Aircraft braking system electro-hydraulic pressure servo valve health assessment method based on feature space distance

Abstract

The invention discloses an aircraft braking system electro-hydraulic pressure servo valve health assessment method based on a characteristic space distance, and belongs to the field of electro-hydraulic pressure servo valve assessment. The method comprises the following steps: extracting two degradation sensitive characteristics of maximum output pressure and dead zone current in a full life cycle current-pressure characteristic curve of a servo valve; after dimension difference is eliminated through Z-score standardization, two-dimensional feature vectors are constructed and mapped into feature space state points; an ideal health center point is determined by initial health state data, an Euclidean distance between a state point and the center point is calculated as a degradation distance, and then a fault threshold is set according to statistics of a failure sample or a health sample, so that quantitative judgment of the health state is realized. According to the method, a multi-physical degradation mechanism is fused, the one-sidedness of single feature evaluation is overcome, the sensitivity of early weak degradation and the early warning robustness are improved, the false alarm and missing alarm rate is reduced, and a reliable basis is provided for condition-based maintenance of a brake system.

Description

Technical Field

[0001] This invention belongs to the field of electro-hydraulic pressure servo valve evaluation, and particularly relates to a health evaluation method for electro-hydraulic pressure servo valves in aircraft braking systems based on characteristic spatial distance. Background Technology

[0002] The electro-hydraulic pressure servo valve in an aircraft braking system is a core actuator for brake pressure regulation, and its performance directly determines the efficiency of anti-skid braking and takeoff and landing safety. In recent years, with the promotion of aviation predictive health management (PHM) technology, the industry has attempted to use physical models, signal processing, or data-driven methods to monitor the condition of servo valves. These methods generally rely on comparing a single feature (such as average pressure or RMS current) with a fixed threshold, triggering maintenance alerts through out-of-limit alarms. Meanwhile, some studies have introduced neural networks or support vector machines to map raw monitoring data into health indicators, enabling the tracking of degradation trends. However, in engineering applications, these solutions still adhere to the "single feature + empirical threshold" framework, failing to incorporate multiple key parameters that best reflect valve core wear, jamming, and leakage mechanisms into a unified quantitative system. This results in assessment results that are sensitive to fluctuations in operating conditions, have poor interpretability, and are difficult to meet the needs of high-precision health assessment under long-term service conditions.

[0003] The current technical challenges are as follows: Servo valve degradation is a slow process involving multiple coupled mechanisms such as wear, leakage, and jamming. A single feature cannot cover all degradation directions, and the significant differences in the dimensions of different features mean that direct fusion amplifies the contribution of high-level signals, masking subtle but crucial degradation information. Furthermore, existing threshold settings rely on expert experience or single failure samples, lacking statistically robust thresholds, which easily leads to false alarms or missed alarms. Therefore, there is an urgent need for a health assessment method that can simultaneously consider multiple physical degradation characteristics, eliminate the influence of dimensions, and provide objective distance measurements to improve the accuracy of detecting long-term performance degradation and the reliability of early warnings. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a health assessment method for electro-hydraulic pressure servo valves in aircraft braking systems based on characteristic spatial distance, comprising:

[0005] Obtain the current-pressure characteristic curve of the servo valve throughout its entire life cycle;

[0006] The maximum output pressure value and the dead zone current value are extracted from the characteristic curve and denoted as the maximum pressure characteristic value and the dead zone current characteristic value, respectively.

[0007] The maximum pressure characteristic and the dead zone current characteristic are standardized to obtain the standardized maximum pressure and the standardized dead zone current, respectively.

[0008] The standardized maximum pressure and the standardized dead zone current are combined into a two-dimensional feature vector, and this two-dimensional feature vector is mapped to a state point in the feature space.

[0009] An ideal health center point is preset in the feature space. The ideal health center point is composed of the standardized maximum pressure and the standardized dead zone current under the initial health state.

[0010] Calculate the Euclidean distance between the state point and the ideal health center point as the degradation distance;

[0011] The degradation distance is compared with a preset fault threshold, and the health status of the servo valve is determined based on the comparison result.

[0012] Optionally, the extraction of the maximum pressure feature includes:

[0013] During each test cycle, the output pressure value of the servo valve under the rated control current is obtained;

[0014] The output pressure value is taken as the maximum pressure characteristic quantity.

[0015] Optionally, the extraction of the dead-zone current characteristic quantities includes:

[0016] Within each test cycle, obtain the minimum input current value at which the servo valve output pressure begins to rise;

[0017] The minimum input current value is used as the dead zone current characteristic quantity.

