A large oil hydraulic press hydraulic pump fault diagnosis device and method

Abstract

This invention discloses a fault diagnosis device and method for a large hydraulic press hydraulic pump, belonging to the field of hydraulic component performance testing and fault diagnosis technology. The device includes a test bench and a computer. The test bench integrates multiple sensors and a data acquisition module, simulating various load and speed conditions by adjusting throttle valves, relief valves, and variable frequency motors to acquire parameters such as flow rate, pressure, and leakage of the hydraulic pump in real time. The computer includes dedicated analysis software and a convolutional neural network diagnostic module to visualize the acquired data and perform intelligent fault identification. Through multi-condition simulation and deep learning algorithms, this invention can accurately measure the dynamic decay process of the hydraulic pump, achieving early fault warning and diagnosis, effectively improving diagnostic accuracy and system operational safety, and providing data support for predictive maintenance of hydraulic pumps.

Description

Technical Field

[0001] This invention belongs to the field of hydraulic component performance testing and fault diagnosis technology, specifically relating to a fault diagnosis device and method for a large hydraulic press hydraulic pump. Background Technology

[0002] As industrial equipment develops towards intelligence and high reliability, the performance of hydraulic pumps, a key component of the hydraulic system and a core power transmission device, directly affects the overall operational stability of the machine. Current technologies for testing hydraulic pumps largely rely on traditional testing platforms, but these platforms generally suffer from limited testing capabilities and difficulty in reflecting dynamic characteristics under actual operating conditions. Specifically, existing technologies are mostly limited to steady-state parameter measurements, failing to analyze the dynamic response characteristics of hydraulic pumps under coupled conditions of varying loads and speeds, and lacking online monitoring capabilities for key degradation indicators such as internal leakage and volumetric efficiency. Furthermore, traditional methods are insufficient in identifying weak early-stage fault signals, making it difficult to provide timely warnings of potential failures, which can easily lead to sudden shutdowns or even system contamination.

[0003] Therefore, there is an urgent need to provide a hydraulic pump testing device and method that can realize multi-condition simulation, multi-parameter acquisition, and intelligent fault diagnosis function, so as to make up for the technical defects of conventional testing methods in tracking performance degradation trends and provide a reliable basis for the full life cycle management of hydraulic pumps. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a fault diagnosis device and method for hydraulic pumps in large hydraulic presses. This invention can accurately acquire core static and dynamic performance profiles of hydraulic pumps, such as flow-pressure characteristic curves and volumetric efficiency. By simulating complex load conditions and pressure shocks, it accurately measures the dynamic degradation process of hydraulic pump performance, overcoming the parameter limitations of online detection. This enables comprehensive, multi-parameter, and in-depth performance evaluation and fault diagnosis of hydraulic pumps, providing crucial data for predictive maintenance.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a fault diagnosis device for a large hydraulic press hydraulic pump, comprising a test bench and a computer terminal; the test bench is used to perform performance testing on the hydraulic pump under test, and integrates multiple high-precision sensors and data acquisition modules to acquire flow and pressure parameters of the hydraulic pump in real time during operation; the computer terminal is electrically connected to the actuators on the test bench through a programmable controller to control the operation of the test bench and realize the simulation and performance testing of the hydraulic pump under different working conditions; the computer terminal is electrically connected to the data acquisition module to acquire the detection data of each sensor; the computer terminal is equipped with a convolutional neural network diagnostic module and a dedicated analysis software module, the convolutional neural network diagnostic module performs fault diagnosis of the hydraulic pump based on the data acquired by the high-precision sensors through a convolutional neural network model, and the dedicated analysis software module is used to visualize the test data.

[0006] Furthermore, the test bench includes an oil tank, a drive motor, and a hydraulic pump under test; the oil tank is equipped with a level gauge and a thermometer; the drive motor is connected to the hydraulic pump under test via a coupling to drive the hydraulic pump under test; the suction port of the hydraulic pump under test is connected to the oil tank via a second filter; the leakage port of the hydraulic pump under test is connected to the oil tank via a first filter and a first check valve in sequence, and a first flow sensor is installed on the leakage oil pipeline to monitor changes in internal leakage in real time; the outlet port of the hydraulic pump under test is connected to a second check valve, a two-position three-way directional valve, a throttle valve, a fourth shut-off valve, a second flow sensor, an overflow valve, and a third filter in sequence, and then connected to the oil tank.

