Hydraulic system throttling control method based on hydraulic control one-way valve

Abstract

The invention discloses a hydraulic system throttling control method based on a hydraulic control one-way valve, which relates to the technical field of throttling control, and comprises the following steps of: in an initialization stage of starting a hydraulic system, setting an optimal initial throttling opening for a throttling valve based on flow data of current oil; the method comprises the following steps: performing high-frequency real-time acquisition on flow state data of oil through a sensor deployed at the upstream of a throttling area, and establishing a set of flow state data system based on a time sequence; continuously collected flow state data are divided according to time windows, and an analysis set of oil flow data is constructed. According to the method, the dynamic flowing state of oil is monitored and deeply analyzed in real time, the load state is intelligently predicted in combination with a support vector machine model, and the opening degree of the throttling valve is adjusted in real time; the problems that in the prior art, response of a throttling control strategy to a high-load working condition is delayed, and oil temperature rise and viscosity abnormal reduction are caused by serious local pressure loss of a throttling area are effectively solved.

Description

Technical Field

[0001] The invention relates to the technical field of throttling control, and in particular to a throttling control method for a hydraulic system based on a hydraulically controlled one-way valve. Background Art

[0002] The throttling control of the hydraulic system based on the hydraulic control check valve refers to a control method that realizes the precise regulation of the flow and movement speed of the hydraulic actuator (such as a hydraulic cylinder or a hydraulic motor) by setting a hydraulic control check valve in the hydraulic circuit and cooperating with a throttling element. In this system, the hydraulic control check valve allows free flow in one direction, while it needs to rely on the control oil pressure to open in the other direction, thereby limiting the flow direction of the oil. When used in conjunction with a throttle valve, a throttle valve can be set in the controlled direction of the hydraulic control check valve so that the oil can flow in a controlled manner during the reflux or oil inlet process, thereby realizing one-way throttling, two-way throttling or speed control. The advantage of this method is that it can avoid the problems of excessive back pressure and unstable flow control that may occur in traditional throttling circuits, improve the energy efficiency and control accuracy of the system, and is widely used in hydraulic systems that require two-way speed regulation or load-sensitive control.

[0003] The prior art has the following deficiencies: In existing hydraulic systems, the opening adjustment of the throttling circuit is usually controlled based on the oil flow rate, and lacks an effective response to the dynamic load change characteristics of the oil during the actual working process. Especially under high-load conditions, the pressure of the oil increases significantly before entering the throttling area. If the throttling adjustment is still performed according to the current flow control strategy, it is easy to cause a large instantaneous pressure loss in the throttling area. Since this pressure loss will be quickly converted into heat energy, the local oil temperature in the throttling area will rise sharply in a short period of time, which will cause the oil viscosity to drop sharply and the lubrication performance to drop sharply. When the lubrication effect cannot be maintained, the metal contact surface of the hydraulic component will face a dry friction state, which is prone to abnormal wear and surface ablation. In severe cases, it may lead to a decrease in system performance or even equipment damage. Therefore, the existing throttling control method is difficult to adapt to high-load dynamic change scenarios, and there are technical problems such as high energy consumption, severe heat loss, and reduced component reliability.

[0004] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0005] The purpose of the present invention is to provide a throttling control method for a hydraulic system based on a hydraulically controlled one-way valve. By real-time monitoring and in-depth analysis of the dynamic flow state of the oil, combined with a support vector machine model, intelligent prediction of the load state is performed and the throttle valve opening is adjusted in real time. The problem of the throttling control strategy in the prior art being slow to respond to high-load conditions and serious local pressure loss in the throttling area leading to abnormal temperature rise and viscosity reduction of the oil is effectively overcome, so as to solve the problems in the above-mentioned background technology.

[0006] In order to achieve the above object, the present invention provides the following technical solution: a throttling control method of a hydraulic system based on a hydraulically controlled one-way valve, comprising the following steps: In the initialization stage of the hydraulic system startup, an optimal initial throttle opening is set for the throttle valve based on the current oil flow data; During the operation of the system, the flow state data of the oil is collected in real time at high frequency through sensors deployed upstream of the throttling area, and a set of flow state data system based on time series is established to provide the original basis for subsequent modeling and analysis; The continuously collected flow state data is divided into time windows to construct an analysis set of oil flow data. Within each fixed time window, the features reflecting the oil flow under high load are extracted from the analysis set. The features are deeply processed using feature engineering technology to perform a preliminary quantitative evaluation of the current load state of the oil. The deeply processed features are input as feature vectors into the support vector machine model that has completed offline training to predict the current oil load state and determine whether the oil that is about to reach the throttling area is in an abnormal load state; When the support vector machine model predicts that there is a high load risk for the oil, the load severity is evaluated based on the output results of the model, and the initial throttling opening of the throttle valve is adaptively and dynamically adjusted based on the evaluation output. When it is predicted that the oil flow is in a high load state, the throttle valve opens wider in advance to alleviate the upcoming pressure difference mutation, and maintains the optimal initial throttling opening when the oil flow is in a light load state. Under the premise of ensuring stable system flow and sensitive response, the throttling loss and energy consumption are minimized to improve the overall efficiency and operating economy of the hydraulic system.

[0007] Preferably, the optimal initial throttle opening refers to an initial throttle valve opening value that is most suitable for the current working conditions and is comprehensively determined in the initial stage of hydraulic system operation, combining the current real-time oil flow data, system load expectations, and throttling control targets.

