Hydraulic brake system oil pipe leakage fault detection method, device, equipment and storage medium

Abstract

This application discloses a method, apparatus, equipment, and storage medium for detecting oil pipe leakage faults in a hydraulic braking system, relating to the field of engineering machinery technology. The method includes: constructing a first dynamic mathematical model of key components in the hydraulic braking system of a loader, the key components including an accumulator and an electro-hydraulic proportional reverse brake valve; constructing a second dynamic mathematical model of the pressure relief valve in the hydraulic braking system; importing the first and second dynamic mathematical models into preset simulation software to obtain an oil pipe leakage fault simulation model, and generating normal and abnormal data using the oil pipe leakage fault simulation model; and detecting oil pipe leakage faults based on the normal and abnormal data using an adaptive self-organizing mapping network. This application can accurately detect oil pipe leakage faults in the hydraulic braking system of a loader even when there is a lack of labeled fault data, inaccurate brake pressure control, and no theoretical basis for threshold selection.

Description

Technical Field

[0001] This application relates to the field of engineering machinery technology, and in particular to methods, devices, equipment and storage media for detecting oil pipe leakage faults in hydraulic braking systems. Background Technology

[0002] In underground mining operations, the WJ-6 internal combustion loader, as a core trackless transport equipment, relies heavily on the reliability of its hydraulic braking system, which directly impacts personnel safety and production efficiency. The hydraulic braking system transmits braking force through high-pressure hydraulic fluid. Leaks in the hydraulic lines, especially internal leaks (where hydraulic oil flows abnormally from the high-pressure chamber to the low-pressure chamber), can lead to decreased braking pressure, delayed response, and even brake failure. Under complex operating conditions of high load, high dust, and high humidity, problems such as aging hydraulic lines, loose joints, or seal failures frequently occur, significantly increasing the risk of leakage.

[0003] Current technologies for detecting leaks in hydraulic systems mainly include manual inspection, flow monitoring, and data-driven anomaly detection methods. Manual inspection relies on operators periodically observing for external leaks such as oil stains or a drop in oil level. Flow monitoring methods use high-precision flow sensors installed in pipelines to compare the theoretical flow rate with the actual flow rate to determine if a leak has occurred. In addition, some studies have attempted to introduce machine learning methods, such as using supervised learning models to classify and train labeled normal and faulty samples to identify abnormal states.

[0004] However, existing methods face multiple technical bottlenecks in practical applications. Manual inspections cannot detect internal leaks and lack continuity and real-time capability; flow sensor solutions are costly, complex to install, and have limited sensitivity to minute leaks; supervised learning methods heavily rely on large amounts of high-quality labeled fault data, while leak fault samples are scarce in real industrial scenarios, and actively injecting faults poses safety risks and equipment damage costs; unsupervised methods, while not requiring labels, typically rely on empirically set thresholds for anomaly detection, lacking theoretical basis and prone to false alarms or missed alarms. Furthermore, most existing systems still use mechanical foot pedals as braking inputs, whose output pressure is greatly influenced by operator habits, making it difficult to establish stable and reproducible normal operating condition benchmarks, further restricting the accuracy and robustness of fault detection. Therefore, accurately detecting oil pipe leaks in the hydraulic braking system of loaders in the absence of labeled fault data, inaccurate braking pressure control, and a lack of theoretical basis for threshold selection has become an urgent problem to be solved. Summary of the Invention

[0005] The purpose of this application is to provide a method, device, equipment and storage medium for detecting oil pipe leakage faults in hydraulic braking systems, aiming to solve the technical problem of accurately detecting oil pipe leakage faults in the hydraulic braking systems of loaders when there is a lack of labeled fault data, inaccurate braking pressure control and no theoretical basis for threshold selection.

[0006] To achieve the above objectives, this application proposes a method for detecting oil pipe leakage faults in a hydraulic braking system, the method comprising:

[0007] Calculate the target pressure of the accumulator in the hydraulic braking system, and filter and limit the target pressure to obtain the limited accumulator pressure;

[0008] Based on the valve core opening degree of the electro-hydraulic proportional reverse brake valve in the hydraulic braking system and the preset polynomial fitting coefficients, a mapping relationship between the valve core opening degree and the braking pressure is established to obtain the theoretical braking pressure of the electro-hydraulic proportional reverse brake valve.

[0009] Based on the accumulator pressure after the amplitude limit, the preset nonlinear scaling index, and the theoretical braking pressure, a mathematical function of the actual output braking pressure of the electro-hydraulic proportional reverse braking valve is constructed to obtain the first dynamic mathematical model.

[0010] A second dynamic mathematical model of the pressure relief valve is constructed based on the leakage and abnormal pressure drop of the pressure relief valve in the hydraulic braking system.

[0011] The first dynamic mathematical model and the second dynamic mathematical model are imported into a preset simulation software to obtain a pipeline leakage fault simulation model, and normal data and abnormal data are generated using the pipeline leakage fault simulation model.

[0012] Based on the normal data and the abnormal data, an adaptive self-organizing mapping network is used to detect pipeline leaks.

[0013] In one embodiment, the step of calculating the target pressure of the accumulator in the hydraulic braking system and filtering and limiting the target pressure to obtain the limited accumulator pressure includes:

[0014] Obtain the pressure of the accumulator at the previous moment, and calculate the target pressure of the accumulator using the explicit Euler method based on the previous pressure, the preset pressure compensation pressure, and the preset pressure loss.

[0015] The target pressure is processed using a first-order low-pass filter to obtain the actual output pressure of the energy storage device.

[0016] The actual output pressure of the accumulator is limited according to the preset upper pressure limit and the preset lower pressure limit to obtain the limited accumulator pressure.

[0017] The first dynamic mathematical model is represented as follows:

[0018]

[0019] in, This refers to the actual output braking pressure of the electro-hydraulic proportional reverse brake valve. This refers to the theoretical braking pressure. This refers to the accumulator pressure after the amplitude is limited. This refers to the preset pressure upper limit. This refers to the preset nonlinear scaling index.

[0020] In one embodiment, the step of constructing a second dynamic mathematical model of the pressure relief valve based on the leakage and abnormal pressure drop of the pressure relief valve in the hydraulic braking system includes:

[0021] The leakage of the pressure relief valve in the hydraulic braking system is calculated based on the preset maximum pressure relief capacity and the preset pressure relief growth time.

[0022] The initial opening of the pressure relief valve is subjected to dead zone processing based on a preset dead zone threshold to obtain the effective opening.

[0023] The effective opening is normalized to obtain the normalized opening.

[0024] The abnormal pressure drop is calculated based on the normalized opening degree and the preset fitting index;

[0025] Based on the leakage amount, the abnormal pressure drop, the effective opening degree, and the preset dead zone threshold, a second dynamic mathematical model of the pressure relief valve is constructed.

[0026] The second dynamic mathematical model is expressed as follows:

[0027]

[0028] in, This refers to the abnormal pressure drop. This refers to the effective opening degree. This refers to the amount of leakage. This refers to the preset dead zone threshold.

[0029] In one embodiment, the step of detecting pipeline leaks using an adaptive self-organizing mapping network based on the normal data and the abnormal data includes:

[0030] The normal data and the abnormal data are integrated and divided into training set, validation set and test set according to a preset ratio;

[0031] An adaptive self-organizing mapping network is constructed based on a preset neighborhood radius, a preset learning rate, and preset grid parameters. The input features of the adaptive self-organizing mapping network are the valve core opening of the electro-hydraulic proportional reverse braking valve and the system braking pressure.

[0032] The adaptive self-organizing map network is iteratively trained based on the training set to determine the best matching unit;

[0033] The optimal matching unit and the neuron weights in the neighborhood of the optimal matching unit are updated by a Gaussian neighborhood function, while the preset learning rate and the preset neighborhood radius decay exponentially to obtain the trained adaptive self-organizing map network.

[0034] The detection results are obtained by performing pipeline leak fault detection based on the validation set, the test set, and the trained adaptive self-organizing map network.

[0035] In one embodiment, the step of detecting pipeline leaks based on the validation set, the test set, and the trained adaptive self-organizing map network to obtain detection results includes:

[0036] The validation set is input into the trained adaptive self-organizing map network, and the first quantization error of each validation sample to the corresponding best matching unit is calculated.

[0037] Based on the first quantization error and the true label of the verification sample, the anomaly detection threshold is determined using the Youden index threshold optimization method.

[0038] Input the test set into the trained adaptive self-organizing map network and calculate the second quantization error for each test sample;

[0039] When the second quantization error is less than the anomaly detection threshold, the detection result is that there is no oil pipe leakage fault;

[0040] When the second quantization error is greater than or equal to the anomaly detection threshold, the detection result is that there is an oil pipe leak.

[0041] In one embodiment, the step of importing the first dynamic mathematical model and the second dynamic mathematical model into preset simulation software to obtain a pipeline leakage fault simulation model includes:

[0042] Import the first dynamic mathematical model and the second dynamic mathematical model into the preset simulation software;

[0043] Based on the variable transfer relationship between the actual output braking pressure in the first dynamic mathematical model and the abnormal pressure drop in the second dynamic mathematical model, an initial joint simulation model of the target output braking pressure is constructed.

[0044] The initial joint simulation model is simulated in the preset simulation software to obtain preliminary simulation results;

[0045] Obtain real experimental test data, and calculate the fitting accuracy between the real experimental test data and the preliminary simulation results based on the Pearson correlation coefficient and the coefficient of determination;

[0046] When the fitting accuracy is less than a preset accuracy threshold, the internal parameters of the initial co-simulation model are adjusted, and the process returns to the step of simulating the initial co-simulation model in the preset simulation software to obtain preliminary simulation results.

[0047] When the fitting accuracy is greater than or equal to the preset accuracy threshold, the initial joint simulation model is used as the simulation model for oil pipe leakage fault.

[0048] In one embodiment, the step of generating normal and abnormal data using the pipeline leak fault simulation model includes:

[0049] In the oil pipe leakage fault simulation model, the pressure relief valve opening of the second dynamic mathematical model is set to the first preset opening value, and the oil pipe leakage fault simulation model is controlled to run according to the valve core opening of multiple different electro-hydraulic proportional reverse brake valves to obtain normal data.