[0018] Optionally, the standardization process includes:

[0019] The mean and standard deviation of the maximum pressure characteristic sequence are calculated respectively to obtain the mean and standard deviation of the maximum pressure.

[0020] The mean and standard deviation of the dead zone current characteristic sequence are calculated respectively to obtain the dead zone current mean and dead zone current standard deviation;

[0021] Subtracting the mean of the maximum pressure from each maximum pressure characteristic and then dividing by the standard deviation of the maximum pressure yields the standardized maximum pressure.

[0022] The normalized dead zone current is obtained by subtracting the mean dead zone current from each dead zone current characteristic and then dividing by the standard deviation of the dead zone current.

[0023] Optionally, the construction of the two-dimensional feature vector includes:

[0024] The standardized maximum pressure is used as the first dimension component;

[0025] The standardized dead zone current is used as the second dimension component;

[0026] The first dimension component and the second dimension component are combined to form a two-dimensional feature vector.

[0027] Optionally, determining the ideal health center point includes:

[0028] The standardized maximum pressure and standardized dead zone current corresponding to multiple test cycles of the servo valve in its initial healthy state are selected.

[0029] The average value of the standardized maximum pressure is calculated as the first dimension component of the health center.

[0030] The average value of the standardized dead zone current is calculated as the second dimension component of the health center.

[0031] The first dimension component of the health center is combined with the second dimension component of the health center to form an ideal health center point.

[0032] Optionally, the calculation of the degradation distance includes:

[0033] Obtain the standardized maximum pressure and standardized dead zone current at the current state point;

[0034] Calculate the difference between the standardized maximum pressure at the current state point and the first dimension component of the ideal health center point to obtain the first difference;

[0035] Calculate the difference between the standardized dead zone current at the current state point and the second-dimensional component at the ideal health center point to obtain the second difference;

[0036] The square root of the sum of the squares of the first and second differences is used to obtain the degradation distance.

[0037] Optionally, the fault threshold setting includes:

[0038] Obtain the standardized maximum pressure and standardized dead zone current of multiple healthy servo valves in the initial state;

[0039] Calculate the Euclidean distance between each health state point and the ideal health center point to form a set of health distances;

[0040] Calculate the mean and standard deviation of the set of healthy distances to obtain the mean and standard deviation of the healthy distances;

[0041] The fault threshold is obtained by adding three times the standard deviation of the healthy distance to the mean healthy distance.

[0042] On the other hand, the present invention also provides an electronic device including a memory, a processor, and a computing program stored in the memory and executable on the processor, wherein the processor implements the method when executing the computing program.

[0043] On the other hand, the present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method.

[0044] Compared with the prior art, the present invention has the following advantages and technical effects:

[0045] This invention extracts two degradation features directly related to wear and jamming: "maximum output pressure" and "dead zone current." After Z-score standardization, a two-dimensional feature space is constructed. Each detection is mapped to a state point in this space, and the Euclidean distance between this point and the "ideal health center" is calculated as the degradation distance. Furthermore, a distance threshold is statistically set using failure or health samples to quantitatively determine the health status of the servo valve. The resulting technical effects are: in multi-mechanism coupled degradation scenarios, the evaluation results monotonically increase with small fluctuations, are sensitive to early, subtle degradation, and the threshold setting is statistically robust, significantly reducing false alarm and missed alarm rates, providing reliable data support for subsequent condition-based maintenance. Attached Figure Description

[0046] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0047] Figure 1 This is a characteristic curve evolution diagram of an embodiment of the present invention;

[0048] Figure 2 This is a graph showing the evolution of the maximum pressure characteristics in an embodiment of the present invention;

[0049] Figure 3 This is a diagram illustrating the evolution of dead zone characteristics in an embodiment of the present invention.

[0050] Figure 4 This is a diagram illustrating the evolution of the degradation index according to an embodiment of the present invention.

[0051] Figure 5 This is a health index degradation curve diagram according to an embodiment of the present invention;

[0052] Figure 6 This is a feature space degradation trajectory diagram according to an embodiment of the present invention;

[0053] Figure 7 This is a graph showing the change of the Euclidean distance D with the number of working cycles in an embodiment of the present invention.