[0008] Preferably, the flow state data of the oil is collected in real time at high frequency by a sensor deployed upstream of the throttling area, and the specific steps are as follows: Deploy various types of high-precision sensors at key locations upstream of the throttling area, including pressure sensors, flow sensors, temperature sensors, and vibration sensors; Secondly, based on the control requirements of the hydraulic system, the high-frequency sampling rate of the sensor (e.g., millisecond or sub-millisecond level) and the data synchronization mechanism are set to ensure that the multi-dimensional flow state information is accurately recorded at the same timestamp; When the system is running, each sensor collects the flow data such as the pressure value, flow value, temperature change, vibration frequency, etc. of the oil passing through the area in real time, and immediately sends this data to the control center through the data transmission unit; Finally, after receiving the data, the control center will perform preliminary storage and real-time organization, and form a data stream collection with a unified timestamp and standard format, providing high-quality, continuous raw data information for subsequent feature extraction and intelligent prediction.

[0009] Preferably, within each fixed time window, features reflecting the flow of oil under high load are extracted from the analysis set, the extracted features include the ratio of the instantaneous peak of the flow curve to the average flow and the proportion of the energy of high-order harmonics (such as the third and above harmonics) to the fundamental wave energy after spectral analysis of the flow signal, and the extracted features are deeply processed using feature engineering technology to generate flow fluctuation peak reference values ​​and flow harmonic instability reference values, respectively, and the current load state of the oil is preliminarily quantitatively evaluated by the flow fluctuation peak reference values ​​and the flow harmonic instability reference values.

[0010] Preferably, the deeply processed flow fluctuation peak reference value and flow harmonic instability reference value are input as feature vectors into a support vector machine model that has completed offline training, a load perception index is generated by the support vector machine model, and the current load state of the oil is intelligently predicted based on the load perception index to determine whether the oil that is about to reach the throttling area is in an abnormal load state.

[0011] Preferably, the load perception index outputted in the prediction phase of the support vector machine model is compared with a set load reference threshold to determine whether the current oil is in a high load risk state. The specific steps are as follows: If the load sensing index is greater than or equal to the load reference threshold, the oil flow condition upstream of the throttling area is classified as a high load risk; if the load sensing index is less than the load reference threshold, the oil flow condition upstream of the throttling area is classified as a light load state.

[0012] Preferably, when the support vector machine model predicts that the oil has a high load risk, the initial throttle opening of the throttle valve is adaptively and dynamically adjusted in combination with the evaluation output. The specific steps are as follows: When the current oil flow is identified as a high load risk, a throttling correction factor is generated to make a positive correction to the initial throttling opening of the throttle valve. The throttling correction factor is gradually enhanced according to the degree to which the load perception index exceeds the load reference threshold to prevent sudden changes from causing system instability. The generated expression is: ,in: The load perception index generated by the support vector machine model based on the feature vector reflects the flow pressure environment that the current oil is subjected to. The load reference threshold set for the system is used to divide the high load and light load states. It is determined based on historical data and fault feature engineering. is the throttling correction factor, which is used to increase the initial throttling opening. To correct the gain coefficient and adjust the opening correction amplitude, it is usually set according to the maximum safe pressure difference of the equipment. To correct the exponent and adjust the sensitivity of the response curve, it is suitable for nonlinear amplification or suppression (such as exponential growth and buffer adjustment). Represents the degree of load overload, which serves as a reference for regulation intensity; The currently set starting throttling opening is updated according to the throttling correction factor, the throttling opening is adaptively adjusted, and the controller sends an opening adjustment command to the throttle valve actuator. The expression for the adaptive adjustment of the throttling opening is: ,in: is the initial throttling opening set originally, It is the actual throttle opening after adaptive adjustment, which is used for actual execution.

[0013] Preferably, within a fixed time window, the specific steps of generating a flow fluctuation peak reference value after deep processing the ratio of the instantaneous peak of the flow curve to the average flow using feature engineering technology are as follows: Collect data at set sampling intervals within a fixed time window N flow data points to construct a continuous flow curve function ,in Represents the real-time traffic at time t (within a fixed time window); In order to identify instantaneous spike fluctuations, the joint criterion of first-order derivative mutation and second-order derivative sign conversion is used in the time series to identify the peak point set. The identification expression is: ,in: is the first-order derivative of the flow curve, indicating the flow rate change rate, is the second-order derivative of the flow curve, indicating the acceleration of flow change and used to identify the "maximum value". is the mutation threshold, which indicates the minimum allowed mutation rate value. is the identified peak point; Define the peak weight factor for each peak point , which is used to evaluate the intensity of the peak point relative to the global traffic environment. The expression of the peak weight factor is: ,in: is a nonlinear reinforcement coefficient (e.g., 2 to 3), which is used to amplify the weight of peak points with large change rates and highlight the influence of traffic mutations under high load. j It is used to traverse the peak point set The index variable represents each index of all time points identified as “spikes”; Based on the extracted peak point set The corresponding peak weight factor , calculate the flow fluctuation peak reference value, the calculation expression is: ,in: is the instantaneous flow value at the peak moment, is the maximum flow value in the current fixed time window, is the peak response enhancement factor (usually 1 to 3), which is used to nonlinearly emphasize high flow peaks. It is the final generated traffic fluctuation peak reference value.

