characterized in that, The typical operating conditions include at least two of the following types: The reference operating condition is that the medium under the reference operating condition is room temperature clean water and the pipeline pressure difference is at the standard rated value. High viscosity conditions, wherein the viscosity of the medium under high viscosity conditions is higher than a preset viscosity threshold, and the torque change rate is within a preset smooth range; Under high pressure differential conditions, the pipeline pressure differential is higher than a preset pressure differential threshold, and the amplitude of the fluid dynamic torque is higher than a preset torque threshold. Cavitation conditions, wherein the torque data signal under cavitation conditions contains noise oscillations within a specific frequency range; In the aging and wear condition, the inherent mechanical friction torque data under the aging and wear condition deviates from the factory reference data by a preset offset threshold over the entire stroke range.

6. The electric actuator flow control method according to claim 5, characterized in that, The set of fluid dynamic parameters includes flow coefficients; the sub-step of S5 includes: S51. Extract the flow coefficient from the synthetic parameters of the actual operating conditions; S52. Based on the current valve opening, calculate the instantaneous flow rate in reverse using fluid dynamics principles; S53. The instantaneous flow rate is directly output through the communication interface or display module of the electric actuator.

7. The electric actuator flow control method according to claim 6, characterized in that, The sub-step of S5 further includes: S54. Receive target traffic command; S55. Update the parameters of the preset inverse mapping model of flow rate and valve opening using the current actual working condition synthesis parameters; S56. Calculate the target valve opening required to achieve the target flow command; S57. Drive the electric actuator to the target valve opening degree.

8. A computer device, characterized in that, It includes: One or more processors; Memory; One or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, the one or more applications being configured to: perform the electric actuator flow control method according to any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that, The storage medium stores at least one instruction, at least one program, code set, or instruction set, wherein the at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by a processor to implement the electric actuator flow control method as described in any one of claims 1 to 7.

Images