A verification system and method for an intelligent hydraulic pump state diagnosis model

Abstract

The present application provides a verification system and method for an intelligent hydraulic pump state diagnosis model. The system provided by the present application includes a database module, a real-time simulation module and a state diagnosis module connected in sequence; the real-time simulation module is used to call the database module based on the verification requirements of the state diagnosis model, generate sensor signals based on the acquired target real test data, and transmit the sensor signals to the state diagnosis module; the state diagnosis module is used to collect target sensor signals in the sensor signals based on the verification requirements of the state diagnosis model, determine the state of the intelligent hydraulic pump based on the target sensor signals, and verify the performance of the state diagnosis model based on the intelligent hydraulic pump state. The system and method provided by the present application are used to improve the convenience of intelligent hydraulic pump state diagnosis and the authenticity of data through a model verification system that combines software and hardware, further improve the confidence and test efficiency of model verification, and provide a solid foundation for the development of intelligent electromechanical products.

Description

Technical Field

[0001] The present application relates to the technical field of model verification, and in particular to a verification system and method for an intelligent hydraulic pump state diagnosis model. Background Art

[0002] With the development of intelligent mechatronics technology, the combination of artificial intelligence and mechatronics is becoming closer and closer. More and more hydraulic pumps will develop into intelligent hydraulic pumps, becoming an important direction for enterprise automation, intelligence and efficiency. In the design and development process of intelligent hydraulic pumps, it is not only necessary for the intelligent hydraulic pump to automatically complete the preset tasks, but also to continuously learn and optimize its own operation strategies according to environmental changes and data accumulation. This requires the intelligent hydraulic pump to diagnose the status of existing data quickly and accurately.

[0003] In the past development process, the rapidity and accuracy of the intelligent hydraulic pump status diagnosis model were mostly verified in two ways. One is to build the status diagnosis model in the simulation stage, and then use simulation data to train and optimize the model; the other is to conduct system status diagnosis tests based on the product test data after the intelligent hydraulic pump is developed, and then optimize iteratively. Although the former has the characteristics of fast simulation and quick adjustment, the verification result is not confident because it is simply verified using simulation data without verifying the real data in the physical hardware and software system. After the product is developed, it is still necessary to optimize and adjust the model; although the latter can directly use real data to complete the verification of the status diagnosis function, it has the disadvantages of complex process, inconvenient collection, huge cost and low efficiency.

[0004] In order to overcome the shortcomings of the above two methods, the present invention proposes a verification system for an intelligent hydraulic pump state diagnosis model. The convenience of intelligent hydraulic pump state diagnosis and the authenticity of data are improved through a model verification system that combines software and hardware, and the confidence and test efficiency of model verification are further improved, thereby providing a solid foundation for the development of intelligent electromechanical products. Summary of the invention

[0005] In view of this, the present application provides a verification system and method for an intelligent hydraulic pump status diagnosis model, which is used to improve the convenience of intelligent hydraulic pump status diagnosis and the authenticity of data through a model verification system that combines software and hardware, and further improve the confidence and test efficiency of model verification, thereby providing a solid foundation for the development of intelligent mechatronic products.

[0006] Specifically, the present application is implemented through the following technical solutions:

[0007] In a first aspect, the present application provides a verification system for a state diagnosis model of an intelligent hydraulic pump, the verification system comprising a database module, a real-time simulation module and a state diagnosis module mounted on a physical chip connected in sequence; the state diagnosis module comprises a state diagnosis model;

[0008] The database module includes a plurality of data units, each of which contains data of different types, which at least include simulation data and test data; the simulation data is collected during the intelligent hydraulic pump modeling process, and the simulation data is used to train the state diagnosis model; the test data includes historical test data and real test data, and the historical test data is collected during the historical test of the intelligent hydraulic pump; the real test data is obtained by combining the split results of the simulation data and the split results of the historical test data under the same working condition and state, and the splitting methods of different data types are different, and the real test data is used to verify the performance of the state diagnosis model;

[0009] The real-time simulation module includes a real-time simulation machine and a simulation output card connected to each other, wherein the real-time simulation machine is used to call the database module based on the verification requirements of the state diagnosis model, and generate a sensor digital signal based on the acquired target real test data, and the simulation output card is used to convert the sensor digital signal into a sensor analog signal and transmit it to the state diagnosis module;

[0010] The state diagnosis module is used to collect the target sensor signal in the sensor simulation signal based on the verification requirements of the state diagnosis model, determine the state of the intelligent hydraulic pump based on the target sensor signal, and verify the performance of the state diagnosis model based on the intelligent hydraulic pump state.

