A method and system for monitoring leaks in a hydraulic system

Abstract

The present application relates to the field of hydraulic system monitoring, and more particularly to a kind of hydraulic system's leakage monitoring method and system, method includes: obtaining the pressure data of hydraulic servo system valve core zero state;With each pressure data point as center to construct local analysis window, the residual component is obtained by local mean decomposition processing to the pressure data, the local mean decomposition processing includes using adaptive sliding step to data point;The average decline rate of the residual component is calculated;In response to the average decline rate is greater than set threshold, generate alarm signal.The present application improves the sensitivity and accuracy of the detection of hydraulic valve core zero small internal leakage, effectively avoids false alarm and miss.

Description

Technical Field

[0001] This invention relates to the field of hydraulic system monitoring, and more particularly to a method and system for monitoring leaks in hydraulic systems. Background Technology

[0002] Hydraulic servo systems, as the core unit for precise control in coal mining machinery and heavy hydraulic equipment, directly affect the efficiency and safety of mining operations due to their operational stability. Theoretically, when the valve core of a hydraulic servo system is in the zero-position state, the oil circuit should be completely cut off, maintaining constant pressure in the actuator. However, in the harsh working conditions underground, the valve core is highly susceptible to internal leakage at the zero position due to minor wear or oil contamination. This internal leakage is initially highly insidious, manifesting as a slow pressure decay that is extremely difficult to detect, but it can lead to equipment pressure holding failure, sluggish action, and even major safety accidents such as hydraulic support instability. Therefore, achieving early and accurate monitoring of valve core internal leakage at the zero position has extremely high practical value. A Local Mean Decomposition (LMD) algorithm for non-stationary signal processing is available. It can adaptively decompose complex signals into several product functions, exhibiting excellent performance in extracting nonlinear trend features and is often used in the field of mechanical fault diagnosis.

[0003] When using the LMD algorithm to analyze zero-point pressure data of a hydraulic servo system, the algorithm typically employs a fixed sliding step size to smooth local extrema during the calculation of the local mean function and envelope estimation function. This fixed sliding step size mechanism is ill-suited to the signal characteristics of hydraulic systems in complex electromagnetic environments. Electromagnetic interference in industrial settings usually manifests as high-frequency, dense clutter, requiring a larger sliding step size to eliminate its effects. However, the true internal leakage of a hydraulic servo system is often accompanied by subtle pressure trend changes or physically significant pressure surges occurring at the moment of valve spool movement. This necessitates a smaller sliding step size to preserve its edge characteristics and details. If the sliding step size is set too small, electromagnetic interference may be mistakenly identified as valid signal characteristics and preserved, leading to false alarms from the monitoring system. Conversely, if the sliding step size is set too large, the true internal leakage trend or pressure surge will be excessively smoothed and lost, resulting in missed alarms. Summary of the Invention

[0004] To address the problem that the LMD algorithm, which uses a fixed sliding step size, is difficult to adapt to the needs of monitoring hydraulic systems, this invention provides a method and system for monitoring leakage in hydraulic systems.

[0005] In a first aspect, the present invention provides a method for monitoring leakage in a hydraulic system, employing the following technical solution: A method for detecting leakage in a hydraulic system, the method comprising: Acquire pressure data of the valve core in the zero-position state of the hydraulic servo system; construct a local analysis window centered on each pressure data point, and perform local mean decomposition processing on the pressure data to obtain residual components. The local mean decomposition processing includes using an adaptive sliding step size for the data points; the calculation method of the adaptive sliding step size includes: calculating the effective noise dominance factor of each data point; the effective noise dominance factor is positively correlated with the local fluctuation density; the local fluctuation density is positively correlated with the number of sign changes of the first-order difference and the sum of squares of the second-order difference within the local analysis window; calculate the average descent rate of the residual components; and generate an alarm signal in response to the average descent rate exceeding a set threshold.

[0006] By constructing an effective noise-dominant factor related to local fluctuation density and second-order difference characteristics, high-frequency electromagnetic interference and actual pressure changes in signals can be automatically identified. Increasing the step size in noise-dominant regions for strong noise reduction, and decreasing the step size in trend-changing regions caused by leakage or impact to preserve details, improves the sensitivity and accuracy of detecting minute internal leaks at the zero position of hydraulic valve cores in complex industrial electromagnetic environments, effectively avoiding false alarms and missed alarms.