[0050] The pressure relief valve opening of the second dynamic mathematical model is adjusted to multiple different second preset opening values, and the oil pipe leakage fault simulation model is controlled to run according to the input of multiple valve core opening values ​​to obtain abnormal data.

[0051] Furthermore, to achieve the above objectives, this application also proposes a hydraulic braking system oil pipe leakage fault detection device, the device comprising:

[0052] An accumulator pressure processing module is used to calculate the target pressure of the accumulator in the hydraulic braking system, and to filter and limit the target pressure to obtain the limited accumulator pressure.

[0053] The proportional valve theoretical pressure module is used to establish a mapping relationship between the valve core opening and the braking pressure based on the valve core opening of the electro-hydraulic proportional reverse brake valve in the hydraulic braking system and a preset polynomial fitting coefficient, so as to obtain the theoretical braking pressure of the electro-hydraulic proportional reverse brake valve.

[0054] The first dynamic model construction module is used to construct a mathematical function of the actual output braking pressure of the electro-hydraulic proportional reverse braking valve based on the accumulator pressure after the limit, the preset nonlinear scaling index and the theoretical braking pressure, so as to obtain the first dynamic mathematical model.

[0055] The second dynamic model construction module is used to construct a second dynamic mathematical model of the pressure relief valve based on the leakage and abnormal pressure drop of the pressure relief valve in the hydraulic braking system.

[0056] The simulation module is used to import the first dynamic mathematical model and the second dynamic mathematical model into a preset simulation software to obtain a pipeline leakage fault simulation model, and to generate normal data and abnormal data using the pipeline leakage fault simulation model.

[0057] The fault detection module is used to detect oil pipe leaks based on the normal data and the abnormal data through an adaptive self-organizing mapping network.

[0058] In addition, to achieve the above objectives, this application also proposes a hydraulic braking system oil pipe leakage fault detection device, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the hydraulic braking system oil pipe leakage fault detection method as described above.

[0059] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the hydraulic braking system oil pipe leakage fault detection method as described above.

[0060] In addition, to achieve the above objectives, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the hydraulic braking system oil pipe leakage fault detection method described above.

[0061] One or more technical solutions proposed in this application have at least the following technical effects:

[0062] First, the target pressure of the accumulator is calculated, and the inertial delay of its actual response is simulated by a first-order low-pass filter. Then, a limiting process is performed using preset upper and lower pressure limits to obtain the limited accumulator pressure, making the model closer to the dynamic characteristics of the real system. Second, a fifth-order polynomial mapping relationship is established based on the valve core opening of the electro-hydraulic proportional reverse braking valve and preset polynomial fitting coefficients to obtain the theoretical braking pressure, accurately characterizing the nonlinear relationship between valve control input and output pressure. Next, the actual output braking pressure function of the electro-hydraulic proportional reverse braking valve is constructed by combining the limited accumulator pressure, the preset nonlinear scaling exponent, and the theoretical braking pressure, forming the first dynamic... A dynamic mathematical model is first constructed to effectively reflect the impact of pressure source limitations on braking output. Then, a second dynamic mathematical model is built based on the leakage amount and abnormal pressure drop of the pressure relief valve to simulate oil pipe leakage faults of different degrees. Subsequently, the two models are integrated into Simulink simulation software to obtain an oil pipe leakage fault simulation model. This model is then used to generate a dataset covering normal and various abnormal operating conditions, providing high-fidelity samples for algorithm training. Finally, based on the generated normal and abnormal data, an adaptive self-organizing map network is used for unsupervised learning, and the anomaly judgment threshold is optimized in combination with the validation set to achieve automatic detection of oil pipe leakage faults. This application can accurately detect oil pipe leakage faults in the hydraulic braking system of a loader in the absence of labeled fault data, inaccurate braking pressure control, and lack of theoretical basis for threshold selection. Attached Figure Description

[0063] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0064] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0065] Figure 1 This is a flowchart illustrating an embodiment of the hydraulic braking system oil pipe leakage fault detection method of this application.

[0066] Figure 2 This is a flowchart illustrating Embodiment 2 of the hydraulic braking system oil pipe leakage fault detection method of this application;

[0067] Figure 3 This is a schematic diagram of a simulation model of an oil pipe leakage fault provided in Embodiment 2 of the hydraulic braking system oil pipe leakage fault detection method of this application.

[0068] Figure 4This is a schematic diagram of the brake pressure change curve provided in Embodiment 2 of the hydraulic braking system oil pipe leakage fault detection method of this application;

[0069] Figure 5 A schematic diagram of accumulator pressure change provided in Embodiment 2 of the hydraulic braking system oil pipe leakage fault detection method of this application;

[0070] Figure 6 This is a schematic diagram of the simulation model generating data provided in Embodiment 2 of the hydraulic braking system oil pipe leakage fault detection method of this application;

[0071] Figure 7 A simplified flowchart illustrating the oil pipe leakage fault detection method for the hydraulic braking system provided in Embodiment 2 of this application;

[0072] Figure 8 This is a schematic diagram of the module structure of the oil pipe leakage fault detection device for the hydraulic braking system according to an embodiment of this application;

[0073] Figure 9 This is a schematic diagram of the equipment structure of the hardware operating environment involved in the oil pipe leakage fault detection method of the hydraulic braking system in the embodiments of this application.

[0074] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0075] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0076] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0077] It should be noted that the executing entity of this application embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or an electronic device or fault detection system capable of realizing the above functions. The following uses a fault detection system as an example to describe this embodiment and the subsequent embodiments.

[0078] Based on this, embodiments of this application provide a method for detecting oil pipe leakage faults in a hydraulic braking system, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the hydraulic braking system oil pipe leakage fault detection method of this application.

[0079] In this embodiment, the method for detecting oil pipe leakage faults in the hydraulic braking system includes steps S10 to S60:

[0080] Step S10: Calculate the target pressure of the accumulator in the hydraulic braking system, and filter and limit the target pressure to obtain the limited accumulator pressure.

[0081] It should be noted that a loader is a self-propelled trackless engineering machine used for loading and transporting materials in underground or open-pit mines. A typical example is the WJ-6 internal combustion loader, which is driven by a diesel engine and equipped with a hydraulic system to achieve its operating functions. The hydraulic braking system refers to the braking device assembly in the loader that uses hydraulic oil as the working medium and transmits pressure to achieve vehicle deceleration or stopping. It mainly includes components such as accumulators, brake valves, oil pipes, and brakes, and its performance directly affects driving safety. An accumulator is a pressure vessel in the hydraulic braking system used to store and release hydraulic energy. It is usually composed of an air bladder or spring and a hydraulic chamber. When the system pressure is insufficient, it releases pre-stored high-pressure oil to maintain stable braking pressure. The target pressure refers to the idealized accumulator pressure value calculated using the explicit Euler method based on the previous pressure, the replenishment pressure, and the pressure loss. It represents the pressure level that the system should achieve under conditions of no dynamic delay.

[0082] As an example, the step of calculating the target pressure of the accumulator in the hydraulic braking system and filtering and limiting the target pressure to obtain the limited accumulator pressure includes: obtaining the pressure of the accumulator at the previous moment; calculating the target pressure of the accumulator using the explicit Euler method based on the previous moment pressure, a preset pressure compensation pressure, and a preset pressure loss; processing the target pressure using a first-order low-pass filter to obtain the actual output pressure of the accumulator; and limiting the actual output pressure of the accumulator according to a preset upper pressure limit and a preset lower pressure limit to obtain the limited accumulator pressure.

[0083] It should be noted that the pressure at the previous moment refers to the hydraulic pressure value inside the accumulator recorded in the sampling period prior to the current simulation time step, used as the initial condition for the dynamic model to calculate the current state. The preset replenishment pressure represents the pressurization capacity that the hydraulic pump can provide when replenishing oil to the accumulator. The preset pressure loss includes the sum of pressure drops caused by valves, pipelines, etc., during normal system operation (such as the normal pressure drop of an electro-hydraulic proportional reverse brake valve) and additional pressure drops caused by leakage under abnormal operating conditions. The first-order low-pass filter is a mathematical model that simulates the inertial response characteristics of a physical system. Its output has a smoothing hysteresis effect on input changes, used here to reflect the dynamic delay behavior of the accumulator's actual pressure not being able to instantaneously follow changes in the target pressure. The actual output pressure refers to the accumulator pressure value after processing by the first-order low-pass filter, more closely resembling the continuous and smooth pressure change process presented in a real hydraulic system due to fluid inertia and component response limitations. The preset pressure upper limit and preset pressure lower limit refer to the boundary values ​​of the allowable range of accumulator pressure set to ensure system safety and normal operation, corresponding to the maximum working pressure the system can withstand and the minimum pressure required to maintain effective braking, respectively.

[0084] Understandably, this involves accumulator modeling:

[0085] The accumulator serves as the power source for the entire hydraulic braking system, replenishing pressure by filling with fluid when the system pressure is insufficient. The target pressure of the accumulator is calculated using the explicit Euler method with the following formula. This includes the pressure of the accumulator at the previous moment. and the current time step Internal pressure changes. The simpler, explicit Euler method is used for calculation, which is computationally less complex and intuitive, making it more suitable for real-time simulation. Pressure is replenished by filling the hydraulic pump. (Preset pressure compensation) and pressure loss The net inflow is calculated based on the difference between the preset pressure loss and the actual pressure, reflecting the change in pressure and having a clear physical meaning; and the effective liquid capacity of the accumulator is also introduced. ( This relates pressure changes to volume changes, which aligns with the basic principles of hydraulic systems (the larger the liquid volume, the smaller the pressure change caused by the same net inflow).

[0086]

[0087]

[0088] in, Preset pressure, unit MPa; The pressure replenishment rate is expressed in MPa / s; when the accumulator pressure is lower than the preset lower pressure limit... This means that pressure is replenished. The preset pressure loss is expressed in MPa, including normal pressure drop (the decrease in braking pressure caused by different opening degrees of the electro-hydraulic proportional valve, reflecting internal valve losses, and representing the actual accumulator pressure). (Actual output pressure) and the actual output braking pressure of the electro-hydraulic proportional reverse brake valve The difference between them) and the abnormal pressure drop caused by the oil pipe leak. (Additional pressure loss due to pipeline leakage) Identifying the source of the loss helps distinguish the characteristics and impacts of both. Pressure difference. The value implies a conversion factor for the dimension of flow (i.e., the flow rate corresponding to a unit pressure difference).