[0014] Preferably, within a fixed time window, the specific steps of using feature engineering technology to deeply process the proportion of high-order harmonic (such as third and higher harmonic) energy to fundamental wave energy to generate a flow harmonic instability reference value are as follows: Perform fast Fourier transform (FFT) on the collected flow signal in a fixed time window to obtain its spectrum representation, which is specifically expressed as: ,in, Indicates frequency components The corresponding amplitude (or energy intensity), For the k The harmonic frequency, n Indicates the maximum order (or maximum frequency index) in the spectrum after fast Fourier transform of the flow signal. Here corresponds to the fundamental (i.e. the main period of the lowest frequency), and is the high-order harmonic component. The high-order harmonic energy ratio vector is defined as: , m Indicates the highest harmonic order (or the maximum index of the frequency component) selected for analysis. This vector represents the energy proportion of each high-order harmonic component relative to the fundamental wave, which is a direct reflection of the nonlinearity and disturbance energy in the flow waveform. Through this normalization method with the numerator as the high frequency and the denominator as the fundamental frequency, the influence of the change of the flow amplitude itself can be effectively avoided, and only the relative characteristics of the disturbance mode and the system vibration can be focused on; In order to quantify the overall impact of high-order harmonics in the spectrum on system stability, a nonlinear weighted model based on frequency exponential decay is introduced to construct the flow harmonic instability reference value, and its formula is defined as: ,in: For the p The normalized ratio of the energy of the first harmonic to the fundamental wave, Logarithmic weighting based on harmonic order is used to enhance the weight of medium and high order harmonics. is the exponential weighting term for harmonic frequencies, normalized by the frequency ratio , which can eliminate the spectrum inconsistency caused by specific sampling rate and sampling length. is the frequency sensitivity adjustment factor, For the p The harmonic frequency, is the highest frequency in the spectrum, It is the reference value of the final output flow harmonic instability. The larger the value, the more severe the flow disturbance and the more unstable the system.

[0015] In the above technical solution, the technical effects and advantages provided by the present invention are: The present invention monitors and deeply analyzes the dynamic flow state of oil in real time, combines the support vector machine model to perform intelligent prediction of load state and adjust the throttle valve opening in real time, effectively overcoming the problems in the prior art of throttling control strategy responding slowly to high-load conditions and serious local pressure loss in the throttling area, which leads to abnormal temperature rise and viscosity reduction of the oil. Through forward-looking adaptive adjustment of the throttling opening, it can significantly reduce the instantaneous pressure loss in the throttling area while ensuring the flow stability and dynamic response sensitivity of the hydraulic system, reduce the resulting rapid heat release phenomenon, avoid the deterioration of the lubrication performance of the oil and the early wear or ablation of the hydraulic components, effectively improve the overall efficiency and economy of the hydraulic system operation, extend the service life of key components of the system, and reduce the cost of equipment downtime and maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0017] Figure 1 The present invention is a method flow chart of a throttling control method for a hydraulic system based on a hydraulically controlled one-way valve. DETAILED DESCRIPTION

[0018] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that the description of the present disclosure will be more comprehensive and complete, and the concept of the example embodiments will be fully conveyed to those skilled in the art.

[0019] The present invention provides Figure 1 The throttling control method of the hydraulic system based on the hydraulically controlled one-way valve shown comprises the following steps: In the initialization stage of the hydraulic system startup, an optimal initial throttle opening is set for the throttle valve based on the current oil flow data; The optimal initial throttle opening refers to the initial throttle valve opening value that is most suitable for the current working conditions, which is determined by combining the current real-time oil flow data, system load expectations, and throttling control targets at the initial stage of hydraulic system operation. The initial throttle opening not only affects the flow state of the oil when it initially passes through the throttling area, but is also directly related to the pressure distribution, energy loss, and temperature rise trend of the system in the early stages. Therefore, choosing a scientific and reasonable initial throttle opening helps to reduce local pressure difference fluctuations and oil shocks while ensuring the start-up responsiveness of the actuator, and effectively alleviates the potential risks of overheating and viscosity reduction in the early stage of throttling.

[0020] During the operation of the system, the flow state data of the oil is collected in real time at high frequency through sensors deployed upstream of the throttling area, and a set of flow state data system based on time series is established to provide the original basis for subsequent modeling and analysis; The flow state data of the oil is collected in high frequency and in real time by using sensors deployed upstream of the throttling area. The specific steps are as follows: Deploy various types of high-precision sensors at key locations upstream of the throttling area, including pressure sensors, flow sensors, temperature sensors, and vibration sensors; Secondly, based on the control requirements of the hydraulic system, the high-frequency sampling rate of the sensor (e.g., millisecond or sub-millisecond level) and the data synchronization mechanism are set to ensure that the multi-dimensional flow state information is accurately recorded at the same timestamp; When the system is running, each sensor collects the flow data such as the pressure value, flow value, temperature change, vibration frequency, etc. of the oil passing through the area in real time, and immediately sends this data to the control center through the data transmission unit; Finally, after receiving the data, the control center will perform preliminary storage and real-time organization, and form a data stream collection with a unified timestamp and standard format, providing high-quality, continuous raw data information for subsequent feature extraction and intelligent prediction.

[0021] The continuously collected flow state data is divided into time windows to construct an analysis set of oil flow data. Within each fixed time window, the features reflecting the oil flow under high load are extracted from the analysis set. The features are deeply processed using feature engineering technology to perform a preliminary quantitative evaluation of the current load state of the oil. When the continuously collected flow state data is divided into time windows, it is necessary to ensure that each time window can completely cover a typical oil flow behavior cycle, that is, it is representative and stable in the time dimension and can capture the main dynamic characteristic changes of the oil during this period. Specifically, the length of the time window should meet three conditions: First, the sampling density is high enough to reflect small fluctuations and abnormal trends; Second, the window period should match the dynamic response characteristics of the system. It should not be too short to cause data fragmentation and insufficient features, nor too long to cause information confusion and reduced timeliness. Third, the data within the window has a certain statistical consistency and flow continuity to avoid crossing multiple operating stages, thereby ensuring the accuracy and discriminability of subsequent feature extraction.