[0054] Figure 8 This is a schematic diagram of the method flow according to an embodiment of the present invention. Detailed Implementation

[0055] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0056] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0057] Example 1

[0058] like Figure 8 As shown, this embodiment provides a health assessment method for electro-hydraulic pressure servo valves of aircraft braking systems based on feature spatial distance, including:

[0059] Obtain the current-pressure characteristic curve of the servo valve throughout its entire life cycle;

[0060] The maximum output pressure value and the dead zone current value are extracted from the characteristic curve and denoted as the maximum pressure characteristic value and the dead zone current characteristic value, respectively.

[0061] The maximum pressure characteristic and the dead zone current characteristic are standardized to obtain the standardized maximum pressure and the standardized dead zone current, respectively.

[0062] The standardized maximum pressure and the standardized dead zone current are combined into a two-dimensional feature vector, and this two-dimensional feature vector is mapped to a state point in the feature space.

[0063] An ideal health center point is preset in the feature space. The ideal health center point is composed of the standardized maximum pressure and the standardized dead zone current under the initial health state.

[0064] Calculate the Euclidean distance between the state point and the ideal health center point as the degradation distance;

[0065] The degradation distance is compared with a preset fault threshold, and the health status of the servo valve is determined based on the comparison result.

[0066] Specifically:

[0067] The feature space distance-based evaluation method offers an alternative perspective for quantifying the health status of servo valves. The core idea is to map each operating state of the servo valve to a geometric point in a multi-dimensional feature space. By calculating the distance between this point and a predefined "ideal health center point," the degree of deviation from the healthy state is objectively measured. This method is widely used in pattern recognition and anomaly detection. Its advantages lie in the fact that it eliminates the need for complex index fusion and weight allocation, resulting in a more objective evaluation process that comprehensively reflects the common deviation effects of multiple features.

[0068] We can construct a feature space with each performance degradation feature as the coordinate axis. In this study, it is a two-dimensional feature space with the maximum pressure Pmax as one axis and the dead zone current Id as the other axis. In this space, each feature pair extracted from each test, i.e., the health state vector vk=[Pmax,k,Id,k]T, corresponds to a unique coordinate point. This point comprehensively describes the health state of the servo valve in the k-th test cycle.

[0069] Wherein, the feature vector in the initial healthy state is:

[0070] ;

[0071] We can define this as the "ideal health center" in this space. Theoretically, the feature points corresponding to all healthy servo valves should cluster near this health center, forming a "health cluster." As the performance of the servo valve degrades due to factors such as wear and jamming, Pmax gradually decreases, Id gradually increases, and its corresponding health state vector vk will drift in this two-dimensional space, forming a degradation trajectory. This trajectory will gradually move away from the health center. Therefore, the geometric distance from this state point to the health center can serve as a very intuitive and quantifiable "degradation indicator" or "abnormal score." The smaller the distance, the healthier the state; the larger the distance, the more severe the degradation.

[0072] Before directly calculating the distance, a crucial step is to standardize the original feature data. This is necessary because of the differences in dimensions and magnitudes between different features. As described in Section 3.1, the numerical magnitude of Pmax is much larger than that of Id. If the Euclidean distance is calculated directly in the original feature space, the result will be almost entirely dominated by Pmax, which has a much larger range of numerical variation, while the contribution of changes in Id to the total distance will be negligible, which is clearly unreasonable.

[0073] To eliminate the influence of different units of measurement and ensure that each feature has equal weight in distance calculation, this study employs Z-score standardization. The purpose of Z-score standardization is to transform the original data distribution into a standard normal distribution with a mean of 0 and a standard deviation of 1. Its mathematical formula is:

[0074] ;

[0075] Where x is the original measurement of a feature, μ is the mean of that feature in the entire lifecycle dataset, and σ is its standard deviation. For a feature sequence X={x1,x2,...,xm} containing m samples, its mean and standard deviation are calculated as follows:

[0076] ;

[0077] ;

[0078] This transformation is applied to the P_{max} and I_d data sequences respectively, resulting in two new standardized feature sequences, P_{std} and I_{std}. After standardization, both features are transformed to the same dimensionless scale, making them comparable and laying the foundation for subsequent distance calculations.

[0079] In the standardized feature space, we can accurately calculate the distance from each state point to the health center.

[0080] Define standardized health centers: First, standardize the health benchmarks Phealthy and Ihealthy using the Z-score formula mentioned above to obtain the standardized "health center points" Pstd, healthy, Istd,healthy.