[0089] The target pressure will not be reached instantly. By utilizing the inertial delay characteristics of the actual pressure response of the accumulator as shown in the following formula, and employing a first-order low-pass filter, the pressure is ensured to smoothly approach the target value, preventing jumps and more closely resembling the behavior of the real system.

[0090]

[0091] Real accumulator pressure Limiting, Ensure that the value is not negative and does not exceed the preset pressure limit. ; is a time constant, measured in seconds (s), representing the response rate to changes in pressure.

[0092] Step S20: Based on the valve core opening degree of the electro-hydraulic proportional reverse brake valve in the hydraulic braking system and the preset polynomial fitting coefficient, establish the mapping relationship between the valve core opening degree and the braking pressure to obtain the theoretical braking pressure of the electro-hydraulic proportional reverse brake valve.

[0093] It should be noted that an electro-hydraulic proportional reverse brake valve is a control element that uses electrical signals to control the valve core opening, thereby proportionally adjusting the flow and pressure of hydraulic oil. Its "reverse" characteristic means that the brake pressure decreases as the control signal increases, while the actual brake pressure decreases. It can replace the traditional foot pedal to achieve digital and precise control of the braking process. Valve core opening refers to the positional offset of the movable valve core relative to the valve body, usually expressed as a percentage or normalized value. It directly determines the cross-sectional area of ​​the hydraulic oil passage, thus controlling the brake pressure output. Preset polynomial fitting coefficients refer to a set of fixed parameters obtained by fitting experimental data, used to construct a fifth-order polynomial function to accurately describe the nonlinear relationship between valve core opening and theoretical brake pressure. Braking pressure refers to the actual hydraulic pressure exerted by the hydraulic braking system on the brake (such as a brake caliper or brake shoe), a key physical quantity for achieving vehicle deceleration or stopping. The mapping relationship refers to the deterministic correspondence between valve core opening and brake pressure established through a mathematical function; that is, given a certain opening value, the corresponding brake pressure value can be uniquely calculated using this function. Theoretical braking pressure refers to the braking pressure value calculated under ideal conditions (ignoring factors such as pressure source limitations) based solely on the valve core opening and preset polynomial fitting coefficients, reflecting the static characteristics of the electro-hydraulic proportional reverse braking valve itself.

[0094] Understandably, this involves modeling an electro-hydraulic proportional reverse braking valve:

[0095] The electro-hydraulic proportional reverse brake valve controls the filling and releasing of oil in the hydraulic system by adjusting the valve core opening, thus affecting the braking pressure. Considering the superposition of physical effects such as mechanical friction between the valve core and valve body, a fifth-order polynomial is used to fit the valve core opening *x* and its theoretical output braking pressure. The highly nonlinear relationship between them can flexibly capture the nonlinear characteristics between them, while avoiding overfitting due to excessive complexity. The calculation formula is as follows, where a, b, c, d, e, and f are all polynomial fitting coefficients (unit MPa).

[0096]

[0097] The actual output braking pressure of the electro-hydraulic proportional reverse brake valve can be calculated using the following formula. Controllable Not exceeding its pressure source ( ), and introduce a non-linear scaling exponent. This enhances its sensitivity to pressure source anomalies (accumulator pressure drop). (Rapidly decreasing), on the other hand, ensuring that the output pressure is not negative, which is consistent with physical reality.

[0098] Step S30: Based on the accumulator pressure after the limit, the preset nonlinear scaling index and the theoretical braking pressure, construct a mathematical function of the actual output braking pressure of the electro-hydraulic proportional reverse braking valve to obtain the first dynamic mathematical model.

[0099] It should be noted that the preset nonlinear scaling exponent is a power-law parameter used to adjust the dependence of the actual output braking pressure on the accumulator pressure. Its introduction enhances the model's sensitivity to pressure source anomalies (such as accumulator pressure drops) and ensures that the output conforms to physical constraints. The actual output braking pressure refers to the true braking pressure value ultimately output by the electro-hydraulic proportional reverse braking valve after comprehensively considering the accumulator pressure after limiting, the theoretical braking pressure, and the nonlinear scaling effect. This serves as the key output variable of the entire hydraulic braking system. The first dynamic mathematical model refers to the set of dynamic equations that mathematically describe the input-output relationship of key components in the hydraulic braking system in the time domain. It reflects the physical behavior of variables such as pressure, flow rate, and displacement as they change over time during operation.

[0100] It is understandable that the first dynamic mathematical model can be represented as follows:

[0101]

[0102] in, This refers to the actual output braking pressure of the electro-hydraulic proportional reverse brake valve. This refers to the theoretical braking pressure. This refers to the accumulator pressure after the amplitude is limited. This refers to the preset pressure upper limit. This refers to the preset nonlinear scaling index.

[0103] Step S40: Construct a second dynamic mathematical model of the pressure relief valve based on the leakage and abnormal pressure drop of the pressure relief valve in the hydraulic braking system;

[0104] It should be noted that a pressure relief valve is a virtual or equivalent control element in a hydraulic braking system used to simulate oil pipe leakage faults. It releases a portion of the hydraulic oil through a controllable opening, artificially introducing pressure loss similar to real leakage to reproduce different degrees of oil pipe leakage in a simulation environment. Leakage amount refers to the equivalent hydraulic oil loss simulated by the pressure relief valve at the current simulation moment, which gradually increases over time to reflect the development process of the oil pipe leakage fault. Abnormal pressure drop refers to the additional pressure loss caused by the simulated oil pipe leakage, distinct from the pressure drop during normal system operation, and is a key indicator for determining whether a leakage fault has occurred in the braking system. The second dynamic mathematical model is a mathematical expression describing the dynamic relationship between the pressure relief valve opening and the resulting pressure loss. This model can progressively simulate the time evolution characteristics of the leakage process and reflect the nonlinear flow-pressure drop behavior of the pressure relief valve. It is specifically used to generate abnormal simulation data that conforms to physical laws in the absence of real fault data.

[0105] As an example, the step of constructing a second dynamic mathematical model of the pressure relief valve based on the leakage and abnormal pressure drop of the pressure relief valve in the hydraulic braking system includes: calculating the leakage of the pressure relief valve in the hydraulic braking system based on a preset maximum pressure relief capacity and a preset pressure relief growth time; performing dead-zone processing on the initial opening of the pressure relief valve based on a preset dead-zone threshold to obtain an effective opening; normalizing the effective opening to obtain a normalized opening; calculating the abnormal pressure drop based on the normalized opening and a preset fitting index; and constructing a second dynamic mathematical model of the pressure relief valve based on the leakage, the abnormal pressure drop, the effective opening, and the preset dead-zone threshold.

[0106] It should be noted that the preset maximum pressure relief capacity refers to the maximum pressure loss that the pressure relief valve can cause when fully open, used to characterize the upper limit of system pressure drop under the most severe leakage conditions. The preset pressure relief growth time refers to the length of time required from the start of pressure relief to the pressure relief capacity reaching the preset maximum pressure relief capacity, used to control the gradual rate of leakage, making the pressure drop in the simulation more consistent with the real physical process. The preset dead zone threshold is a critical value for the opening of the pressure relief valve; below this value, the system does not produce effective pressure relief, used to simulate the response hysteresis phenomenon caused by mechanical clearance or sealing characteristics in real hydraulic components. The initial opening refers to the opening value corresponding to the original control input signal of the pressure relief valve before any processing, usually set by the fault simulation strategy. The effective opening refers to the actual opening value used in the pressure relief calculation after applying dead zone processing to the initial opening; that is, it is considered effective only when the initial opening exceeds the preset dead zone threshold. The normalized opening refers to the dimensionless value after linearly mapping the effective opening to the interval [0, 1], facilitating subsequent unified calculations and improving the model's universality and stability. The preset fitting index refers to the nonlinear power function exponent (1.5 in this embodiment) used when calculating abnormal pressure drop. It is used to characterize the nonlinear relationship between the opening degree of the pressure relief valve and the pressure drop, so that the adjustment is precise at small opening degree and the pressure relief is rapid at large opening degree.

[0107] It is understandable that a leak in the hydraulic braking system's oil lines causes a decrease in the system's output braking pressure. The actual leak does not occur instantaneously. The leakage amount within the current time interval t can be calculated using the following formula. The simulation of brake line leakage faults is progressive, avoiding numerical shocks and preventing pressure collapse from causing simulation instability. This is the preset pressure relief increase time, which represents the time required for the pressure relief capacity to reach its maximum value. This indicates the preset maximum pressure relief capacity.

[0108]

[0109] Adjust the pressure relief valve opening. Simulating different degrees of brake line leakage, the abnormal pressure drop is calculated using the following formula. This conforms to the nonlinear characteristics of a pressure relief valve. Firstly, the effective opening degree of the pressure relief valve... Dead zone processing simulates the dead zone characteristics of a real pressure relief valve. When the valve opening exceeds a preset dead zone threshold... The system will only respond when the signal is normalized, avoiding accidental pressure relief caused by minute control signals. The signal is then normalized and mapped to [0,1] for easier calculation. A fitting index of 1.5 is chosen to ensure the pressure relief valve has fine adjustment capabilities at smaller openings and rapid pressure relief at larger openings, resulting in a fast response speed. The second dynamic mathematical model is expressed as follows:

[0110]

[0111] in, This refers to the abnormal pressure drop. This refers to the effective opening degree. This refers to the amount of leakage. This refers to the preset dead zone threshold.

[0112] Step S50: Import the first dynamic mathematical model and the second dynamic mathematical model into the preset simulation software to obtain the oil pipe leakage fault simulation model, and use the oil pipe leakage fault simulation model to generate normal data and abnormal data.