[0011] The second aspect of the present application provides a verification method for an intelligent hydraulic pump state diagnosis model, and the verification method for an intelligent hydraulic pump state diagnosis model is applied to a verification system for an intelligent hydraulic pump state diagnosis model as described in any one of the first aspects of the present application; the method comprises:

[0012] The state diagnosis module collects a target sensor signal from the sensor signals based on the verification requirement of the state diagnosis model;

[0013] The real-time simulation module calls the database module based on the verification requirements of the state diagnosis model, simulates and generates target sensor signals based on the acquired target real test data, and transmits the target sensor signals to the state diagnosis module;

[0014] The state diagnosis module determines a state of the intelligent hydraulic pump based on the target sensor signal, and verifies performance of the state diagnosis model based on the intelligent hydraulic pump state.

[0015] The verification system and method of the intelligent hydraulic pump state diagnosis model provided by the present application, through the combination of software and hardware, at the software level, the real-time simulator can call the database module on demand, use the simulation data to quickly simulate the sensor digital signals under various working conditions, and convert them into analog signals with the simulation output card, which not only saves the cost and time of building a complex physical experimental environment, but also provides accurate input for model verification with highly repeatable and controllable simulation sensor signals, meeting the needs of various test scenarios. At the hardware level, the physical chip equipped with the state diagnosis model is responsible for collecting analog signals and diagnosing in real time. Its high-speed calculation and stability ensure the timeliness and accuracy of diagnosis, close to the real industrial scene, and make the verification results more credible. The database module contains a variety of data, which can generate a large amount of test data of different working conditions after splitting and combining. The real-time simulator can also be flexibly configured to simulate extremely complex working conditions. At the same time, the state diagnosis module feeds back the diagnosis results in real time, which is convenient for rapid adjustment and optimization, and improves the adaptability of the model to the state diagnosis of intelligent hydraulic pumps under various working conditions. It has the advantages of convenient simulation verification and reliable physical verification, effectively improves the convenience of state diagnosis, data authenticity, enhances the confidence of model verification and test efficiency, and lays a solid foundation for the development of intelligent electromechanical products. The present application uses both simulation data and test data to verify the intelligent hydraulic pump state diagnosis model. In the simulation stage of the intelligent hydraulic pump state diagnosis model, simulation verification is performed through simulation data. In the physical verification stage of the intelligent hydraulic pump state diagnosis model, physical verification is performed through real test data. Compared with the previous model verification only through single simulation data or test data, the present application not only has the convenience and speed of simulation verification, but also has the advantages of high credibility of physical verification. It can quickly and reliably perform state diagnosis of intelligent hydraulic pumps, improve the convenience of intelligent hydraulic pump state diagnosis and the authenticity of data, and further improve the confidence of model verification and test efficiency, thereby providing a solid foundation for the development of intelligent electromechanical products. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A schematic diagram of the structure of a verification system for an intelligent hydraulic pump state diagnosis model provided in Example 1 of the present application;

[0017] Figure 2 This is a flow chart of a verification method for the intelligent hydraulic pump status diagnosis model provided in Example 2 of the present application. DETAILED DESCRIPTION

[0018] Here, exemplary embodiments are described in detail, and examples thereof are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application.

[0019] The terms used in this application are only for the purpose of describing specific embodiments and are not intended to limit this application. The singular forms of "a", "said" and "the" used in this application are also intended to include plural forms, unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used in this article refers to and includes any or all possible combinations of one or more associated listed items.

[0020] It should be understood that although the terms first, second, third, etc. may be used in the present application to describe various information, these information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of the present application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0021] Specific embodiments are given below to introduce the technical solution of the present application in detail.

[0022] Figure 1 This is a schematic diagram of the structure of the verification system of the intelligent hydraulic pump status diagnosis model provided in Example 1 of the present application. Figure 1 The verification system provided in this embodiment includes a database module, a real-time simulation module and a state diagnosis module mounted on a physical chip connected in sequence; the state diagnosis module includes a state diagnosis model;

[0023] The database module includes a plurality of data units, each of which contains data of different types, which at least include simulation data and test data; the simulation data is collected during the intelligent hydraulic pump modeling process, and the simulation data is used to train the state diagnosis model; the test data includes historical test data and real test data, and the historical test data is collected during the historical test of the intelligent hydraulic pump; the real test data is obtained by combining the split results of the simulation data and the split results of the historical test data under the same working condition and state, and the splitting methods of different data types are different, and the real test data is used to verify the performance of the state diagnosis model;

[0024] The real-time simulation module includes a real-time simulation machine and a simulation output card connected to each other, wherein the real-time simulation machine is used to call the database module based on the verification requirements of the state diagnosis model, and generate a sensor digital signal based on the acquired target real test data, and the simulation output card is used to convert the sensor digital signal into a sensor analog signal and transmit it to the state diagnosis module;

[0025] The state diagnosis module is used to collect the target sensor signal in the sensor simulation signal based on the verification requirements of the state diagnosis model, determine the state of the intelligent hydraulic pump based on the target sensor signal, and verify the performance of the state diagnosis model based on the intelligent hydraulic pump state.