[0081] Calculate the Euclidean distance: For each standardized feature point (P_{std, current}, I_{std, current}) obtained from each test, calculate its Euclidean distance to the standardized healthy center. Degenerate distance D:

[0082] ;

[0083] The degradation distance D is a non-negative scalar, and its value itself can serve as a quantitative health indicator. Unlike HI, distance D has no fixed upper bound, and its absolute value's physical meaning is not as intuitive as HI. However, its trend is very clear: throughout the servo valve's lifespan, as performance degrades, the value of D exhibits a clear, monotonically increasing trend. It is worth noting that besides Euclidean distance, other distance metrics exist, such as Mahalanobis distance, which can consider the correlation between features. However, in this study, since the physical degradation mechanisms of P_{max}$ and $I_d are relatively independent, and Z-score normalization has decoupled the data to some extent, using the more intuitive and computationally simple Euclidean distance is reasonable and effective.

[0084] To enable the degradation distance D to be used for practical fault early warning, a reasonable fault threshold Dthreshold needs to be set for it. When the detected distance D exceeds this threshold, the system can issue an alarm, indicating that the servo valve has entered an unhealthy state. There are generally two methods for setting the threshold:

[0085] Failure-based determination method: Using our "run-to-failure" experimental data, the feature points of the last test (i.e., the failure state) are standardized, and the corresponding degradation distance Dfailure is calculated. This distance value, or the value multiplied by a safety factor (e.g., 0.9), can be used as the final failure threshold. This method directly utilizes prior knowledge from physical experiments, and the defined threshold has a clear physical meaning.

[0086] Statistical approach based on healthy samples: If initial test data from multiple healthy servo valves is available, the distance distribution from all healthy sample points to the "ideal health center" can be calculated. Assuming this distribution approximates a certain statistical distribution (such as a normal distribution), a threshold can be set based on statistical principles. For example, the classic 3-sigma principle can be used, setting the threshold at the "mean + 3 standard deviations" of the healthy distance distribution. Any point exceeding this threshold can be considered a low-probability event, indicating significant performance degradation. This method is particularly useful when explicit failure data is lacking.

[0087] On the other hand, this embodiment also provides an electronic device, including a memory, a processor, and a computing program stored in the memory and executable on the processor, wherein the processor implements the method when executing the computing program.

[0088] On the other hand, this embodiment also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method.

[0089] Example 2

[0090] This embodiment provides a health assessment method for electro-hydraulic pressure servo valves in aircraft braking systems based on feature spatial distance, including:

[0091] To verify the effectiveness and feasibility of the constructed health assessment theoretical framework, this study conducted a full life-cycle performance degradation experiment of servo valves on a specially built hydraulic servo system test platform. This chapter aims to comprehensively verify and analyze the proposed algorithm based on the collected measured data. The entire analysis process will start with the most intuitive physical phenomena, gradually delve into the quantitative characteristics and evolution laws, and finally present and compare the evaluation results of the two health assessment models, and conduct preliminary numerical verification of the effectiveness of the online application strategy.

[0092] The performance degradation experiments conducted in this study were performed on a hydraulic testing platform capable of accurately simulating the actual working load and conditions of a servo valve. The core test object was an electro-hydraulic servo valve with a rated working pressure of 8 MPa and a rated control current of 8 mA. To accurately capture its performance changes, high-precision pressure sensors were installed at the critical ports of the servo valve, along with corresponding current sensors and a data acquisition system. To obtain complete data on the servo valve from its healthy state to complete failure, we designed and executed an accelerated aging experiment. During the intervals between two performance tests, the servo valve underwent continuous high-frequency operating cycles at 10 Hz to simulate and accelerate its wear and fatigue during long-term service. The entire accelerated aging experiment accumulated over one million (1.188 × 10⁶) operating cycles, during which a series of performance snapshot datasets from the initial healthy state to the near-failure state were periodically acquired, providing a precise metric for subsequent correlation analysis between performance degradation and actual workload.

[0093] The degradation of servo valve performance is most directly reflected in the evolution of its current-pressure characteristic curve. Figure 1This process is clearly illustrated in the figure, which selects three representative snapshots from the early, middle, and late stages of the experiment. It is evident from the figure that the characteristic curve exhibits significant and physically consistent degradation as the number of working cycles accumulates. First, the overall "peak" of the curve continuously decreases, intuitively indicating a decline in its maximum pressure output capability at rated current. Second, the "starting point" of the curve on the X-axis continuously shifts to the right, meaning a larger initial current is required for the valve to produce an effective response, i.e., its dead zone is increasing. Furthermore, it can be observed that the slope of the linear segment of the late-stage curve appears slightly lower than that of the early stage, potentially indicating that the valve's pressure gain is also affected. This series of visualized physical phenomena strongly demonstrates that using maximum pressure (Pmax) and dead zone current (Id) as core degradation characteristics is reasonable and effective.