[0007] Preferably, the expression for the effective noise dominance factor is: ; In the formula, This represents the effective noise dominance factor at the i-th pressure data point. This represents the local fluctuation density at the i-th pressure data point. , These represent the values ​​of the j-th and (j+1)-th pressure data points, respectively. The hyperparameters are defined as follows: exp() represents an exponential function with base e, and N represents half the length of the local analysis window.

[0008] By introducing an exponential function that incorporates the algebraic sum of the differences between adjacent points and the ratio of their absolute values, the effective noise dominance factor is calculated, effectively utilizing the trend consistency characteristics of the signal. This calculation method can distinguish between real hydraulic shocks that fluctuate dramatically but have directionality and disordered electromagnetic noise, preventing the algorithm from mistaking real physical pressure changes for noise and performing excessive smoothing, thus further ensuring the integrity of leakage feature extraction.

[0009] Preferably, the expression for the local fluctuation density is:

[0010] In the formula, This represents the local fluctuation density at the i-th pressure data point. , , These represent the values ​​of the (j-1), j-th, and j+1-th pressure data points, respectively. This is an indicator function that returns 1 if the condition within the parentheses is true, and 0 otherwise. denoted as the second-order difference value at the j-th pressure data point, and norm represents the normalization function.

[0011] By combining the number of sign changes of the first-order difference with the sum of squares of the second-order difference to quantify the density of local fluctuations, the oscillation frequency and energy intensity of the signal can be accurately captured at the microscopic level. The second-order difference can reflect the sharpness of waveform peaks, thus more accurately identifying interference areas with concentrated high-frequency energy. This provides a reliable quantitative basis for the subsequent accurate calculation of the adaptive step size and enhances the ability to perceive complex high-frequency noise.

[0012] Preferably, the expression for the adaptive sliding step size is:

[0013] In the formula, This represents the optimized adaptive sliding step size at the i-th pressure data point. This represents the minimum sliding step size preset by the LMD algorithm. This represents the maximum sliding step size preset by the LMD algorithm. This represents the effective noise dominance factor at the i-th pressure data point. This represents the rounding function.

[0014] By performing weighted calculations and rounding between preset minimum and maximum sliding step sizes, the generated sliding step size is ensured to have a reasonable range in a physical sense, and can be smoothly and dynamically adjusted as the local characteristics of the signal change, thus guaranteeing the stability and computational efficiency of the algorithm.

[0015] Preferably, the minimum sliding step size ranges from [2, 5]; the maximum sliding step size ranges from [20, 30].

[0016] This range setting prevents noise residue caused by excessively small step sizes, and also prevents signal distortion caused by excessively large step sizes, ensuring optimal performance matching of the algorithm in practical engineering applications.

[0017] Preferably, the method for constructing a local analysis window is as follows: taking the corresponding data point as the center, extract N points before and after it to form a window of length [missing information]. The local analysis window.

[0018] A local analysis window construction method was established with the current data point as the center and N points before and after it. This ensures that there are enough data samples to statistically analyze local features when calculating fluctuation density and noise factor. At the same time, the window length is moderate, which avoids the local transient features being averaged due to the window being too large, thus balancing statistical accuracy and temporal resolution.

[0019] Preferably, the threshold is set to 0.05 MPa / min.

[0020] A specific threshold for the average rate of pressure drop was set, providing a clear quantitative standard for determining whether internal leakage exists.

[0021] Preferably, the method for acquiring pressure data is as follows: a pressure sensor is installed at the rodless chamber interface of the hydraulic cylinder of the hydraulic servo system; when the hydraulic servo system is in a pressure-holding condition, the pressure sensor is used to collect pressure signals, and the pressure signals are converted from analog to digital to obtain the pressure data.

[0022] The rodless chamber is a key stress-bearing part of the hydraulic support. Data collected here can most directly reflect the locking performance of the valve core. Combined with the data collection conditions under pressure-holding conditions, the validity and relevance of the data can be guaranteed from the source, eliminating interference during dynamic operation.

[0023] Preferably, the frequency range for acquiring pressure signals using the pressure sensor is [1.5, 2.5] kHz.

[0024] By limiting the frequency range of the acquired signals to a higher frequency band, richer details of transient pressure fluctuations can be captured compared to conventional low-frequency acquisition.