[0113] It should be noted that the preset simulation software refers to Simulink (a model-based multi-domain simulation and model-based design tool) developed by MathWorks, used for modeling, simulating, and analyzing dynamic systems. In this embodiment, it serves as a platform integrating the first and second dynamic mathematical models. The oil pipe leakage fault simulation model is a digital simulation system built in Simulink that can fully reflect the dynamic behavior of the WJ-6 internal combustion loader's hydraulic braking system under normal operating conditions and different leakage levels. It outputs key variables such as braking pressure by coupling the mathematical models of the accumulator, electro-hydraulic proportional reverse brake valve, and pressure relief valve. Normal data refers to the simulation dataset composed of operating parameters such as system braking pressure and accumulator pressure generated by the electro-hydraulic proportional reverse brake valve under different opening commands when the pressure relief valve opening is set to zero (i.e., no leakage) in the oil pipe leakage fault simulation model. This dataset represents the behavioral characteristics of the hydraulic braking system in a healthy state. Abnormal data refers to the system operation data generated in the oil pipe leakage fault simulation model when the pressure relief valve is set to have a non-zero effective opening (simulating different degrees of oil pipe leakage), which includes abnormal pressure drop and pressure response changes. It is used to characterize the abnormal behavior mode when the hydraulic braking system experiences leakage fault.

[0114] As an example, the step of importing the first dynamic mathematical model and the second dynamic mathematical model into a preset simulation software to obtain a pipeline leakage fault simulation model includes: importing the first dynamic mathematical model and the second dynamic mathematical model into the preset simulation software; constructing an initial joint simulation model of the target output braking pressure based on the variable transfer relationship between the actual output braking pressure in the first dynamic mathematical model and the abnormal pressure drop in the second dynamic mathematical model; simulating the initial joint simulation model in the preset simulation software to obtain preliminary simulation results; acquiring real experimental test data and calculating the fitting accuracy between the real experimental test data and the preliminary simulation results based on the Pearson correlation coefficient and the coefficient of determination; when the fitting accuracy is less than a preset accuracy threshold, adjusting the internal parameters of the initial joint simulation model and returning to the step of simulating the initial joint simulation model in the preset simulation software to obtain preliminary simulation results; when the fitting accuracy is greater than or equal to the preset accuracy threshold, using the initial joint simulation model as the pipeline leakage fault simulation model.

[0115] It should be noted that the variable transitivity refers to the physical and logical coupling between the actual output braking pressure from the first dynamic mathematical model and the abnormal pressure drop from the second dynamic mathematical model. The latter, as a pressure loss term, is subtracted from the former, jointly determining the final braking pressure output of the system. The target output braking pressure refers to the final braking pressure value of the hydraulic braking system after considering the impact of pipe leakage. Its calculation logic is the actual output braking pressure minus the abnormal pressure drop, representing the key state variable output by the simulation model. The initial co-simulation model refers to the uncalibrated system model formed after initially connecting the first and second dynamic mathematical models in Simulink according to the variable transitivity. It has not yet undergone parameter optimization and only reflects the theoretical structure. The preliminary simulation results refer to the system response data obtained by running the simulation software after applying a specific input (such as a valve core opening step signal) to the initial co-simulation model, including timing outputs such as the target output braking pressure and accumulator pressure. Real experimental test data refers to the actual braking pressure and accumulator pressure, and other operational data, collected by sensors on a WJ-6 internal combustion loader or test bench under the same control input (such as the opening degree of the electro-hydraulic proportional reverse brake valve), used to verify the accuracy of the model. The Pearson correlation coefficient is a statistical indicator that measures the degree of linear correlation between real experimental test data and simulation results, ranging from [-1, 1]. The closer the absolute value is to 1, the more consistent the trends of the two. The coefficient of determination (CCO) is also relevant. Pearson correlation coefficient (PRC) is an index used to evaluate the ability of a simulation model to explain variations in real data. Its value ranges from [0, 1], with values ​​closer to 1 indicating a higher goodness of fit. Fit accuracy refers to the degree to which the reference simulation model approximates real experimental test data, as reflected by the Pearson correlation coefficient and the coefficient of determination. It serves as a quantitative standard for judging the usability of the model. The preset accuracy threshold is a manually set lower limit for fit accuracy, used to determine whether the reference simulation model meets engineering application requirements.

[0116] Understandably, using Simulink simulation software, based on the established mathematical equations and variable relationships between components of the accumulator, electro-hydraulic proportional reverse brake valve, and pressure relief valve, these components are combined into a whole to obtain a simulation model of oil pipe leakage fault in the loader's hydraulic braking system. The final output braking pressure of the system is... (i.e., the target output braking pressure), calculated using the following formula, is the actual output braking pressure of the electro-hydraulic proportional reverse brake valve. Abnormal pressure drop caused by brake line leakage difference.

[0117]

[0118] The simulation model is used to simulate the electro-hydraulic proportional reverse brake valve opening degree and system braking pressure under normal conditions (pressure relief valve opening degree is 0). The relationship between the changes, and the pressure changes of the accumulator as a result of pressure relief during accumulator charging and oil cooling. The process of change. By comparing and analyzing real experimental test data and simulation results, the internal parameters of the simulation model are adjusted.

[0119] And using the Pearson correlation coefficient r and the coefficient of determination The goodness of fit between real experimental data and simulation results is calculated to evaluate the accuracy of the simulation model. The Pearson correlation coefficient r measures the degree of linear correlation between real experimental data and simulation results. The value of [-1,1] is closer to 1, indicating a stronger linear correlation between the two and a better fit of the simulation model. It measures the goodness of fit of the simulation model. The closer the value is to 1, the better the simulation model fits the real data.

[0120]

[0121]

[0122] Where n is the total number of samples. This is the actual value. These are simulation values. and These are the sample means of real data and simulation results, respectively.

[0123] As an example, the steps of generating normal and abnormal data using the oil pipe leakage fault simulation model include: in the oil pipe leakage fault simulation model, setting the pressure relief valve opening of the second dynamic mathematical model to a first preset opening value, and controlling the operation of the oil pipe leakage fault simulation model according to the valve core opening of multiple different electro-hydraulic proportional reverse brake valves to obtain normal data; adjusting the pressure relief valve opening of the second dynamic mathematical model to multiple different second preset opening values, and controlling the operation of the oil pipe leakage fault simulation model according to the input of multiple valve core opening values ​​to obtain abnormal data.

[0124] It should be noted that the pressure relief valve opening degree is an input variable used in the second dynamic mathematical model to control the simulated leakage intensity. It represents the degree to which the pressure relief valve is open, and its value determines the magnitude of the abnormal pressure drop, thus reflecting the severity of the oil pipe leakage. The first preset opening value refers to the pressure relief valve opening value set when generating normal data. It is usually zero or below the dead zone threshold to ensure that the system has no additional leakage, thereby simulating the hydraulic braking system in a fault-free normal operating state. The second preset opening value refers to a series of pressure relief valve opening values ​​greater than the dead zone threshold set when generating abnormal data. These values ​​are used to simulate oil pipe leakage fault conditions of different degrees, with each value corresponding to a specific leakage intensity.

[0125] Understandably, in the oil pipe leakage fault simulation model, the pressure relief valve opening of the second dynamic mathematical model is first set to the first preset opening value to ensure that the system does not introduce additional leakage. Then, multiple different valve core opening values ​​of electro-hydraulic proportional reverse brake valves are input sequentially to drive the simulation model to run and collect the system braking pressure data output in each run, thereby obtaining normal data reflecting normal operating conditions. Next, the pressure relief valve opening of the second dynamic mathematical model is adjusted to multiple different second preset opening values. Under each pressure relief valve opening, multiple valve core opening values ​​are input again to control the oil pipe leakage fault simulation model to run and simultaneously collect the corresponding system braking pressure data, thereby generating abnormal data that can characterize different leakage severity levels.

[0126] Step S60: Based on the normal data and the abnormal data, oil pipe leakage fault detection is performed through an adaptive self-organizing mapping network.

[0127] It should be noted that the adaptive self-organizing map network refers to an improved self-organizing map (SOM) neural network. It introduces an adaptive mechanism on the basis of the traditional SOM, which can dynamically adjust training parameters such as learning rate and neighborhood radius according to the distribution characteristics of input data. It also combines supervised threshold optimization strategies (such as the Youden index method) to set a discrimination boundary for quantization error, thereby achieving effective differentiation between normal and abnormal patterns under the premise of unlabeled training. It is suitable for oil pipe leakage fault detection tasks based on normal and abnormal data.

[0128] Understandably, the process involves several steps. First, normal and abnormal data are integrated and divided into training, validation, and test sets according to a preset ratio. Then, using the valve core opening of the electro-hydraulic proportional reverse braking valve and the system braking pressure as input features, the adaptive self-organizing map network is iteratively trained using the training set. In each iteration, the best matching unit for each sample is determined, and the weights of that unit and its neighbors are updated using a Gaussian neighborhood function. Simultaneously, the preset learning rate and preset neighborhood radius decay exponentially until training is complete. Next, the validation set is input into the trained network, the quantization error of each sample is calculated, and the optimal anomaly detection threshold is determined using the Youden exponential threshold optimization method, combined with the true label (normal or abnormal). Finally, the test set is input into the same network, and the quantization error of each test sample is calculated. If the error is less than the anomaly detection threshold, it is determined that there is no oil pipe leak; otherwise, it is determined that there is an oil pipe leak, thus completing the oil pipe leak detection.

[0129] This embodiment provides a method for detecting oil pipe leakage faults in a hydraulic braking system. First, the target pressure of the accumulator is calculated, and the inertial delay of its actual response is simulated using a first-order low-pass filter. Then, a limiting process is performed using preset upper and lower pressure limits to obtain the limited accumulator pressure, making the model closer to the dynamic characteristics of the real system. Second, a fifth-order polynomial mapping relationship is established based on the valve core opening of the electro-hydraulic proportional reverse braking valve and preset polynomial fitting coefficients to obtain the theoretical braking pressure, accurately characterizing the nonlinear relationship between valve control input and output pressure. Next, the actual braking pressure of the electro-hydraulic proportional reverse braking valve is constructed by combining the limited accumulator pressure, the preset nonlinear scaling exponent, and the theoretical braking pressure. The system outputs a braking pressure function to form a first dynamic mathematical model, effectively reflecting the impact of pressure source limitations on braking output. Then, a second dynamic mathematical model is constructed based on the leakage rate of the pressure relief valve and abnormal pressure drop to simulate different degrees of pipe leakage faults. Subsequently, the two models are integrated into Simulink simulation software to obtain a pipe leakage fault simulation model. This model is then used to generate a dataset covering normal and various abnormal operating conditions, providing high-fidelity samples for algorithm training. Finally, based on the generated normal and abnormal data, an adaptive self-organizing map network is used for unsupervised learning, and the anomaly detection threshold is optimized using a validation set to achieve automatic detection of pipe leakage faults. This embodiment can accurately detect pipe leakage faults in the hydraulic braking system of a loader even when labeled fault data is lacking, braking pressure control is inaccurate, and threshold selection lacks theoretical basis.