[0022] By setting the time window reasonably, the analysis set can retain key dynamic changes while having good data structure, which is convenient for model identification and processing.

[0023] In each fixed time window, features reflecting the flow of oil under high load are extracted from the analysis set. The extracted features include the ratio of the instantaneous peak of the flow curve to the average flow and the proportion of the energy of high-order harmonics (such as the third and above harmonics) to the fundamental wave energy after spectral analysis of the flow signal. The extracted features are deeply processed using feature engineering technology to generate flow fluctuation peak reference values ​​and flow harmonic instability reference values, respectively. The flow fluctuation peak reference values ​​and flow harmonic instability reference values ​​are used to make a preliminary quantitative evaluation of the current load state of the oil.

[0024] In the upstream of the throttling area, if the ratio of the instantaneous peak to the average flow value in the flow curve increases significantly, it usually indicates that there is a high-load oil flow phenomenon in this area. This is because under high-load conditions, when the oil flows through certain key parts of the system (such as hydraulic control valves, variable load cylinders, etc.), the local flow velocity will be unstable due to factors such as instantaneous pressure increase, valve dynamic adjustment hysteresis or pipeline compression effect, resulting in sharp fluctuation characteristics of the flow signal, that is, the "peak" is obviously prominent. Under normal flow conditions, the flow rate of the oil changes relatively smoothly, and the difference between the instantaneous peak and the average value is small. Therefore, the increase in the ratio of the peak to the average flow not only means that there is a dynamic disturbance inside the system, but is also often closely related to abnormal dynamics such as nonlinear flow caused by high load pressure and force changes at the throttling port, which can be used as an important sensitive feature for identifying high-load flow.

[0025] The specific steps of generating a reference value for a flow fluctuation peak by deeply processing the ratio of the instantaneous peak of the flow curve to the average flow rate using feature engineering technology within a fixed time window are as follows: Collect data at set sampling intervals within a fixed time window N flow data points to construct a continuous flow curve function ,in Represents the real-time traffic at time t (within a fixed time window); In order to identify instantaneous spike fluctuations, the joint criterion of first-order derivative mutation and second-order derivative sign conversion is used in the time series to identify the peak point set. The identification expression is: ,in: is the first-order derivative of the flow curve, indicating the flow rate change rate, is the second-order derivative of the flow curve, indicating the acceleration of flow change and used to identify the "maximum value". is the mutation threshold, which indicates the minimum allowed mutation rate value. is the identified peak point; The first-order derivative and second-order derivative information of the flow curve are used to determine the maximum point of the flow curve in combination with the sign change of the second-order derivative. At the same time, the change amplitude of the first-order derivative at the extreme point is constrained. When the mutation amplitude of the first-order derivative is greater than the set threshold This method can accurately extract those instantaneous abnormal flow fluctuation points caused by flow instability, load changes or other disturbances in the time series. It has good sensitivity and robustness and can provide an effective data basis for subsequent load feature extraction and unstable flow identification.

[0026] Define the peak weight factor for each peak point , which is used to evaluate the intensity of the peak point relative to the global traffic environment. The expression of the peak weight factor is: ,in: is a nonlinear reinforcement coefficient (e.g., 2 to 3), which is used to amplify the weight of peak points with large change rates and highlight the influence of traffic mutations under high load. j It is used to traverse the peak point set The index variable represents each index of all time points identified as “spikes”; The core function of this step is to accurately extract peak points with clear physical meaning from the original traffic sequence and evaluate their significance through weights.

[0027] Based on the extracted peak point set The corresponding peak weight factor , calculate the flow fluctuation peak reference value, the calculation expression is: ,in: is the instantaneous flow value at the peak moment, is the maximum flow value in the current fixed time window, is the peak response enhancement factor (usually 1 to 3), which is used to nonlinearly emphasize high flow peaks. The peak reference value of the flow fluctuation finally generated; The traffic fluctuation peak reference value integrates two core dimensions: one is the significance of the peak change rate (through The first is the relative intensity of the flow at the peak moment (reflected by a nonlinear normalization term). This can effectively avoid the failure of asymmetric fluctuations or a small number of extreme peaks, and is particularly suitable for highly sensitive identification of flow signals under dynamic loads. The higher the final flow fluctuation peak reference value, the stronger the high-load peak flow behavior of the oil in the current window, which can be used as an important input factor for load anomaly prediction.

[0028] It can be seen from the flow fluctuation peak reference value that within a fixed time window, the larger the performance value of the flow fluctuation peak reference value generated by deep processing the ratio of the instantaneous peak of the flow curve to the average flow using feature engineering technology, the higher the oil flow load upstream of the throttling area. The flow fluctuation peak reference value can sensitively reflect the strong fluctuation characteristics of the oil flow under high load conditions by multi-dimensionally strengthening the instantaneous peaks in the flow curve and comprehensively considering the change rate of the peaks, the relative flow amplitude and its nonlinear characteristics. When the performance value of the flow fluctuation peak reference value is large, it means that there are multiple high-amplitude and high-steepness flow mutation points in the current time window, reflecting the rapid flow velocity fluctuation and instantaneous impact flow caused by the strong load of the oil, which is usually closely related to the system being in a high load, high pressure difference or throttling unstable state. On the contrary, when the flow fluctuation peak reference value is small, it means that the flow curve is relatively stable, the peak change is not significant, and the system is in a low load or stable working state.