[0025] Secondly, the present invention provides a leakage monitoring system for a hydraulic system, which adopts the following technical solution: A leakage monitoring system for a hydraulic system includes a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a leakage monitoring method for a hydraulic system as described above is implemented.

[0026] The above-mentioned method for monitoring leakage in a hydraulic system is used to generate a computer program, which is then stored in a memory for loading and execution by a processor. This allows for the creation of a system based on the memory and processor, making it convenient to use.

[0027] The present invention has the following technical effects: 1. By introducing an adaptive sliding step size mechanism, this invention solves the problem of balancing denoising and feature preservation in traditional LMD algorithms. Specifically, the method first identifies the local fluctuation density of the signal to preliminarily determine its complexity. Addressing the special case where real pressure surges and electromagnetic interference in hydraulic systems are extremely similar in fluctuation density, leading to misjudgments, this invention further introduces trend consistency analysis. By distinguishing between chaotic oscillations and directional accumulation, a robust effective noise dominance factor is constructed. Based on this factor, the sliding step size is automatically increased in regions dominated by electromagnetic interference to powerfully filter out noise; in regions dominated by real internal leakage or pressure surges, the sliding step size is automatically decreased to fully preserve the physical characteristics of the signal. This dynamic adjustment strategy improves the decomposition accuracy of the LMD algorithm in highly turbulent industrial environments, enabling the monitoring system to sensitively detect minute internal leakage at the valve core zero position, effectively avoiding false alarms caused by environmental interference, and improving the maintenance efficiency and operational safety of hydraulic systems in coal mine machinery.

[0028] 2. This invention constructs an effective noise-dominant factor that includes the characteristics of local fluctuation density and trend consistency. By dynamically adjusting the sliding step size through this factor, the step size is increased to remove noise at disordered electromagnetic noise locations and decreased to preserve fidelity at locations with real hydraulic shock or leakage trends. This effectively solves the problem of difficulty in extracting weak internal leakage signals under high-frequency interference and improves the monitoring accuracy. Attached Figure Description

[0029] Figure 1 This is a flowchart of a leakage monitoring method for a hydraulic system according to the present invention.

[0030] Figure 2 This is a comparison diagram of the effects of the present invention and the prior art. Detailed Implementation

[0031] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0032] This invention discloses a method for monitoring leakage in a hydraulic system, referring to... Figure 1 This includes the following steps: S1: Collect pressure data of the valve core in the zero position state of the hydraulic servo system.

[0033] High-frequency pressure sensors are installed at critical pressure-holding points in the hydraulic servo system, such as at the rodless chamber interface of the hydraulic cylinder. When the hydraulic equipment is in standby or pressure-holding mode, i.e., the control system commands the valve spool to be in the zero position, and data acquisition is initiated. The acquisition frequency is set to 2kHz to ensure that transient pressure fluctuations can be captured. The acquisition duration is one complete pressure-holding cycle, for example, 30 minutes. Subsequently, a high-precision analog-to-digital converter is used to convert the analog signal into a digital pressure data sequence to obtain the pressure data to be analyzed.

[0034] S2: Analyze the local fluctuation characteristics and directional evolution characteristics of the pressure data, and calculate the optimized adaptive sliding step size for each data point when decomposing the pressure data using the LMD algorithm.

[0035] S21: Analyze the intensity of oscillations in the pressure data at the micro level to obtain the local fluctuation density at each pressure data point.

[0036] The purpose of analyzing the local fluctuation density at each pressure data point in this step is to preliminarily quantify the complexity of the signal surrounding the current data point. In hydraulic monitoring scenarios, electromagnetic interference typically manifests as frequent up-and-down signal jumps within a very short time, exhibiting significant curvature. Therefore, during analysis, the more frequently the data sign changes within the neighborhood of a data point, and the larger the second-order difference of the data, the more severe the signal oscillation in that area, the greater the likelihood of it being affected by high-frequency interference, and the greater the corresponding local fluctuation density.