[0130] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to that in Embodiment 1 above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 2 , Figure 2 This is a flowchart illustrating the second embodiment of the hydraulic braking system oil pipe leakage fault detection method of this application. Step S40 of the hydraulic braking system oil pipe leakage fault detection method includes steps S41 to S45:

[0131] Step S41: Integrate the normal data and the abnormal data, and divide them into training set, validation set and test set according to a preset ratio.

[0132] It should be noted that the preset ratio refers to the sample quantity allocation ratio among the training set, validation set, and test set, which is set manually during the data partitioning process. In this embodiment, it is 6:2:2. The training set, validation set, and test set refer to three non-overlapping subsets of data, which are obtained by dividing the integrated normal and abnormal data according to the standard machine learning process: the training set is used to train the adaptive self-organizing map network to learn the data distribution pattern under normal working conditions; the validation set is used to evaluate the detection performance at different thresholds after training and to determine the optimal anomaly detection threshold using the Youden index; the test set is used to independently evaluate the model's generalization ability and detection effect on data that did not participate in the training and parameter tuning process.

[0133] It is understandable that the training set, validation set, and test set all contain both normal and abnormal data.

[0134] Step S42: Construct an adaptive self-organizing mapping network based on a preset neighborhood radius, a preset learning rate, and preset grid parameters. The input features of the adaptive self-organizing mapping network are the valve core opening of the electro-hydraulic proportional reverse braking valve and the system braking pressure.

[0135] It should be noted that the preset neighborhood radius refers to the initial neighborhood range set during the initialization phase of the adaptive self-organizing map network. It is used to determine the size of the neuron region around the best-matching unit that participates in weight updates, and its value decays exponentially during training. The preset learning rate refers to the weight adjustment step size parameter set in the early stages of network training. It controls the speed at which neuron weights move towards the input sample in each iteration. This parameter also gradually decays according to a preset time constant during training to ensure model convergence. The preset grid parameters refer to the topological configuration of the two-dimensional competition layer in the adaptive self-organizing map network, including the number of grid rows and columns (i.e., m×n neurons), used to define the resolution and expressive power of the mapping space. The system braking pressure refers to the actual pressure value that the hydraulic braking system ultimately outputs and acts on the braking actuator. In this embodiment, it is obtained by subtracting the abnormal pressure drop caused by the pressure relief valve from the actual output braking pressure of the electro-hydraulic proportional reverse braking valve. It is a key state variable reflecting whether the system is leaking.

[0136] Understandably, firstly, a two-dimensional topology containing m×n neurons is created based on preset grid parameters (i.e., the neuron grid is set to m rows and n columns), and a weight vector with the same dimension as the input feature is initialized for each neuron. Secondly, preset neighborhood radius and preset learning rate are set for the initial training stage, and both are configured to decay exponentially with the number of iterations to ensure that the network can quickly organize the global structure in the early stage and finely adjust the local mapping in the later stage. Finally, the valve core opening degree of the electro-hydraulic proportional reverse braking valve and the system braking pressure are used as two-dimensional input vectors as input features of the adaptive self-organizing mapping network, so that the network can autonomously learn the joint distribution pattern of opening degree and pressure under normal working conditions based on these data in subsequent training, thereby establishing a reliable benchmark for anomaly detection.

[0137] Step S43: Iteratively train the adaptive self-organizing map network based on the training set to determine the best matching unit.

[0138] It should be noted that the Best Matching Unit (BMU) refers to the neuron with the smallest Euclidean distance between the current input sample (composed of the valve core opening of the electro-hydraulic proportional reverse braking valve and the system braking pressure) and the weight vectors of all neurons in the grid during each iteration of the adaptive self-organizing map network training. This neuron best represents the features of the current input sample.

[0139] Understandably, the self-organizing map network model is trained iteratively using the training dataset. For the random input sample in the k-th iteration... k = 1, 2, ..., K, where K is the total number of iterations. This represents the valve core opening of the electro-hydraulic proportional reverse brake valve for the k-th input sample. Let represent the system braking pressure of the k-th input sample. Calculate its Euclidean distance to all neurons in the grid using the following formula. Through a competition mechanism, find the BMU that is closest to it to ensure that the input sample is represented by the most similar neuron.

[0140]

[0141] in, This represents the component of the valve opening in the weight vector of the neuron in the i-th row and j-th column of the network at the k-th iteration; This represents the component of the system braking pressure in the weight vector of the neuron in the i-th row and j-th column of the network during the k-th iteration.

[0142] Step S44: Update the weights of the best matching unit and the neurons in the neighborhood of the best matching unit using a Gaussian neighborhood function, while simultaneously causing the preset learning rate and the preset neighborhood radius to decay exponentially, thus obtaining the trained adaptive self-organizing map network.

[0143] It should be noted that the Gaussian neighborhood function is a two-dimensional Gaussian distribution centered on the BMU. Neurons closer to the BMU receive larger update weights, thus ensuring the orderliness and local smoothness of the topology. Neuron weights refer to an adjustable parameter vector with the same dimension as the input features maintained by each neuron in the adaptive self-organizing map network. In this embodiment, it is a two-dimensional vector, corresponding to the valve core opening of the electro-hydraulic proportional reverse braking valve and the system braking pressure, respectively. This weight is continuously adjusted through training to gradually approximate the distribution pattern of the input data.

[0144] It is understandable that the Gaussian neighborhood function is used. The weights of neurons in the BMU and its neighborhood are updated, adjusting the weight vectors of the BMU and its neighboring neurons toward the input sample. Similar input samples activate nearby neurons in the self-organizing map network model grid, forming a topologically ordered mapping. is the decay function of the neighborhood radius.

[0145]

[0146]

[0147] The neighborhood function is calculated using the Gaussian function as shown in the following formula. Define the degree of influence of BMU on its neighboring neurons to ensure the topology preservation property of the self-organizing map network model. These are the coordinates of neuron (i, j) in a two-dimensional grid. () represents the coordinates of the BMU in a two-dimensional grid. is the decay function of the neighborhood radius.

[0148]

[0149] in, Represents the neighborhood radius.

[0150] The learning rate and neighborhood radius decay exponentially with each training iteration to ensure model convergence. The decay function of the learning rate is calculated using the following formula. decay function of neighborhood radius This allows the learning rate and neighborhood radius to gradually decrease during training, ensuring that the model quickly organizes the global structure in the early stages and makes local adjustments in the later stages, avoiding oscillations and improving convergence and stability. and These are the initial neighborhood radius (the initial value of the preset neighborhood radius) and the initial learning rate (the initial value of the preset learning rate), respectively. and These are the time constants for the neighborhood radius and the learning rate, respectively, which control the decay rate of both.

[0151]

[0152]

[0153] Where k represents the current iteration number.

[0154] Step S45: Detect pipeline leaks based on the validation set, the test set, and the trained adaptive self-organizing map network to obtain detection results.

[0155] It should be noted that the detection result refers to the final fault judgment conclusion output by comparing the second quantization error with the anomaly detection threshold after each sample in the test set is input into the trained adaptive self-organizing map network.

[0156] As an example, the step of detecting pipeline leaks based on the validation set, the test set, and the trained adaptive self-organizing map network to obtain a detection result includes: inputting the validation set into the trained adaptive self-organizing map network and calculating a first quantization error from each validation sample to the corresponding best matching unit; determining an anomaly detection threshold using the Youden exponent threshold optimization method based on the first quantization error and the true label of the validation sample; inputting the test set into the trained adaptive self-organizing map network and calculating a second quantization error for each test sample; obtaining a detection result indicating no pipeline leak when the second quantization error is less than the anomaly detection threshold; and obtaining a detection result indicating a pipeline leak when the second quantization error is greater than or equal to the anomaly detection threshold.

[0157] The first quantization error refers to the Euclidean distance between each sample in the validation set and its corresponding BMU weight vector after inputting each sample into the trained adaptive self-organizing map network. It measures the degree of deviation of the sample from the normal pattern. The true label refers to the known fault state identifier for each sample in the validation set during the data generation phase, set by the simulation model: if the sample comes from a condition where the pressure relief valve opening is the first preset value (no leakage), the label is "normal"; if it comes from a condition where the valve opening is the second preset value (leakage), the label is "abnormal". The Youden index threshold optimization method is a supervised threshold selection strategy based on the validation set. By iterating through the recall and false positive rate (FPR) corresponding to different thresholds, the Youden index is calculated, and the threshold that maximizes this index is selected as the optimal anomaly detection threshold to achieve the best balance between false negatives and false positives. The anomaly detection threshold is the critical quantization error value used to determine whether a sample is abnormal. It is determined on the validation set by the Youden index threshold optimization method and serves as the decision boundary distinguishing between normal and leakage fault states. The second quantization error refers to the Euclidean distance between each sample in the test set and its corresponding best-matching unit, calculated after inputting each sample into the trained adaptive self-organizing map network. Its purpose is to compare it with the anomaly detection threshold to generate the final fault detection result.

[0158] The optimal anomaly detection threshold is determined using the supervised learning-based Youden exponent threshold optimization method: for input samples The Euclidean distance to the best matching unit after K iterations is calculated as the quantization error, i.e., the anomaly detection index. Normal samples usually have small quantization errors, while abnormal samples have large quantization errors because they have different patterns from normal samples.

[0159]

[0160] Where QE represents the quantization error (dimensionless, only the values ​​of the variables on the left side of the equation are used in the calculation); x refers to the valve core opening value of the electro-hydraulic proportional reverse brake valve. This represents the weight value of BMU in the valve core opening dimension after the Kth iteration; This represents the weight value of BMU in the system braking pressure dimension after the Kth iteration.

[0161] Set candidate threshold ,if If the error is not specified, the predicted label for the input sample is 1 (anomaly); otherwise, the predicted label is 1 (anomaly). If the threshold is too high, the predicted label for the input sample will be 0 (normal). The setting of the candidate threshold directly affects the prediction performance of the anomaly detection algorithm. If the candidate threshold is set too high, anomalies may be missed; if the candidate threshold is set too low, normal samples will be misclassified as anomalies.