[0007] Furthermore, the outlet of the two-position three-way directional valve is connected to a pressure gauge and a pressure sensor via a second shut-off valve and a third shut-off valve, respectively, for real-time monitoring of system pressure fluctuations.

[0008] Furthermore, the test bench simulates different load conditions by adjusting the overflow valve and throttle valve, and achieves dynamic speed adjustment by combining the frequency conversion control of the drive motor, so as to complete the performance test of the hydraulic pump under various working conditions.

[0009] Furthermore, a measuring branch is also connected to the leaking oil pipeline. The measuring branch is connected to a detachable measuring cup through a first shut-off valve. By opening the first shut-off valve, the leaking oil is led to the measuring cup for manual measurement to verify the accuracy of the first flow sensor.

[0010] Furthermore, the dedicated analysis software module can automatically generate performance graphs including flow-pressure curves and volumetric efficiency curves, and supports historical data comparison and trend analysis.

[0011] The present invention also provides a method for diagnosing faults in a large hydraulic press pump based on the above-mentioned device, comprising the following steps:

[0012] The test bench is controlled by a computer to simulate various load and speed conditions, and the dynamic performance of the hydraulic pump under test is tested.

[0013] The hydraulic pump's operating data is collected in real time by multiple high-precision sensors integrated on the test bench. The sensing data includes flow data and pressure data.

[0014] The collected sensor data is input into a pre-trained convolutional neural network model, which automatically extracts features and classifies faults from the input sensor data, and outputs the fault diagnosis results of the hydraulic pump.

[0015] Furthermore, the simulated multiple load conditions include:

[0016] Throttling loading mode: Switch the two-position three-way directional valve to the right position and open the fourth shut-off valve so that the hydraulic oil forms back pressure through the throttle valve. Change the system pressure by adjusting the opening of the throttle valve.

[0017] Overflow loading mode: Switch the two-position three-way directional valve to the left position, close the fourth shut-off valve, and achieve overload protection and extreme condition simulation by adjusting the pressure set by the overflow valve.

[0018] Furthermore, it also includes a sensor calibration step: collecting the leaked oil using a measuring cup within a specified time, weighing it with an electronic balance, and then converting the weight into a volumetric flow rate to verify the accuracy of the first flow sensor.

[0019] Furthermore, the convolutional neural network model includes a convolutional layer, a pooling layer, and a fully connected layer connected in sequence for processing sensor data; the fault diagnosis result includes fault type and fault degree, and the fault type includes bearing wear, blade damage, distributor plate wear, or cavitation.

[0020] Compared with existing technologies, this invention has the following advantages: It achieves intelligent identification and classification of typical hydraulic pump faults through deep learning algorithms, effectively improving diagnostic accuracy and automation levels; combined with the multi-condition testing capabilities of the test bench, a comprehensive fault sample library can be constructed, providing high-quality data support for convolutional neural network training, further enhancing model generalization ability and practical application reliability; through real-time monitoring and intelligent diagnosis, early fault warnings for hydraulic pumps are achieved, reducing the risk of sudden shutdowns and improving system operational safety and maintenance efficiency. It also supports historical data tracing and diagnostic result comparison, facilitating the analysis of fault evolution trends. Attached Figure Description

[0021] Figure 1 This is a block diagram illustrating the implementation principle of the large hydraulic press hydraulic pump fault diagnosis device provided in this embodiment of the invention.

[0022] Figure 2 This is a hydraulic schematic diagram of the test bench in an embodiment of the present invention;

[0023] Figure 3 This is an architecture diagram of the convolutional neural network model in an embodiment of the present invention.