[0037] Taking each pressure data point being analyzed as the center, extract N points before and after it, for example, N=15, to form a data point of length [missing information]. The local analysis window is then used to calculate the local fluctuation density at each pressure data point, expressed as:

[0038] In the formula, This represents the local fluctuation density at the i-th pressure data point. , , These represent the values ​​of the (j-1), j-th, and j+1-th pressure data points, respectively. This is an indicator function. It takes the value 1 when the condition in parentheses is true (i.e., the direction of data change at point j is reversed, forming a peak or trough), and takes the value 0 otherwise. This represents the second-order difference value at the j-th pressure data point, and norm represents the normalization function, specifically a linear normalization function.

[0039] In the formula The larger the value, the more times the direction of pressure changes within the local analysis window of the i-th data point, indicating more frequent oscillations and a more unstable signal with greater local fluctuation density. (In the formula...) The larger the value, the greater the energy of the second derivative of the data within the local window, meaning the sharper the peak of the waveform, indicating stronger high-frequency energy contained in the signal, further confirming the greater density of local fluctuations.

[0040] S22: Analyze the cumulative trend characteristics of pressure data within a local range, and combine the local fluctuation density to optimize and obtain the effective noise dominance factor at each pressure data point.

[0041] In hydraulic systems, sudden internal leakage or external load impacts can cause drastic changes in pressure data within a short period, leading to an overestimation of local fluctuation density. If the sliding step size is increased solely based on local fluctuation density, the LMD algorithm will smooth out these real physical impacts, resulting in missed detections. Therefore, optimization is needed to address this situation. The difference between hydraulic shock and electromagnetic noise lies in the fact that while hydraulic shock changes drastically, it has a clear directionality, exhibiting a consistent trend within a local area; whereas electromagnetic noise is disordered, with local changes canceling each other out and lacking a trend. Therefore, the higher the proportion of the algebraic sum of the differences between all adjacent points within a local window relative to the sum of their absolute differences, the more consistent the direction of change in that area, and the greater the likelihood of a real hydraulic shock. In this case, even with a high local fluctuation density, the noise dominance factor should be reduced to prevent over-smoothing. Conversely, the lower the proportion of the algebraic sum of the differences between all adjacent points within a local window relative to the sum of their absolute differences, the more likely the data is oscillating in place, and the greater the likelihood of electromagnetic noise, resulting in a larger effective noise dominance factor. Based on this principle, the effective noise dominance factor at each pressure data point is calculated, and the expression is:

[0042] In the formula, This represents the effective noise dominance factor at the i-th pressure data point. This represents the local fluctuation density at the i-th pressure data point. , These represent the values ​​of the j-th and (j+1)-th pressure data points, respectively. The hyperparameters are set. To prevent the denominator from being 0, exp() represents an exponential function with base e.

[0043] In the formula This value reflects the consistency of local data trends. A larger value indicates a strong, monotonically increasing or decreasing trend within the local window, suggesting a higher probability of genuine hydraulic shocks and a smaller effective noise factor. This means that for genuine hydraulic shocks, the algorithm will subsequently allocate a smaller sliding step size to preserve their waveform characteristics. Conversely, a smaller value indicates chaotic data changes, suggesting a higher probability of electromagnetic noise and a larger effective noise factor. The algorithm will then allocate a larger sliding step size to filter out this noise.

[0044] S23: Based on the effective noise dominance factor at each pressure data point, calculate the optimized adaptive sliding step size at each data point when using the LMD algorithm.

[0045] The larger the effective noise dominance factor at each data point, the more disordered the electromagnetic interference in the current region is. To obtain a clean local mean function, the LMD algorithm requires a larger sliding step size for strong smoothing. Conversely, the smaller the effective noise dominance factor, the more stable the signal in the current region or the more important physical abrupt changes it contains. The LMD algorithm requires a smaller sliding step size to finely depict the signal profile. Based on this principle, the optimized adaptive sliding step size at each pressure data point is calculated, expressed as:

[0046] In the formula, This represents the optimized adaptive sliding step size at the i-th pressure data point. This represents the minimum sliding step size preset by the LMD algorithm. , This represents the maximum sliding step size preset by the LMD algorithm. , This represents the effective noise dominance factor at the i-th pressure data point. This represents the rounding function.

[0047] S3: Use the optimized LMD algorithm to decompose the pressure data, extract leakage trend items, and perform anomaly monitoring.