[0029] When the flow signal is spectrally analyzed upstream of the throttling area, it is found that the energy of high-order harmonics (such as the third harmonic and above) accounts for a high proportion of the fundamental wave energy, which can usually be regarded as an important signal of high-load oil flow upstream of the throttling area. The principle is that under high-load conditions, the flow of oil in the hydraulic system is often accompanied by dynamic phenomena such as nonlinear disturbances, frequent pressure fluctuations, valve micro-vibration, or sudden changes in the flow channel. These factors will introduce more frequency components into the flow signal, especially the enhancement of high-frequency components, causing the flow signal to shift from a stable state dominated by the fundamental wave to a complex waveform containing multiple harmonics. The higher the energy of the high-order harmonics, the stronger the irregularity and disturbance of the flow in the system, which is often closely related to high-load fluid characteristics (such as high-pressure drive, valve response hysteresis, pipeline oscillation, etc.). Therefore, the increase in the proportion of high-order harmonics is a reliable frequency domain characteristic indicator, which has good sensitivity and engineering practicality for identifying abnormal oil flow states under high-load conditions.

[0030] The specific steps of using feature engineering technology to deeply process the proportion of high-order harmonics (such as third-order and above harmonics) energy to fundamental wave energy in a fixed time window to generate a flow harmonic instability reference value are as follows: Perform fast Fourier transform (FFT) on the collected flow signal in a fixed time window to obtain its spectrum representation, which is specifically expressed as: ,in, Indicates frequency components The corresponding amplitude (or energy intensity), For the k The harmonic frequency, n Indicates the maximum order (or maximum frequency index) in the spectrum after fast Fourier transform of the flow signal. Here corresponds to the fundamental (i.e. the main period of the lowest frequency), and ( ) is the high-order harmonic component. The high-order harmonic energy ratio vector is defined as: , m Indicates the highest harmonic order (or the maximum index of the frequency component) selected for analysis. This vector represents the energy proportion of each high-order harmonic component relative to the fundamental wave, which is a direct reflection of the nonlinearity and disturbance energy in the flow waveform. Through this normalization method with the numerator as the high frequency and the denominator as the fundamental frequency, the influence of the change of the flow amplitude itself can be effectively avoided, and only the relative characteristics of the disturbance mode and the system vibration can be focused on; The core function of this step is to extract the energy characteristics of the disturbance structure and provide a preliminary “frequency domain map” for the risk of instability of the system flow state.

[0031] In order to quantify the overall impact of high-order harmonics in the spectrum on system stability, a nonlinear weighted model based on frequency exponential decay is introduced to construct the flow harmonic instability reference value, and its formula is defined as: ,in: For the p The normalized ratio of the energy of the first harmonic to the fundamental wave, Logarithmic weighting based on harmonic order is used to enhance the weight of medium and high order harmonics. is the exponential weighting term for harmonic frequencies, normalized by the frequency ratio , which can eliminate the spectrum inconsistency caused by specific sampling rate and sampling length. is the frequency sensitivity adjustment factor, For the p The harmonic frequency, is the highest frequency in the spectrum, It is the reference value of the flow harmonic instability of the final output. The larger the value, the more severe the flow disturbance and the more unstable the system. This step comprehensively considers the contribution of different-order harmonics to the system stability through nonlinear frequency weighted fusion, and flexibly controls the sensitivity to high-frequency components by adjusting the frequency sensitivity adjustment factor. It has stronger discrimination and engineering adaptability, and can accurately reflect whether the current oil flow presents a non-steady state under high load, thereby achieving a rapid quantitative assessment of potential risks of the system.

[0032] It can be seen from the flow harmonic instability reference value that within a fixed time window, the flow harmonic instability reference value generated after deep processing of the proportion of high-order harmonic (such as 3rd and above harmonic) energy to fundamental energy using feature engineering technology is larger, indicating that the oil flow load upstream of the throttling area is higher. This is because in the hydraulic system, high load state is usually accompanied by stronger system disturbances, nonlinear responses and throttling pressure fluctuations, resulting in more high-frequency and non-stationary components in the oil flow signal. These components are manifested in the spectrum as a significant enhancement of high-order harmonic energy, especially when the proportion of 3rd and above harmonic energy relative to fundamental energy increases, the flow state of the system has changed from stability to instability. After extracting these high-order harmonic energies through feature engineering and combining them with frequency weights for deep processing, the constructed flow harmonic instability reference value becomes an indicator that comprehensively reflects the disturbance intensity and spectrum complexity. Therefore, the larger the flow harmonic instability reference value is, the more significant the high-order disturbance in the flow is and the more violent the system fluctuation is, which is usually highly correlated with the higher flow load borne by the upstream oil. On the contrary, if the flow harmonic instability reference value is smaller, it means that the flow process is relatively stable and the system is operating in a low load or stable load state.

[0033] The deeply processed features are input as feature vectors into the support vector machine model that has completed offline training to predict the current oil load state and determine whether the oil that is about to reach the throttling area is in an abnormal load state; The deeply processed flow fluctuation peak reference value and flow harmonic instability reference value are input as feature vectors into the support vector machine model that has completed offline training. The load perception index is generated by the support vector machine model. The current load state of the oil is intelligently predicted based on the load perception index to determine whether the oil that is about to reach the throttling area is in an abnormal load state.