[0024] In the diagram: 1. Pump under test; 2. Drive motor; 3.1. First filter; 3.2. Second filter; 3.3. Third filter; 4.1. First flow sensor; 4.2. Second flow sensor; 5.1. First check valve; 5.2. Second check valve; 6.1. First shut-off valve; 6.2. Second shut-off valve; 6.3. Third shut-off valve; 6.4. Fourth shut-off valve; 7. Measuring cup; 8. Pressure gauge; 9. Pressure sensor; 10. Relief valve; 11. Thermometer; 12. Level gauge; 13. Oil tank; 14. Two-position three-way directional valve; 15. Throttling valve. Detailed Implementation

[0025] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0026] It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of this application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0027] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0028] like Figure 1As shown, this embodiment provides a fault diagnosis device for a large hydraulic press hydraulic pump. The entire device consists of two parts: a test bench and a computer. The test bench is used to perform performance tests on the hydraulic pump under test. It integrates multiple high-precision sensors and a data acquisition module to acquire flow and pressure parameters during the pump's operation in real time. The computer is electrically connected to the actuators on the test bench via a programmable controller to control the test bench's operation and simulate and test the hydraulic pump under different operating conditions. The computer is also electrically connected to the data acquisition module to acquire the detection data from each sensor. The computer is equipped with a convolutional neural network diagnostic module and a dedicated analysis software module. The convolutional neural network diagnostic module uses a convolutional neural network model to diagnose faults in the hydraulic pump based on the data acquired by the high-precision sensors. The dedicated analysis software module is used to visualize the test data. The outlet of the two-position three-way directional valve 14 is also connected to a pressure gauge 8 and a pressure sensor 9 via a second shut-off valve 6.2 and a third shut-off valve 6.3, respectively, for real-time monitoring of system pressure fluctuations.

[0029] like Figure 2 As shown, the test bench includes an oil tank 13, a drive motor 2, and a hydraulic pump under test 1. A level gauge 12 and a thermometer 11 are installed inside the oil tank 13. The drive motor 2 is connected to the hydraulic pump under test 1 via a coupling to drive the pump. The suction port of the hydraulic pump under test 1 is connected to the oil tank 13 via a second filter 3.2. The leakage port of the hydraulic pump under test 1 is connected to the oil tank 13 sequentially via a first filter 3.1 and a first check valve 5.1, and a first flow sensor 4.1 is installed on the leakage line to monitor changes in internal leakage caused by wear within the pump body in real time. The outlet port of the hydraulic pump under test 1 is sequentially connected to a second check valve 5.2, a two-position three-way directional valve 14, a throttle valve 15, a fourth shut-off valve 6.4, a second flow sensor 4.2, an overflow valve 10, and a third filter 3.3, and then connected to the oil tank 13. The leaking oil pipeline is also connected to a measuring branch, which is connected to a detachable measuring cup 7 via a first shut-off valve 6.1. By opening the first shut-off valve 6.1, the leaking oil is led to the measuring cup 7 for manual measurement to verify the accuracy of the first flow sensor.

[0030] The test bench simulates different load conditions by adjusting the overflow valve and throttle valve, and achieves dynamic speed adjustment by combining the frequency conversion control of the drive motor, so as to complete the performance test of the hydraulic pump under various working conditions.

[0031] When throttle valve 13 is used for loading, the two-position three-way directional valve 14 switches to the right position, the fourth shut-off valve 6.4 opens, and the hydraulic oil forms back pressure through throttle valve 13 to load the tested pump 1. Adjusting the opening of throttle valve 13 can change the system pressure, thereby obtaining pressure-flow characteristic curves under different loads. When relief valve 10 is used for loading, the two-position three-way directional valve 14 switches to the left position, the fourth shut-off valve 6.4 closes, and overload protection and extreme condition simulation are achieved by adjusting the pressure set by relief valve 10.

[0032] When measuring the leakage oil volume using measuring cup 7, the leaked oil must be collected within a specified time. After weighing with an electronic balance, the volumetric flow rate is converted to verify the monitoring accuracy of the first flow sensor 4.1 and ensure the reliability of the internal leakage data. After calibration, the first shut-off valve 6.1 is closed and the measuring cup 7 is disconnected. The first flow sensor 4.1 then measures the leakage volume. During the test, the computer software collects data from each sensor in real time, performs real-time analysis and storage, and automatically plots performance graphs such as flow-pressure curves and volumetric efficiency curves to achieve various performance tests of the hydraulic pump.