[0048] The sliding step size parameter in the LMD algorithm is set to the dynamic sequence calculated in the above steps. (M is the total number of data points), using an adaptive sliding step size Calculate the local mean function and envelope estimation function to decompose the pressure data into several product functions (PF) and a residual component. Among them, high-frequency electromagnetic interference will be separated into the first few PF components, while the pressure change trend caused by internal leakage at the zero position of the valve core will mainly remain in the residual component. In the middle, for the residual components Perform slope analysis and calculate The average rate of decrease within the monitoring period (i.e., one complete pressure holding cycle in step S1). Set an internal leakage rate threshold. .like If the system detects a valve core leak at zero position, it will determine that the hydraulic servo system has an internal leakage at the valve core zero position and will immediately send an alarm signal to the central control center.

[0049] like Figure 2 As shown in the figure, this diagram visually compares the processing effects of the present invention and the prior art under noisy conditions. The curve corresponding to the prior art represents the traditional fixed-step LMD algorithm. Due to the fixed parameters, it exhibits significant spurious oscillations (i.e., mode mixing) in the voltage stabilization region and shows obvious response lag when leakage occurs. The curve corresponding to the present invention automatically increases the step size in the noisy region to suppress oscillations and decreases the step size in the leakage region to accurately track the signal, effectively solving the technical problem of the traditional method's inability to simultaneously achieve noise reduction and fidelity preservation.

[0050] This invention also discloses a leakage monitoring system for a hydraulic system, including a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement a leakage monitoring method for a hydraulic system according to the present invention.

[0051] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0052] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A method for monitoring leakage in a hydraulic system, characterized in that, The method includes: Acquire pressure data of the valve core in the zero-position state of the hydraulic servo system; construct a local analysis window centered on each pressure data point, and perform local mean decomposition processing on the pressure data to obtain residual components. The local mean decomposition processing includes using an adaptive sliding step size for the data points; the calculation method of the adaptive sliding step size includes: calculating the effective noise dominance factor of each data point; the effective noise dominance factor is positively correlated with the local fluctuation density; the local fluctuation density is positively correlated with the number of sign changes of the first-order difference and the sum of squares of the second-order difference within the local analysis window; calculate the average descent rate of the residual components; and generate an alarm signal in response to the average descent rate exceeding a set threshold. The expression for the effective noise dominance factor is: In the formula, This represents the effective noise dominance factor at the i-th pressure data point. This represents the local fluctuation density at the i-th pressure data point. , These represent the values ​​of the j-th and (j+1)-th pressure data points, respectively. The hyperparameters are defined as follows: exp() represents an exponential function with base e, and N represents half the length of the local analysis window. The expression for local fluctuation density is: ； In the formula, This represents the value of the (j-1)th pressure data point. This is an indicator function that returns 1 if the condition within the parentheses is true, and 0 otherwise. represents the second-order difference value at the j-th pressure data point, and norm represents the normalization function; The expression for the adaptive sliding step size is: In the formula, This represents the optimized adaptive sliding step size at the i-th pressure data point. This represents the minimum sliding step size preset by the LMD algorithm. This represents the maximum sliding step size preset by the LMD algorithm. This represents the rounding function.

2. The method for monitoring leakage in a hydraulic system according to claim 1, characterized in that, The minimum sliding step size ranges from [2, 5]; the maximum sliding step size ranges from [20, 30].

3. The method for monitoring leakage in a hydraulic system according to claim 1, characterized in that, The method for constructing a local analysis window is as follows: Using the corresponding data point as the center, extract N points before and after it to form a window of length [length missing]. The local analysis window.

4. The method for monitoring leakage in a hydraulic system according to claim 1, characterized in that, The threshold was set to 0.05 MPa / min.

5. The method for monitoring leakage in a hydraulic system according to claim 1, characterized in that, The method for acquiring pressure data is as follows: a pressure sensor is installed at the rodless chamber interface of the hydraulic cylinder of the hydraulic servo system; when the hydraulic servo system is in pressure holding condition, the pressure sensor is used to collect pressure signals, and the pressure signals are converted from analog to digital to obtain the pressure data.

6. The method for monitoring leakage in a hydraulic system according to claim 5, characterized in that, The frequency range for acquiring pressure signals using the pressure sensor is [1.5, 2.5] kHz.

7. A leakage monitoring system for a hydraulic system, characterized in that, include: A processor and a memory, the memory storing computer program instructions that, when executed by the processor, implement a method for monitoring leakage in a hydraulic system according to any one of claims 1-6.

Images