[0034] The support vector machine model that has completed offline training refers to a machine learning model that has been trained and optimized in advance through a large amount of historical operation data. It no longer relies on real-time training, but has been built in the development stage before operation, and has good generalization ability. It can directly intelligently identify and predict the oil load state during actual operation. As a typical supervised classification model, support vector machine (SVM) is particularly suitable for complex problems with relatively small sample size but high feature dimension and nonlinear classification boundary. In hydraulic systems, since the oil flow state is affected by multiple factors (such as throttling opening, load change, oil temperature change, flow pulsation, etc.), its high-load flow state is often difficult to accurately identify using traditional rules or single threshold judgment methods. Therefore, the introduction of the support vector machine model helps to extract implicit load risk patterns from complex nonlinear feature relationships.

[0035] Specifically, offline training refers to the systematic training of the support vector machine model using a large amount of historical oil flow data collected before the actual operation of the hydraulic system, especially data samples in high-load and non-high-load scenarios. During the training process, the researchers will use the characteristic parameters extracted from the oil flow data analysis set (such as "flow fluctuation peak reference value", "flow harmonic instability reference value", etc.) as model input variables, and input them into the support vector machine for learning with the corresponding labels (such as "high load" or "normal load"). Through continuous iteration and optimization, the SVM model will learn how to find an optimal hyperplane in the feature space to maximize the distinction between sample points of different categories (i.e., different load states). In this process, the model will also automatically select "support vectors", that is, the key training sample points closest to the classification boundary, to determine the position and direction of the classification boundary. These support vectors constitute the core basis for the model to distinguish the input data, so that in practical applications, even in the face of new sample data, accurate predictions can be made.

[0036] The trained SVM model has strong predictive capabilities, and due to its relatively simple model structure and fast inference speed, it is suitable for deployment in real-time control scenarios. A significant advantage of offline training is that the training process can be carried out in a resource-rich environment (such as an engineering background or laboratory), and more complex parameter optimization strategies, larger data sets, and more sophisticated feature engineering methods can be used to improve model accuracy without worrying about the computational burden at runtime. Therefore, when deploying the actual system, it is only necessary to embed the trained model into the system control logic in the form of a mathematical function to achieve low-latency, low-power, and high-precision intelligent prediction of the load state of the newly inflowing oil.

[0037] In this system, the "flow fluctuation peak reference value" and "flow harmonic instability reference value" are used as input feature vectors, which can keenly reflect the nonlinear flow disturbance and frequency distortion characteristics of the oil when it encounters high load. These phenomena are often difficult to intuitively judge with traditional flow values ​​or pressure values. By inputting these complex but discriminative features into the trained SVM model, the model can establish a clear classification boundary between "high load" and "normal load". When a new set of oil state data is input, the system can quickly determine whether its current flow state meets the "abnormal load" characteristics, and then send out a control signal in advance to avoid the violent heat release caused by the sudden change of pressure difference after the oil enters the throttling area.

[0038] "Offline training completed" means that the SVM model no longer needs to be trained or adaptively learned during the application phase, so as to ensure that its prediction response speed meets the high-frequency control requirements of the hydraulic system. This model training-deployment separation method helps to ensure the stability and controllability of the system. It is also convenient to replace model parameters or retrain new model versions when the system is upgraded or new working conditions are expanded, so as to achieve modular maintenance and intelligent iteration. Therefore, the SVM model that has completed offline training is not only an intelligent load identification tool, but also a key component for the upgrade of hydraulic systems to intelligent control. In short, the offline trained support vector machine model is suitable for quickly and reliably judging whether the oil is in an abnormal load state in the hydraulic throttling control scenario with its accurate classification ability and high operating efficiency. When combined with the deeply processed feature input, the model can maximize the "load signal" in the flow data, provide timely and accurate decision support for throttling opening adjustment and system thermal protection mechanism, and significantly improve the intelligent control level and operating stability of the hydraulic system in a high-load dynamic environment.

[0039] The support vector machine model is not specifically limited here, and can realize the reference value of traffic fluctuation peak and flow harmonic instability reference value Perform comprehensive analysis to generate load-aware metrics In order to realize the technical solution of the present invention, the present invention provides a specific implementation method. Load perception index The generated expression is: , where , are the flow fluctuation peak reference values and flow harmonic instability reference value The preset scaling factor of , All are greater than 0. Preset proportional coefficient and It means that in the process of calculating the load perception index, the traffic fluctuation peak reference value is given separately and flow spectrum wave instability index These coefficients are pre-set according to the system's assessment of the impact, sensitivity, and importance of the two features in load perception. In other words, and The contribution ratio of these two features to the final perception index is controlled. For example, if the traffic spike change in the system reflects the high load trend better than the spectrum instability, then ; and vice versa. The setting of preset proportional coefficients is usually based on historical data analysis, empirical judgment or model training optimization results, and is an important parameter to ensure the accuracy, sensitivity and engineering adaptability of load sensing indicators. In actual applications, these coefficients may be adjusted and optimized according to different equipment, system conditions or control objectives to adapt to specific application scenarios.

[0040] It can be seen from the load perception index that within a fixed time window, the greater the performance value of the flow fluctuation peak reference value generated after deep processing of the ratio of the instantaneous peak to the average flow of the flow curve using feature engineering technology, the greater the performance value of the flow harmonic instability reference value generated after deep processing of the proportion of high-order harmonics (such as third and above harmonics) energy to fundamental wave energy using feature engineering technology, that is, the greater the performance value of the load perception index generated when the current oil load state is predicted by the support vector machine model that has completed offline training, the higher the oil flow load upstream of the throttling area, and vice versa.