[0094] To quantify the above evolution process, we extracted two key features, Pmax and Id, from the series of data collected throughout the experiment. Figure 2 The graph shows the complete trajectory of the maximum pressure Pmax as a function of the number of operating cycles. As can be seen from the graph, the degradation of Pmax is not a simple linear process. In the initial stage, the pressure even shows a slight increase, reaching the peak performance level for the entire lifespan (approximately 8.1 MPa). This may correspond to the "break-in period" of a new valve, where internal components reach optimal coordination after initial operation. Subsequently, the pressure enters a long-term, fluctuating downward trend, with a clear overall downward trajectory, eventually decaying to around 7.3 MPa. Figure 3 This illustrates the trajectory of the dead zone current Id as a function of the number of operating cycles. The overall trend is a significant increase, fluctuating from approximately 0.9 mA initially to nearly 1.4 mA at the end of the experiment. Notably, both curves exhibit significant local fluctuations, rather than strictly monotonic changes. This phenomenon realistically reflects the complexity of the physical degradation process; for example, the random movement of internal contaminant particles may cause a sudden increase or decrease in the frictional force of the valve core at a certain moment, resulting in a jump in Id. The macroscopic evolution trends of these two characteristics are perfectly consistent with our theoretical analysis in Chapter 2, providing a solid data foundation for the subsequent construction of a health assessment model.

[0095] The core idea of ​​distance-based evaluation is to abstract the health state of a servo valve as a geometric point in a multi-dimensional feature space. To intuitively verify this idea, we plotted the feature pairs extracted throughout the experiment on a graph such as... Figure 6 In the two-dimensional feature space shown.

[0096] This graph provides a completely new geometric perspective for examining performance degradation. Each point in the graph represents a "snapshot" of the servo valve's health status at a specific moment. We first calculated an "initial health center" (shown as a green pentagram in the graph) based on the data from the initial health window, which represents the servo valve's ideal, unworn performance state. The color bars in the graph represent the evolution over time, gradually transitioning from cool tones (blue) at the beginning of the experiment to warmer tones (yellow / red) at the end.

[0097] The graph clearly shows a degradation trajectory that conforms to the laws of physical degradation. In the initial stage of the experiment, the blue data points representing the healthy state are closely clustered in the upper left of the "initial health center," forming a "health cluster." This region is characterized by high maximum pressure and low dead zone current, which perfectly matches the high performance expected of a healthy servo valve. As the number of working cycles accumulates, the color of the data points gradually warms up, and their positions begin to drift systematically. The overall trend of the trajectory is a gradual shift from the "health zone" in the upper left to the "failure zone" in the lower right. This drift direction has a clear physical meaning: its projection on the Y-axis continuously decreases, corresponding to a continuous decrease in maximum pressure; its projection on the X-axis continuously increases, corresponding to a continuous increase in dead zone current. This graph not only intuitively verifies that the two features we selected can effectively capture the main direction of performance degradation, but more importantly, it materializes the abstract "performance degradation" process into a clearly visible and measurable geometric path in the feature space.

[0098] After visually confirming through feature space graphs that state points move away from the health center over time, the next key step is to accurately quantify this process. Figure 7 The complete curve shows the Euclidean distance D from each state point to the "initial health center" as a function of the number of work cycles.

[0099] This graph successfully reduces the two-dimensional, complex degradation trajectory into a single, quantifiable degradation indicator. As can be observed from the graph, although the curve exhibits significant local fluctuations, its global, macroscopic trend is clearly upward. In the initial stage of the experiment, the degradation distance D remains at a low level of around 0.5, indicating that the servo valve performance is stable at this point, with the state points closely surrounding the healthy center. As the number of work cycles increases, the baseline of the distance value begins to rise slowly, experiencing several sharp peaks throughout the experiment. For example, significant peaks in distance D occur around approximately 2 × 10⁵ and 9.5 × 10⁵ cycles, indicating that the servo valve's performance state at these moments momentarily and significantly deviates from its initial healthy state. This dramatic fluctuation precisely reflects an important characteristic of the distance model: high sensitivity. Compared to the normalized and weighted average HI model, distance D is a more "raw" and direct metric, faithfully reflecting the true fluctuations of Pmax and Id in each test. These fluctuations may correspond to real physical processes, such as the random movement of internal contaminant particles causing a sudden increase in the friction of the valve core at a certain moment, which in turn causes a jump in Id and a surge in the distance value.