[0162] The step of determining the anomaly detection threshold using the Youden index threshold optimization method based on the first quantization error and the true labels of the validation samples includes: establishing a mapping relationship between the first quantization error and the true labels of the validation samples; setting multiple consecutive candidate thresholds, the value range of which is determined based on the maximum and minimum values ​​of the first quantization error; for each candidate threshold, statistically analyzing the corresponding number of correctly detected anomalies, the number of missed detections, the number of false positives, and the number of correctly detected normal samples according to the mapping relationship; calculating the recall rate corresponding to each candidate threshold based on the number of correctly detected anomalies and the number of missed detections, and calculating the false positive rate corresponding to each candidate threshold based on the number of false positives and the number of correctly detected normal samples; calculating the Youden index corresponding to each candidate threshold based on the recall rate and the false positive rate; and using the candidate threshold corresponding to the maximum value of the Youden index as the anomaly detection threshold.

[0163] The mapping relationship refers to a data structure that associates the first quantization error of each sample in the validation set with its true label ("normal" or "abnormal"). This allows for the accurate counting of various detection results based on whether the error exceeds the threshold and matches the true label, given any candidate threshold. Candidate thresholds are a series of threshold points selected evenly or with a certain step size between the minimum and maximum values ​​of the first quantization error for testing. Each candidate threshold represents a possible anomaly determination boundary and is used to evaluate detection performance under different thresholds. The number of correctly detected anomalies, the number of false negatives, the number of false positives, and the number of correctly detected normal samples refer to the four categories of results obtained after binary classification of all samples in the validation set at a certain candidate threshold: the number of correctly detected anomalies (True Positive, TP) is the number of samples that are truly anomalies and are correctly classified as anomalies; the number of false negatives (False Negative, FN) is the number of samples that are truly anomalies but are incorrectly classified as normal; the number of false positives (False Positive, FP) is the number of samples that are truly normal but are incorrectly classified as anomalies; and the number of correctly detected normal samples (True Negative, TN) is the number of samples that are truly normal and are correctly classified as normal.

[0164] This embodiment employs a supervised learning-based Youden index threshold optimization method, using validation set data to determine the optimal anomaly detection threshold. This method requires no complex parameter tuning and is highly interpretable. Recall and FPR corresponding to different thresholds are calculated using quantization error and true labels. Recall represents the proportion of anomaly samples that are correctly detected; the higher the recall, the fewer anomalies are missed. The specific calculation formula is as follows:

[0165]

[0166] The false positive rate (FPR) represents the proportion of normal samples that are falsely reported as abnormal. The lower the FPR, the fewer normal samples are falsely detected.

[0167]

[0168] The Youden index Jthreshold is obtained using the following formula, and the threshold that maximizes the Youden index is selected as the optimal threshold. (Anomaly detection threshold). Balancing the two key metrics for anomaly detection, Recall and FPR, involves maximizing Recall to detect as many anomalies as possible, and minimizing FPR to reduce false positives as much as possible.

[0169] Jthreshold = TPR - FPR

[0170] Detection performance was evaluated using precision, recall, and F1 score:

[0171] The quantization error of the test set data is calculated to obtain the predicted labels, and the performance of the anomaly detection algorithm is evaluated using the test samples. Three evaluation metrics are set, including Recall, Precision, and F1 score.

[0172] Precision indicates the proportion of samples that are actually abnormal among those that are detected as abnormal. The higher the precision, the fewer false positives there are.

[0173]

[0174] In the formula, FP represents the number of normal samples that are misclassified as abnormal (the number of false alarms).

[0175] The F1 score is the harmonic mean of recall and precision, which comprehensively evaluates anomaly detection capability. The higher the F1 score, the better the algorithm's detection performance. The specific calculation formula is as follows:

[0176]

[0177] This embodiment first integrates the normal and abnormal data generated by simulation and divides them into training, validation, and test sets according to a preset ratio to ensure that model training, threshold selection, and performance evaluation are performed on independent and consistently distributed data. Second, an adaptive self-organizing mapping network is constructed based on preset neighborhood radius, preset learning rate, and preset grid parameters. The valve core opening of the electro-hydraulic proportional reverse braking valve and the system braking pressure are used as input features to focus the network on key behavioral variables of the braking system. Then, the network is iteratively trained using the training set. In each iteration, the best matching unit corresponding to the input sample is determined, and the weights of this unit and its neighbors are updated using a Gaussian neighborhood function. Simultaneously, the learning rate and neighborhood radius decay exponentially, allowing the network to gradually form a stable topological mapping of the normal pattern. Finally, based on the validation set, the quantization error is calculated and combined with the real labels. The Youden exponential threshold optimization method is used to determine the anomaly detection threshold, which is then used to judge the quantization error of the test set, outputting the detection result of whether an oil pipe leak fault exists. This embodiment can accurately detect oil pipe leakage faults in the hydraulic braking system of a loader when there is a lack of labeled fault data, inaccurate braking pressure control, and no theoretical basis for threshold selection.

[0178] In one embodiment, a simulation model of hydraulic braking system oil pipe leakage fault is built in Simulink, such as... Figure 3 As shown, Figure 3 This diagram illustrates the simulation model for a hydraulic pipe leakage fault detection method in Embodiment 2 of this application. The model, used to simulate a hydraulic pipe leakage fault in a loader's hydraulic braking system, consists of multiple functional modules. The input is "opening degree x," representing the control signal of the electro-hydraulic proportional reverse brake valve. This signal serves as the input variable for both the accumulator and the electro-hydraulic proportional reverse brake valve. The "opening degree x display" is the visualization output node for this input signal. The accumulator module receives the opening degree x, time Δt, and current pressure. As input, it internally calculates the accumulator pressure at the next moment using dynamic equations. And output it to the subsequent stages, while also The module provides a display; it also includes a feedback loop for dynamic pressure response and stabilization. The electro-hydraulic proportional reverse brake valve module receives the opening degree x and the accumulator pressure. The theoretical braking pressure is calculated based on a pre-defined fifth-order polynomial relationship. The actual output braking pressure is corrected by a nonlinear scaling exponent, resulting in an output... The pressure relief valve module receives the output pressure from the electro-hydraulic proportional reverse brake valve and displays it. and an effective opening degree in the range of 0-100% (Generated by the second dynamic mathematical model), the pressure drop caused by leakage is calculated using a set abnormal pressure drop function. And negative feedback is introduced along the pressure path to form the final system output pressure. .final, The actual system braking pressure, calculated by the subtractor after deducting the abnormal pressure drop, is displayed. The entire process realizes a complete closed-loop simulation from control input to system output, which can be used to generate data under normal and abnormal operating conditions.

[0179] A static test was performed on the electro-hydraulic valves of the loader. With the vehicle stationary, the engine running, and the hydraulic oil temperature reaching operating temperature, and the driver not pressing the foot pedal, a step signal was set for the opening of the electro-hydraulic proportional reverse brake valve, changing in 5% increments from 100% to 0% to obtain the braking pressure output by the system. and accumulator output pressure By comparing and analyzing real experimental test data and simulation results, the internal parameters of the simulation model are adjusted.

[0180] Table 1 Simulation Parameter Table

[0181]

[0182] like Figure 4 As shown, the braking pressure output by the system was measured in the experiment. The simulation model compares the opening degree x of the electro-hydraulic proportional reverse brake valve and the braking pressure under normal conditions (pressure relief valve opening is 0). The curves showing the changes are basically consistent.

[0183] Using the coefficient of determination The Pearson correlation coefficient r was used to calculate the goodness of fit between the real experimental data and the simulation results, and the results are shown in the table below.

[0184]

[0185] A Pearson correlation coefficient (r) exceeding 0.9 indicates a very strong linear correlation between the actual and simulated values, showing a high degree of consistency between them; the coefficient of determination... A value greater than 0.95 indicates that the simulation model has high accuracy and good fitting effect.

[0186] In addition, please refer to Figure 5 , Figure 5 This is a schematic diagram of accumulator pressure change provided in Embodiment 2 of the hydraulic braking system oil pipe leakage fault detection method of this application. The diagram shows the simulated accumulator pressure in the simulation model. The curve shows the changes in pressure during hydraulic pump charging and depressurization processes, including pump oil cooling. The upper part is the accumulator pressure curve, with the horizontal axis representing time and the vertical axis representing accumulator pressure (unit: MPa). The curve starts at an initial pressure of approximately 13 MPa, then rises rapidly during the hydraulic pump charging phase, reaching a maximum of approximately 15.8 MPa, corresponding to the "accumulator upper limit pressure." After that, it enters a constant pressure stage, where the pressure slowly decreases due to system consumption and minor leaks, eventually stabilizing and approaching the "accumulator lower limit pressure" (set at 13 MPa). This curve shows a trend of rapid rise followed by slow decay, reflecting the dynamic balance characteristics of the accumulator between pressure replenishment and depressurization. The lower half is a schematic diagram of the hydraulic pump's operating state, also using time as the horizontal axis to represent the pump's operating modes at different time points: In the initial stage, it is in the "hydraulic pump charging and pressurizing state," where the pump continuously supplies oil to the accumulator, pushing the pressure up; when the pressure reaches the upper limit, the pump switches to the "pump oil cooling and pressurization" state, stopping active pressurization and only maintaining the system's basic pressure, thereby achieving periodic regulation of the accumulator pressure. Both parts together describe the typical response behavior of the accumulator pressure under normal operating conditions.

[0187] Using a high-precision simulation model, different opening degrees of the electro-hydraulic proportional reverse brake valve were set, and the pressure relief valve was adjusted to simulate different degrees of leakage in the brake oil pipe, generating abnormal data. For example... Figure 6 As shown in the figure, the hydraulic braking system output braking pressure generated by the simulation model is as follows: the electro-hydraulic proportional reverse brake valve opening is fixed at 55°, while the pressure relief valve openings are 0° (normal condition) and 35° (simulated leakage). The changes are shown on the horizontal axis, which represents the time series in seconds, and the vertical axis, which represents the braking pressure output by the system. The values ​​range from approximately 4.8 MPa to 6.6 MPa, reflecting the pressure variation trend under different set operating conditions. The graph contains two curves: a dashed line represents the normal operating state when the pressure relief valve opening is 0%, with the braking pressure remaining at a relatively stable level, fluctuating around 6.0 MPa, demonstrating the system's stability under leak-free conditions; the other solid line corresponds to the simulated leakage state when the pressure relief valve opening is 35%, where the braking pressure drops significantly and eventually stabilizes below 5.0 MPa, accompanied by greater pressure fluctuations than under normal conditions, clearly revealing the system pressure loss and instability caused by simulated oil pipe leakage. This comparison of the two sets of data intuitively presents the difference in braking system pressure response under normal and abnormal conditions, providing important reference for fault detection.