[0041] When the support vector machine model predicts that the oil has a high load risk, the load severity is evaluated in combination with the output of the model, and the initial throttle opening of the throttle valve is adaptively and dynamically adjusted in combination with the evaluation output. When the oil flow is predicted to be in a high load state, the throttle valve opens wider in advance to alleviate the upcoming pressure difference mutation, and maintains the best initial throttle opening when the oil flow is in a light load state. Under the premise of ensuring the stability of the system flow and sensitive response, the throttling loss and energy consumption are minimized to the greatest extent, thereby improving the overall efficiency and operation economy of the hydraulic system. Compare the load perception index output by the support vector machine model prediction stage with the set load reference threshold to determine whether the current oil is in a high load risk state. The specific steps are as follows: If the load sensing index is greater than or equal to the load reference threshold, the oil flow condition upstream of the throttling area is classified as a high load risk; if the load sensing index is less than the load reference threshold, the oil flow condition upstream of the throttling area is classified as a light load state.

[0042] When the current oil flow is identified as a high load risk, a throttling correction factor is generated to make a positive correction to the initial throttling opening of the throttle valve. The throttling correction factor is gradually enhanced according to the degree to which the load perception index exceeds the load reference threshold to prevent sudden changes from causing system instability. The generated expression is: ,in: The load perception index generated by the support vector machine model based on the feature vector reflects the flow pressure environment that the current oil is subjected to. The load reference threshold set for the system is used to divide the high load and light load states. It is determined based on historical data and fault feature engineering. is the throttling correction factor, which is used to increase the initial throttling opening. To correct the gain coefficient and adjust the opening correction amplitude, it is usually set according to the maximum safe pressure difference of the equipment. To correct the exponent and adjust the sensitivity of the response curve, it is suitable for nonlinear amplification or suppression (such as exponential growth and buffer adjustment). Represents the degree of load excess and serves as a reference for regulation intensity.

[0043] This step is used to build a quantitative and adaptive throttling correction mechanism, so that the system can adjust the throttling strategy according to the "severity" of high load, with sensitive but smooth response capabilities.

[0044] The currently set starting throttling opening is updated according to the throttling correction factor, the throttling opening is adaptively adjusted, and the controller sends an opening adjustment command to the throttle valve actuator. The expression for the adaptive adjustment of the throttling opening is: ,in: is the initial throttling opening set originally, It is the actual throttle opening after adaptive adjustment, which is used for actual execution; This step is the final execution and implementation control link. By updating the physical opening of the throttle valve in real time, pressure difference buffering is completed before the oil enters the throttling area, effectively suppressing the instantaneous release of heat, reducing energy consumption and wear, and ensuring flow response sensitivity.

[0045] By real-time monitoring and in-depth analysis of the dynamic flow state of the oil, combined with the support vector machine model to intelligently predict the load state and adjust the throttle valve opening in real time, the problem of throttling control strategy in the existing technology responding slowly to high-load conditions and serious local pressure loss in the throttling area, which leads to abnormal temperature rise and viscosity reduction of the oil, is effectively overcome; at the same time, through the forward-looking adaptive adjustment of the throttling opening, the instantaneous pressure loss in the throttling area can be significantly reduced while ensuring the flow stability and dynamic response sensitivity of the hydraulic system, and the rapid release of heat caused by this can be reduced, thereby avoiding the deterioration of the lubrication performance of the oil and the early wear or ablation of the hydraulic components, thereby effectively improving the overall efficiency and economy of the hydraulic system operation, extending the service life of key components of the system, and reducing the cost of equipment downtime and maintenance. It has obvious practical engineering application value of energy saving and consumption reduction as well as safe and stable operation.

[0046] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0047] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

[0048] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A throttling control method for a hydraulic system based on a hydraulically controlled one-way valve, characterized in that: The following steps are involved: In the initialization phase of the hydraulic system startup, an optimal initial throttle opening is set for the throttle valve based on the current oil flow data; By deploying sensors upstream of the throttling area, the flow state data of the oil is collected in real time at high frequency, and a flow state data system based on time series is established; The continuously collected flow state data is divided into time windows to construct an analysis set of oil flow data. Within each fixed time window, the features reflecting the oil flow under high load are extracted from the analysis set. The features are deeply processed using feature engineering technology to perform a preliminary quantitative evaluation of the current load state of the oil. The deeply processed features are input as feature vectors into the support vector machine model that has completed offline training to predict the current oil load state and determine whether the oil that is about to reach the throttling area is in an abnormal load state; When the support vector machine model predicts that there is a high load risk in the oil flow, the load severity is evaluated in combination with the output of the model, and the initial throttle opening of the throttle valve is adaptively and dynamically adjusted in combination with the evaluation output.

2. The throttling control method of a hydraulic system based on a hydraulically controlled one-way valve according to claim 1, characterized in that: The optimal initial throttle opening refers to the initial throttle valve opening value that is most suitable for the current working conditions, which is comprehensively determined in the initial stage of hydraulic system operation, combining the current real-time oil flow data, system load expectations and throttling control targets.

3. The throttling control method of a hydraulic system based on a hydraulically controlled one-way valve according to claim 1, characterized in that: The flow state data of the oil is collected in real time at high frequency by using sensors deployed upstream of the throttling area. The specific steps are as follows: Deploy various types of high-precision sensors at key locations upstream of the throttling area; Based on the control requirements of the hydraulic system, the high-frequency sampling rate of the sensor and the data synchronization mechanism are set to accurately record multi-dimensional flow state information at the same timestamp; The flow data of oil passing through the upstream area is collected in real time through various sensors, and the collected data is sent to the control center through the data transmission unit; After receiving the data, the control center performs preliminary storage and real-time sorting, and forms a data stream collection with a unified timestamp and standard format.