[0033] This embodiment also provides a method for diagnosing faults in the hydraulic pump of a large hydraulic press based on the above-mentioned device, including the following steps:

[0034] S1. The test bench is controlled by a computer to simulate various load conditions and speed conditions to conduct dynamic performance tests on the hydraulic pump under test.

[0035] S2. Real-time acquisition of sensing data during the operation of the hydraulic pump using multiple high-precision sensors integrated on the test bench, including flow data and pressure data.

[0036] S3. The collected sensor data is input into a pre-trained convolutional neural network model. The convolutional neural network model automatically extracts features and classifies faults from the input sensor data, outputting fault diagnosis results for the hydraulic pump. The fault diagnosis results include the fault type (bearing wear, blade damage, distributor plate wear, cavitation, etc.) and the degree of fault. Figure 3 As shown, the convolutional neural network model includes a convolutional layer, a pooling layer, and a fully connected layer connected in sequence, used to process sensor data.

[0037] Before starting the test, check that all pipeline connections are sealed reliably and confirm that the oil level in the tank is above the middle position of the sight glass. Start the drive motor 2 and run it at low speed, gradually increasing the speed to the rated operating condition, and simultaneously adjust the throttle valve 13 or the relief valve 10 to set the load pressure.

[0038] 1. Loading method selection: When using throttle valve 13 for loading, the two-position three-way directional valve 14 switches to the right position, the fourth shut-off valve 6.4 opens, and the hydraulic oil forms back pressure through throttle valve 13 to load the tested pump 1; adjusting the opening of the throttle valve can change the system pressure, thereby obtaining the pressure-flow characteristic curves under different loads; when using relief valve 10 for loading, the two-position three-way directional valve 14 switches to the left position, the fourth shut-off valve 6.4 closes, and overload protection and extreme condition simulation are achieved by adjusting the pressure set by the relief valve.

[0039] 2. Calibration and Measurement of Leakage: When measuring the leakage oil volume using measuring cup 7, open the first shut-off valve 6.1, collect the leaked oil within the specified time, weigh it using an electronic balance, and convert it into volumetric flow rate to verify the monitoring accuracy of the first flow sensor 4.1 and ensure the reliability of the internal leakage data. After calibration, close the first shut-off valve 6.1 and disconnect the measuring cup 7, resuming real-time measurement of the leakage volume by the first flow sensor 4.1.

[0040] 3. Data Acquisition and Diagnosis: The sampling frequency is set via computer software to collect data in real-time from pressure sensor 9, the first flow sensor 4.1, and the second flow sensor 4.2. Combined with temperature monitoring by thermometer 11, oil temperature changes are monitored to ensure the test is conducted within the standard temperature range. Data is recorded after each test point has been running stably for 3 minutes, and this process is repeated to complete multiple operating condition cycles. After the test, the machine is stopped, residual pressure in the pipeline is released, and the data is exported for analysis to generate a test report. The test report includes key parameters of the hydraulic pump under different speed and load combinations, such as flow rate, pressure, leakage, and volumetric efficiency. The data is presented in both graphical and numerical formats for easy comparison and analysis. For abnormal fluctuations or deviations from the design curve...

[0041] The collected sensor data (pressure, flow rate, etc.) is preprocessed (filtered, normalized) and then input into a pre-trained convolutional neural network model. The convolutional neural network model is trained using a historical fault sample database, which contains data on various fault types and degrees, including bearing wear, blade damage, distributor plate wear, and cavitation. The model outputs the current fault type and severity of the hydraulic pump.

[0042] 4. Results Output: Dedicated computer software automatically generates performance graphs such as flow-pressure curves and volumetric efficiency curves, and displays diagnostic results. The system issues warnings for abnormal fluctuations or deviations from the design curves. After testing, a test report is generated, containing key parameters of the hydraulic pump under different speed and load combinations, including flow rate, pressure, leakage, and volumetric efficiency. The data is presented in both graphical and numerical formats.

[0043] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A fault diagnosis device for a large hydraulic press hydraulic pump, characterized in that, The system includes a test bench and a computer. The test bench is used to perform performance testing on the hydraulic pump under test. It integrates multiple sensors and a data acquisition module to acquire flow and pressure parameters of the hydraulic pump in real time. The computer is electrically connected to the actuators on the test bench via a programmable controller to control the test bench's operation and simulate and test the hydraulic pump under different operating conditions. The computer is also electrically connected to the data acquisition module to acquire detection data from each sensor. The computer is equipped with a convolutional neural network diagnostic module and a dedicated analysis software module. The convolutional neural network diagnostic module uses a convolutional neural network model to diagnose faults in the hydraulic pump based on the sensor data, while the dedicated analysis software module is used to visualize the test data.