[0188] An anomaly detection dataset was constructed using a high-precision simulation model, generating a total of 42,007 data entries, including 30,005 normal entries and 12,002 anomaly entries. The dataset was divided into training, validation, and test sets in a 6:2:2 ratio, and a stratified sampling strategy was employed to ensure consistent proportions of each category across subsets. A self-organizing map network was trained using the training set, and the optimal anomaly detection threshold was determined for the validation set using the Youden exponent threshold optimization method. The performance of the adaptive self-organizing map network anomaly detection algorithm was evaluated using precision, recall, and F1 score on the test samples. The results are shown in the table below.

[0189]

[0190] Experimental results show that the adaptive self-organizing map network algorithm used in this application has excellent performance in detecting brake oil pipe leakage faults, with an accuracy of over 97%, few false positives, a recall rate of over 98%, few false negatives, and an F1 score of over 97%, demonstrating robust overall detection performance.

[0191] For example, to help understand the implementation process of the hydraulic braking system oil pipe leakage fault detection method obtained by combining this embodiment with the above embodiment one, please refer to... Figure 7 , Figure 7 A simplified flowchart of a method for detecting oil pipe leakage faults in a hydraulic braking system is provided, specifically:

[0192] First, input the valve spool opening x of the electro-hydraulic proportional reverse brake valve and the system braking pressure. The feature vectors are used as the original input data for subsequent model processing. Then, an adaptive self-organizing network structure is constructed based on preset grid parameters, neighborhood radius, and learning rate, and a weight vector is initialized for each neuron. Subsequently, the network is iteratively trained using the training set. The BMU (Block Memory Unit) is determined by calculating the distance between the input sample and the weights of each neuron, and the weights of the BMU and its neighbors are updated based on a Gaussian neighborhood function. Simultaneously, the learning rate and neighborhood radius decay exponentially during training to achieve topological learning of the joint distribution of opening and pressure under normal operating conditions. Afterward, the validation set is input into the trained network to calculate the quantization error. Combined with its true label, multiple candidate thresholds are iterated, and the recall and false positive rate corresponding to each threshold are calculated. The Youden index is then obtained, and the threshold that maximizes this index is selected as the optimal anomaly detection threshold. Next, each sample in the test set is input into the trained network, and its quantization error QE is calculated; then, the process proceeds to the judgment node "sample quantization error". If the condition is met, the output is "normal" with a prediction label of 0, indicating that no leakage fault was detected; otherwise, the output is "abnormal" with a prediction label of 1, indicating that a leakage fault exists. Finally, the prediction results of all samples are entered into the "Evaluate Algorithm Detection Performance" module to calculate indicators such as accuracy and recall, and to comprehensively measure the actual effect of the fault detection method.

[0193] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the oil pipe leakage fault detection method of the hydraulic braking system of this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0194] This application also provides a device for detecting oil pipe leakage faults in a hydraulic braking system. Please refer to [reference needed]. Figure 8 The hydraulic braking system oil pipe leakage fault detection device includes:

[0195] The accumulator pressure processing module 10 is used to calculate the target pressure of the accumulator in the hydraulic braking system, and to filter and limit the target pressure to obtain the limited accumulator pressure.

[0196] The proportional valve theoretical pressure module 20 is used to establish a mapping relationship between the valve core opening and the braking pressure based on the valve core opening of the electro-hydraulic proportional reverse brake valve in the hydraulic braking system and the preset polynomial fitting coefficient, so as to obtain the theoretical braking pressure of the electro-hydraulic proportional reverse brake valve.

[0197] The first dynamic model construction module 30 is used to construct a mathematical function of the actual output braking pressure of the electro-hydraulic proportional reverse braking valve based on the accumulator pressure after the limit, the preset nonlinear scaling index and the theoretical braking pressure, so as to obtain the first dynamic mathematical model.

[0198] The second dynamic model construction module 40 is used to construct a second dynamic mathematical model of the pressure relief valve based on the leakage and abnormal pressure drop of the pressure relief valve in the hydraulic braking system.

[0199] The simulation module 50 is used to import the first dynamic mathematical model and the second dynamic mathematical model into a preset simulation software to obtain a pipeline leakage fault simulation model, and to generate normal data and abnormal data using the pipeline leakage fault simulation model.

[0200] The fault detection module 60 is used to detect oil pipe leak faults based on the normal data and the abnormal data through an adaptive self-organizing mapping network.

[0201] The hydraulic braking system oil pipe leakage fault detection device provided in this application adopts the hydraulic braking system oil pipe leakage fault detection method in the above embodiments, which can solve the technical problem of how to accurately detect oil pipe leakage faults in the hydraulic braking system of a loader when there is a lack of labeled fault data, inaccurate braking pressure control, and no theoretical basis for threshold selection. Compared with the prior art, the beneficial effects of the hydraulic braking system oil pipe leakage fault detection device provided in this application are the same as those of the hydraulic braking system oil pipe leakage fault detection method provided in the above embodiments, and other technical features in the hydraulic braking system oil pipe leakage fault detection device are the same as those disclosed in the above embodiments, and will not be repeated here.

[0202] This application provides a hydraulic braking system oil pipe leakage fault detection device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the hydraulic braking system oil pipe leakage fault detection method in the above embodiment 1.

[0203] The following is for reference. Figure 9 This document illustrates a structural schematic diagram of a hydraulic braking system oil pipe leakage fault detection device suitable for implementing embodiments of this application. The hydraulic braking system oil pipe leakage fault detection device in this application embodiment may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital radio receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), vehicle terminals (e.g., vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 9 The hydraulic braking system oil pipe leakage fault detection device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0204] like Figure 9As shown, the hydraulic braking system's oil pipe leakage fault detection device may include a processing unit 1001 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in ROM (Read Only Memory) 1002 or a program loaded from storage device 1003 into RAM (Random Access Memory) 1004. RAM 1004 also stores various programs and data required for the operation of the hydraulic braking system's oil pipe leakage fault detection device. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via bus 1005. Input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, LCDs (Liquid Crystal Displays), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the hydraulic braking system's pipe leak fault detection device to wirelessly or wiredly communicate with other devices to exchange data. Although the figure shows a hydraulic braking system's pipe leak fault detection device with various systems, it should be understood that it is not required to implement or possess all the systems shown. More or fewer systems can be implemented alternatively.

[0205] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from ROM 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0206] The hydraulic braking system oil pipe leakage fault detection device provided in this application adopts the hydraulic braking system oil pipe leakage fault detection method in the above embodiments, which can solve the technical problem of how to accurately detect oil pipe leakage faults in the hydraulic braking system of a loader when there is a lack of labeled fault data, inaccurate braking pressure control, and no theoretical basis for threshold selection. Compared with the prior art, the beneficial effects of the hydraulic braking system oil pipe leakage fault detection device provided in this application are the same as those of the hydraulic braking system oil pipe leakage fault detection method provided in the above embodiments, and other technical features of the hydraulic braking system oil pipe leakage fault detection device are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0207] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0208] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the oil pipe leakage fault detection method of the hydraulic braking system in the above embodiments.

[0209] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, RAM (Random Access Memory), ROM (Read Only Memory), EPROM (Erasable Programmable Read Only Memory or Flash Memory), optical fibers, CD-ROM (CD-Read Only Memory), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0210] The aforementioned computer-readable storage medium may be included in the hydraulic braking system's oil pipe leakage fault detection device; or it may exist independently and not be assembled into the hydraulic braking system's oil pipe leakage fault detection device.

[0211] The aforementioned computer-readable storage medium carries one or more programs. When these programs are executed by the hydraulic braking system's oil pipe leakage fault detection device, the device performs the following actions: calculates the target pressure of the accumulator in the hydraulic braking system, and filters and limits the target pressure to obtain the limited accumulator pressure; establishes a mapping relationship between the valve core opening and braking pressure based on the valve core opening of the electro-hydraulic proportional reverse brake valve in the hydraulic braking system and a preset polynomial fitting coefficient, thereby obtaining the theoretical braking pressure of the electro-hydraulic proportional reverse brake valve; and determines the theoretical braking pressure based on the limited accumulator pressure. Based on the pressure, a preset nonlinear scaling exponent, and the theoretical braking pressure, a mathematical function of the actual output braking pressure of the electro-hydraulic proportional reverse braking valve is constructed to obtain a first dynamic mathematical model. A second dynamic mathematical model of the pressure relief valve is constructed based on the leakage and abnormal pressure drop of the pressure relief valve in the hydraulic braking system. The first and second dynamic mathematical models are imported into preset simulation software to obtain a pipe leakage fault simulation model, and normal and abnormal data are generated using this model. Based on the normal and abnormal data, pipe leakage fault detection is performed using an adaptive self-organizing mapping network.

[0212] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including LAN (Local Area Network) or WAN (Wide Area Network)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0213] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0214] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0215] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the hydraulic braking system oil pipe leakage fault detection method described above. This solves the technical problem of accurately detecting oil pipe leakage faults in the hydraulic braking system of a loader when there is a lack of labeled fault data, inaccurate braking pressure control, and no theoretical basis for threshold selection. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the hydraulic braking system oil pipe leakage fault detection method provided in the above embodiments, and will not be repeated here.

[0216] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the hydraulic braking system oil pipe leakage fault detection method described above.

[0217] The computer program product provided in this application can solve the technical problem of accurately detecting oil pipe leakage faults in the hydraulic braking system of a loader when there is a lack of labeled fault data, inaccurate braking pressure control, and no theoretical basis for threshold selection. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the oil pipe leakage fault detection method for the hydraulic braking system provided in the above embodiments, and will not be repeated here.