4. The throttling control method of a hydraulic system based on a hydraulically controlled one-way valve according to claim 1, characterized in that: In each fixed time window, features reflecting the flow of oil under high load are extracted from the analysis set. The extracted features include the ratio of the instantaneous peak of the flow curve to the average flow and the proportion of high-order harmonic energy to fundamental energy after spectral analysis of the flow signal. The extracted features are deeply processed using feature engineering technology to generate flow fluctuation peak reference values ​​and flow harmonic instability reference values, respectively. The flow fluctuation peak reference values ​​and flow harmonic instability reference values ​​are used to make a preliminary quantitative evaluation of the current load state of the oil.

5. The throttling control method of a hydraulic system based on a hydraulically controlled one-way valve according to claim 4, characterized in that: The deeply processed flow fluctuation peak reference value and flow harmonic instability reference value are input as feature vectors into the support vector machine model that has completed offline training. The load perception index is generated by the support vector machine model. The current load state of the oil is intelligently predicted based on the load perception index to determine whether the oil that is about to reach the throttling area is in an abnormal load state.

6. The throttling control method of a hydraulic system based on a hydraulically controlled one-way valve according to claim 5, characterized in that: Compare the load perception index output by the support vector machine model prediction stage with the set load reference threshold to determine whether the current oil is in a high load risk state. The specific steps are as follows: If the load sensing index is greater than or equal to the load reference threshold, the oil flow condition upstream of the throttling area is classified as a high load risk; if the load sensing index is less than the load reference threshold, the oil flow condition upstream of the throttling area is classified as a light load state.

7. The throttling control method of a hydraulic system based on a hydraulically controlled one-way valve according to claim 6, characterized in that: When the support vector machine model predicts that the oil has a high load risk, the initial throttle opening of the throttle valve is adaptively and dynamically adjusted based on the evaluation output. The specific steps are as follows: When the current oil flow is identified as a high load risk, a throttling correction factor is generated to make a positive correction to the initial throttling opening of the throttle valve. The throttling correction factor is enhanced according to the degree to which the load perception index exceeds the load reference threshold. The generated expression is: ,in: The load perception index generated by the support vector machine model based on the feature vector reflects the flow pressure environment that the current oil is subjected to. The load reference threshold set for the system, is the throttling correction factor, which is used to increase the initial throttling opening. To correct the gain coefficient and adjust the opening correction amplitude, To correct the index, adjust the sensitivity of the response curve; The currently set starting throttling opening is updated according to the throttling correction factor, the throttling opening is adaptively adjusted, and the controller sends an opening adjustment command to the throttle valve actuator. The expression for the adaptive adjustment of the throttling opening is: ,in: is the initial throttling opening set originally, It is the actual throttle opening after adaptive adjustment, which is used for actual execution.

8. The throttling control method of a hydraulic system based on a hydraulically controlled one-way valve according to claim 4, characterized in that: The specific steps of generating a reference value for a flow fluctuation peak by deeply processing the ratio of the instantaneous peak of the flow curve to the average flow rate using feature engineering technology within a fixed time window are as follows: Collect data at set sampling intervals within a fixed time window N flow data points to construct a continuous flow curve function ,in represents the real-time traffic at time t; In the time series, the joint criterion of the first-order derivative mutation and the second-order derivative sign conversion is used to identify the peak point set. The expression for identification is: ,in: is the first-order derivative of the flow curve, indicating the flow rate change rate, is the second-order derivative of the flow curve, indicating the acceleration of flow change. is the mutation threshold, which indicates the minimum allowed mutation rate value. is the identified peak point; Define the peak weight factor for each peak point , which is used to evaluate the intensity of the peak point relative to the global traffic environment. The expression of the peak weight factor is: ,in: is the nonlinear enhancement coefficient, which highlights the influence of sudden flow changes under high load. j It is used to traverse the peak point set An index variable representing each index of all time points identified as "spikes"; Based on the extracted peak point set The corresponding peak weight factor , calculate the flow fluctuation peak reference value, the calculation expression is: ,in: is the instantaneous flow value at the peak moment, is the maximum flow value in the current fixed time window, is the peak response enhancement factor, which is used to nonlinearly emphasize high flow spikes. It is the final generated traffic fluctuation peak reference value.

9. The throttling control method of a hydraulic system based on a hydraulically controlled one-way valve according to claim 4, characterized in that: The specific steps of generating the reference value of flow harmonic instability by deeply processing the proportion of high-order harmonic energy to fundamental wave energy using feature engineering technology within a fixed time window are as follows: The collected flow signal is fast Fourier transformed in a fixed time window to obtain its spectrum representation, which is specifically expressed as: ,in, Indicates frequency components The corresponding amplitude, For the k The harmonic frequency, n It represents the maximum order in the spectrum after fast Fourier transform of the flow signal. Corresponding to the fundamental wave, is the high-order harmonic component, , define the high-order harmonic energy ratio vector as: , m Indicates the highest harmonic order selected for analysis; A nonlinear weighted model based on frequency exponential decay is introduced to construct the flow harmonic instability reference value. The constructed expression is: ,in: For the p The normalized ratio of subharmonic to fundamental energy, Logarithmic weighting based on harmonic order is used to enhance the weight of medium and high order harmonics. is the exponential weighting term for the harmonic frequencies, is the frequency sensitivity adjustment factor, For the p The harmonic frequency, is the highest frequency in the spectrum, It is the reference value of flow harmonic instability of the final output.

Images