2. The fault diagnosis device for hydraulic pumps of large hydraulic presses according to claim 1, characterized in that, The test bench includes an oil tank, a drive motor, and a hydraulic pump under test. The oil tank is equipped with a level gauge and a thermometer. The drive motor is connected to the hydraulic pump under test via a coupling to drive its operation. The suction port of the hydraulic pump under test is connected to the oil tank via a second filter. The leakage port of the hydraulic pump under test is connected to the oil tank sequentially via a first filter and a first check valve, and a first flow sensor is installed on the leakage pipeline to monitor internal leakage changes in real time. The outlet port of the hydraulic pump under test is sequentially connected to a second check valve, a two-position three-way directional valve, a throttle valve, a fourth shut-off valve, a second flow sensor, an overflow valve, and a third filter, before being connected to the oil tank.

3. The fault diagnosis device for hydraulic pumps of large hydraulic presses according to claim 2, characterized in that, The outlet of the two-position three-way directional valve is also connected to a pressure gauge and a pressure sensor via a second shut-off valve and a third shut-off valve, respectively, for real-time monitoring of system pressure fluctuations.

4. The fault diagnosis device for hydraulic pumps of large hydraulic presses according to claim 2, characterized in that, The test bench simulates different load conditions by adjusting the overflow valve and throttle valve, and achieves dynamic speed adjustment by combining the frequency conversion control of the drive motor, so as to complete the performance test of the hydraulic pump under various working conditions.

5. The fault diagnosis device for hydraulic pumps of large hydraulic presses according to claim 2, characterized in that, The leaking oil pipeline is also connected to a measuring branch, which is connected to a detachable measuring cup via a first shut-off valve. By opening the first shut-off valve, the leaking oil is led to the measuring cup for manual measurement to verify the accuracy of the first flow sensor.

6. The fault diagnosis device for hydraulic pumps of large hydraulic presses according to claim 1, characterized in that, The dedicated analysis software module can automatically generate performance graphs including flow-pressure curves and volumetric efficiency curves, and supports historical data comparison and trend analysis.

7. A method for diagnosing faults in a large hydraulic press pump based on the device described in any one of claims 1-6, characterized in that, Includes the following steps: The test bench is controlled by a computer to simulate various load and speed conditions, and the dynamic performance of the hydraulic pump under test is tested. The hydraulic pump's operating data is collected in real time by multiple sensors integrated on the test bench. The sensor data includes flow data and pressure data. The collected sensor data is input into a pre-trained convolutional neural network model, which automatically extracts features and classifies faults from the input sensor data, and outputs the fault diagnosis results of the hydraulic pump.

8. The method for diagnosing faults in the hydraulic pump of a large hydraulic press according to claim 7, characterized in that, The simulated various load conditions include: Throttling loading mode: Switch the two-position three-way directional valve to the right position and open the fourth shut-off valve so that the hydraulic oil forms back pressure through the throttle valve. Change the system pressure by adjusting the opening of the throttle valve. Overflow loading mode: Switch the two-position three-way directional valve to the left position, close the fourth shut-off valve, and achieve overload protection and extreme condition simulation by adjusting the pressure set by the overflow valve.

9. The method for diagnosing faults in the hydraulic pump of a large hydraulic press according to claim 7, characterized in that, It also includes a sensor calibration step: collecting the leaked oil using a measuring cup within a specified time, weighing it, and converting it into volumetric flow rate to verify the accuracy of the first flow sensor.

10. The method for diagnosing faults in the hydraulic pump of a large hydraulic press according to claim 7, characterized in that, The convolutional neural network model includes a convolutional layer, a pooling layer, and a fully connected layer connected in sequence, used to process sensor data; the fault diagnosis results include fault type and fault degree, and the fault type includes bearing wear, blade damage, distributor plate wear, or cavitation.

Images