[0218] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A hydraulic brake system oil pipe leakage failure detection method characterized by, The method includes: Calculate the target pressure of the accumulator in the hydraulic braking system, and filter and limit the target pressure to obtain the limited accumulator pressure; Based on the valve core opening degree of the electro-hydraulic proportional reverse brake valve in the hydraulic braking system and the preset polynomial fitting coefficients, a mapping relationship between the valve core opening degree and the braking pressure is established to obtain the theoretical braking pressure of the electro-hydraulic proportional reverse brake valve. Based on the accumulator pressure after the amplitude limit, the preset nonlinear scaling index, and the theoretical braking pressure, a mathematical function of the actual output braking pressure of the electro-hydraulic proportional reverse braking valve is constructed to obtain the first dynamic mathematical model. A second dynamic mathematical model of the pressure relief valve is constructed based on the leakage and abnormal pressure drop of the pressure relief valve in the hydraulic braking system. The first dynamic mathematical model and the second dynamic mathematical model are imported into a preset simulation software to obtain a pipeline leakage fault simulation model, and normal data and abnormal data are generated using the pipeline leakage fault simulation model. Based on the normal data and the abnormal data, an adaptive self-organizing mapping network is used to detect oil pipe leaks. The second dynamic mathematical model is represented as follows: wherein, refers to the abnormal pressure drop, refers to the effective opening of the pressure relief valve, refers to the leakage amount, refers to a preset deadband threshold; The step of detecting oil pipe leaks using an adaptive self-organizing mapping network based on the normal data and the abnormal data includes: The normal data and the abnormal data are integrated and divided into training set, validation set and test set according to a preset ratio; An adaptive self-organizing mapping network is constructed based on a preset neighborhood radius, a preset learning rate, and preset grid parameters. The input features of the adaptive self-organizing mapping network are the valve core opening of the electro-hydraulic proportional reverse braking valve and the system braking pressure. The adaptive self-organizing map network is iteratively trained based on the training set to determine the best matching unit; The optimal matching unit and the neuron weights in the neighborhood of the optimal matching unit are updated by a Gaussian neighborhood function, while the preset learning rate and the preset neighborhood radius decay exponentially to obtain the trained adaptive self-organizing map network. Based on the validation set, the test set, and the trained adaptive self-organizing map network, oil pipe leakage fault detection is performed to obtain the detection result; The step of importing the first dynamic mathematical model and the second dynamic mathematical model into preset simulation software to obtain the oil pipe leakage fault simulation model includes: Import the first dynamic mathematical model and the second dynamic mathematical model into the preset simulation software; Based on the variable transfer relationship between the actual output braking pressure in the first dynamic mathematical model and the abnormal pressure drop in the second dynamic mathematical model, an initial joint simulation model of the target output braking pressure is constructed. The initial joint simulation model is simulated in the preset simulation software to obtain preliminary simulation results; Obtain real experimental test data, and calculate the fitting accuracy between the real experimental test data and the preliminary simulation results based on the Pearson correlation coefficient and the coefficient of determination; When the fitting accuracy is less than a preset accuracy threshold, the internal parameters of the initial co-simulation model are adjusted, and the process returns to the step of simulating the initial co-simulation model in the preset simulation software to obtain preliminary simulation results. When the fitting accuracy is greater than or equal to the preset accuracy threshold, the initial joint simulation model is used as the simulation model for the oil pipe leakage fault. The steps for generating normal and abnormal data using the oil pipe leakage fault simulation model include: In the oil pipe leakage fault simulation model, the pressure relief valve opening of the second dynamic mathematical model is set to the first preset opening value, and the oil pipe leakage fault simulation model is controlled to run according to the valve core opening of multiple different electro-hydraulic proportional reverse brake valves to obtain normal data. The pressure relief valve opening of the second dynamic mathematical model is adjusted to multiple different second preset opening values, and the oil pipe leakage fault simulation model is controlled to run according to the input of multiple valve core opening values ​​to obtain abnormal data.

2. The method of claim 1, wherein, The steps of calculating the target pressure of the accumulator in the hydraulic braking system, and filtering and limiting the target pressure to obtain the limited accumulator pressure include: Obtain the pressure of the accumulator at the previous moment, and calculate the target pressure of the accumulator using the explicit Euler method based on the previous pressure, the preset pressure compensation pressure, and the preset pressure loss. The target pressure is processed using a first-order low-pass filter to obtain the actual output pressure of the energy storage device. The actual output pressure of the accumulator is limited according to the preset upper pressure limit and the preset lower pressure limit to obtain the limited accumulator pressure. The first dynamic mathematical model is represented as follows: in, This refers to the actual output braking pressure of the electro-hydraulic proportional reverse brake valve. This refers to the theoretical braking pressure. This refers to the accumulator pressure after the amplitude is limited. This refers to the preset pressure upper limit. This refers to the preset nonlinear scaling index.

3. The method of claim 1, wherein, The step of constructing the second dynamic mathematical model of the pressure relief valve based on the leakage and abnormal pressure drop of the pressure relief valve in the hydraulic braking system includes: The leakage of the pressure relief valve in the hydraulic braking system is calculated based on the preset maximum pressure relief capacity and the preset pressure relief growth time. The initial opening of the pressure relief valve is subjected to dead zone processing based on a preset dead zone threshold to obtain the effective opening. The effective opening is normalized to obtain the normalized opening. The abnormal pressure drop is calculated based on the normalized opening degree and the preset fitting index; Based on the leakage amount, the abnormal pressure drop, the effective opening degree, and the preset dead zone threshold, a second dynamic mathematical model of the pressure relief valve is constructed.

4. The method of claim 1, wherein, The step of detecting pipeline leaks based on the validation set, the test set, and the trained adaptive self-organizing map network to obtain detection results includes: The validation set is input into the trained adaptive self-organizing map network, and the first quantization error of each validation sample to the corresponding best matching unit is calculated. Based on the first quantization error and the true label of the verification sample, the anomaly detection threshold is determined using the Youden index threshold optimization method. Input the test set into the trained adaptive self-organizing map network and calculate the second quantization error for each test sample; When the second quantization error is less than the anomaly detection threshold, the detection result is that there is no oil pipe leakage fault; When the second quantization error is greater than or equal to the anomaly detection threshold, the detection result is that there is an oil pipe leak.

5. An oil line leakage fault detection device of a hydraulic brake system characterized by comprising: The device includes: An accumulator pressure processing module is used to calculate the target pressure of the accumulator in the hydraulic braking system, and to filter and limit the target pressure to obtain the limited accumulator pressure. The proportional valve theoretical pressure module is used to establish a mapping relationship between the valve core opening and the braking pressure based on the valve core opening of the electro-hydraulic proportional reverse brake valve in the hydraulic braking system and a preset polynomial fitting coefficient, so as to obtain the theoretical braking pressure of the electro-hydraulic proportional reverse brake valve. The first dynamic model construction module is used to construct a mathematical function of the actual output braking pressure of the electro-hydraulic proportional reverse braking valve based on the accumulator pressure after the limit, the preset nonlinear scaling index and the theoretical braking pressure, so as to obtain the first dynamic mathematical model. The second dynamic model construction module is used to construct a second dynamic mathematical model of the pressure relief valve based on the leakage and abnormal pressure drop of the pressure relief valve in the hydraulic braking system; the second dynamic mathematical model is expressed as follows: wherein, refers to the abnormal pressure drop, refers to the effective opening of the pressure relief valve, refers to the leakage amount, refers to a preset deadband threshold; The simulation module is used to import the first dynamic mathematical model and the second dynamic mathematical model into preset simulation software to obtain a pipeline leak fault simulation model, and to generate normal and abnormal data using the pipeline leak fault simulation model. The step of importing the first dynamic mathematical model and the second dynamic mathematical model into the preset simulation software to obtain the pipeline leak fault simulation model includes: importing the first dynamic mathematical model and the second dynamic mathematical model into the preset simulation software; constructing an initial joint simulation model of the target output braking pressure based on the variable transfer relationship between the actual output braking pressure in the first dynamic mathematical model and the abnormal pressure drop in the second dynamic mathematical model; simulating the initial joint simulation model in the preset simulation software to obtain preliminary simulation results; acquiring real experimental test data, and calculating the fitting accuracy between the real experimental test data and the preliminary simulation results based on the Pearson correlation coefficient and the coefficient of determination. When the fitting accuracy is less than a preset accuracy threshold, the internal parameters of the initial co-simulation model are adjusted, and the simulation is returned to the preset simulation software to obtain preliminary simulation results. When the fitting accuracy is greater than or equal to the preset accuracy threshold, the initial co-simulation model is used as the oil pipe leakage fault simulation model. The step of generating normal and abnormal data using the oil pipe leakage fault simulation model includes: in the oil pipe leakage fault simulation model, setting the pressure relief valve opening of the second dynamic mathematical model to a first preset opening value, and controlling the operation of the oil pipe leakage fault simulation model according to the valve core opening of multiple different electro-hydraulic proportional reverse brake valves to obtain normal data; adjusting the pressure relief valve opening of the second dynamic mathematical model to multiple different second preset opening values, and controlling the operation of the oil pipe leakage fault simulation model according to the input of multiple valve core opening values ​​to obtain abnormal data. A fault detection module is used to detect oil pipe leaks using an adaptive self-organizing map network based on normal and abnormal data. The steps of detecting oil pipe leaks using the adaptive self-organizing map network based on normal and abnormal data include: integrating the normal and abnormal data and dividing them into a training set, a validation set, and a test set according to a preset ratio; constructing an adaptive self-organizing map network based on a preset neighborhood radius, a preset learning rate, and preset grid parameters, wherein the input features of the adaptive self-organizing map network are the valve core opening of the electro-hydraulic proportional reverse brake valve and the system braking pressure; iteratively training the adaptive self-organizing map network using the training set to determine the optimal matching unit; updating the optimal matching unit and the neuron weights within its neighborhood using a Gaussian neighborhood function, while simultaneously causing the preset learning rate and the preset neighborhood radius to decay exponentially, resulting in a trained adaptive self-organizing map network; and performing oil pipe leak detection based on the validation set, the test set, and the trained adaptive self-organizing map network to obtain the detection result.

6. An oil line leakage fault detection device of a hydraulic brake system characterized by comprising: The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the method for detecting oil pipe leakage faults in a hydraulic braking system as described in any one of claims 1 to 4.

7. A storage medium, characterized by The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the hydraulic braking system oil pipe leakage fault detection method as described in any one of claims 1 to 4.

Images