[0100] Ultimately, towards the end of the experiment, the central fluctuation of the degradation distance D stabilized above 2.5 and reached a maximum value exceeding 3.0, representing a several-fold increase compared to the initial state. This clearly indicates that the servo valve's performance has undergone irreversible and significant degradation. In summary, the evolution trend of the degradation distance D successfully and quantitatively tracked the entire health deterioration process of the servo valve, proving that the assessment method based on characteristic space distance is an effective and sensitive means of health status assessment.

[0101] The above are merely preferred embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A health assessment method for electro-hydraulic pressure servo valves in aircraft braking systems based on characteristic spatial distance, characterized in that, include: Obtain the current-pressure characteristic curve of the servo valve throughout its entire life cycle; The maximum output pressure value and the dead zone current value are extracted from the characteristic curve and denoted as the maximum pressure characteristic value and the dead zone current characteristic value, respectively. The maximum pressure characteristic and the dead zone current characteristic are standardized to obtain the standardized maximum pressure and the standardized dead zone current, respectively. The standardized maximum pressure and the standardized dead zone current are combined into a two-dimensional feature vector, and this two-dimensional feature vector is mapped to a state point in the feature space. An ideal health center point is preset in the feature space. The ideal health center point is composed of the standardized maximum pressure and the standardized dead zone current under the initial health state. Calculate the Euclidean distance between the state point and the ideal health center point as the degradation distance; The degradation distance is compared with a preset fault threshold, and the health status of the servo valve is determined based on the comparison result.

2. The method according to claim 1, characterized in that, The extraction of the maximum pressure feature includes: During each test cycle, the output pressure value of the servo valve under the rated control current is obtained; The output pressure value is taken as the maximum pressure characteristic quantity.

3. The method according to claim 1, characterized in that, The extraction of dead-zone current characteristics includes: Within each test cycle, obtain the minimum input current value at which the servo valve output pressure begins to rise; The minimum input current value is used as the dead zone current characteristic quantity.

4. The method according to claim 1, characterized in that, The standardization process includes: The mean and standard deviation of the maximum pressure characteristic sequence are calculated respectively to obtain the mean and standard deviation of the maximum pressure. The mean and standard deviation of the dead zone current characteristic sequence are calculated respectively to obtain the dead zone current mean and dead zone current standard deviation; Subtracting the mean of the maximum pressure from each maximum pressure characteristic and then dividing by the standard deviation of the maximum pressure yields the standardized maximum pressure. The normalized dead zone current is obtained by subtracting the mean dead zone current from each dead zone current characteristic and then dividing by the standard deviation of the dead zone current.

5. The method according to claim 1, characterized in that, The construction of the two-dimensional feature vector includes: The standardized maximum pressure is used as the first dimension component; The standardized dead zone current is used as the second dimension component; The first dimension component and the second dimension component are combined to form a two-dimensional feature vector.

6. The method according to claim 1, characterized in that, The determination of the ideal health center point includes: The standardized maximum pressure and standardized dead zone current corresponding to multiple test cycles of the servo valve in its initial healthy state are selected. The average value of the standardized maximum pressure is calculated as the first dimension component of the health center. The average value of the standardized dead zone current is calculated as the second dimension component of the health center. The first dimension component of the health center is combined with the second dimension component of the health center to form an ideal health center point.

7. The method according to claim 1, characterized in that, The calculation of the degradation distance includes: Obtain the standardized maximum pressure and standardized dead zone current at the current state point; Calculate the difference between the standardized maximum pressure at the current state point and the first dimension component of the ideal health center point to obtain the first difference; Calculate the difference between the standardized dead zone current at the current state point and the second-dimensional component at the ideal health center point to obtain the second difference; The square root of the sum of the squares of the first and second differences is used to obtain the degradation distance.

8. The method according to claim 1, characterized in that, The setting of fault thresholds includes: Obtain the standardized maximum pressure and standardized dead zone current of multiple healthy servo valves in the initial state; Calculate the Euclidean distance between each health state point and the ideal health center point to form a set of health distances; Calculate the mean and standard deviation of the set of healthy distances to obtain the mean and standard deviation of the healthy distances; The fault threshold is obtained by adding three times the standard deviation of the healthy distance to the mean healthy distance.

9. An electronic device comprising a memory, a processor, and a computing program stored in the memory and executable on the processor, characterized in that, When the processor executes the computing program, it implements the method of any one of claims 1-8.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1-8.

Images