[0026] For details, please refer to Figure 1 The verification system of the intelligent hydraulic pump state diagnosis model includes a database module, a real-time simulation module and a state diagnosis module, wherein the database module is interconnected with the real-time simulation module, and the real-time simulation module is interconnected with the state diagnosis module. The state diagnosis module is mounted on a physical chip and includes a state diagnosis model. This application aims to verify the performance of the state diagnosis model.

[0027] Furthermore, the database module is used to store data, which includes multiple data units, each of which contains different types of data, and the different types of data may include simulation data and test data. The simulation data is collected during the intelligent hydraulic pump modeling process, and it contains models of intelligent hydraulic pumps in various ways. The corresponding simulation data can be selected according to demand to train the state diagnosis model. The test data includes historical test data and real test data. The historical test data is collected during the historical test of the intelligent hydraulic pump, and it stores offline data such as temperature, pressure, and vibration of the intelligent hydraulic pump. The real test data is obtained by combining the split results of the simulation data and the split results of the historical test data. The real test data is used to verify the performance of the state diagnosis model.

[0028] Please refer to Figure 1 The multiple data units include simulation data units and test data units. The data contained in the simulation data units are simulation data, and the data contained in the test data units are test data. The simulation data units and the test data units are established in the host computer Labview local file.

[0029] Optionally, in combination with the above description, the purpose of this application is to verify the performance of the intelligent hydraulic pump state diagnosis model. Therefore, it is necessary to first train the state diagnosis model and then verify the trained state diagnosis model. In the training stage of the state diagnosis model, the real-time simulation module calls the simulation data in the database module, and the simulation data is input into the state diagnosis model after being processed by the real-time simulation module and the state diagnosis module to train the state diagnosis model; after the state diagnosis model is trained, the real-time simulation module calls the real test data in the database module, and the real test data is input into the state diagnosis model after being processed by the real-time simulation module and the state diagnosis module to verify the performance of the state diagnosis model.

[0030] Specifically, since the real test data can simulate limited working conditions, fault states, and fault signals or fault levels of different dimensions, expanding it can fully improve the richness of samples and improve the accuracy of model verification. For situations under the same working conditions and states, such as the same working conditions and the same component failure, it can be the same fault degree, in order to obtain information of different dimensions, and the real data can be expanded, or it can be to obtain information of different fault degrees, and the real data can be expanded. At this time, the acquisition process of the real test data includes:

[0031] A first matching relationship between the simulation data and the historical verification data is established according to the working conditions and intelligent hydraulic pump states corresponding to the data, wherein each matching pair of the first matching relationship includes the simulation data and the historical verification data, and the working conditions and intelligent hydraulic pump states of the simulation data and the historical verification data in each matching pair are the same, and the set of matching pairs under the first matching relationship is used as the first basic data;

[0032] Splitting each data in the first basic data according to the periodicity of the data in each group of matching pairs, wherein the splitting units of data with different change periods are different;

[0033] The processed first basic data are combined according to the first matching relationship to obtain real test data.

[0034] Under this expansion requirement, each matching pair includes data A and data B. Data A and data B come from simulation data and historical verification data, respectively, that is, they come from different data units, but their working conditions and intelligent hydraulic pump states are the same. For each matching pair, data A and data B are data corresponding to the same intelligent hydraulic pump state under the same working condition. They can be data from different parts or data of different state levels of the same part. For example, if they are all fault states, they can be divided into serious faults and general faults. The data change periodicity of each working condition is inconsistent. First, each data is divided according to its own periodicity, such as unequal division into multiple sub-data such as 1s, 2s, 3s, 4s, or equal division, that is, variable period data and equal-length period data. Then, the data A and B in each matching pair are combined after division to obtain the amplified data in this case. The amplified data can be used as a supplementary sample of the same working condition and the same intelligent hydraulic pump state.

[0035] For the same working condition, different state types and state levels cannot be fully simulated. At this time, the number of samples is limited, which affects the model verification in this case. In this case, the process of obtaining the real test data includes:

[0036] A second matching relationship of all data in each data unit is established according to the working condition and the intelligent hydraulic pump state corresponding to the data, the combination mode of the data units of the data sources in each matching pair in the second matching relationship is not completely the same, the working condition of the two paired data of each matching pair in the second matching relationship is the same but the state is different, and the set of matching pairs is used as the second basic data;

[0037] Determine a segmentation method for each data in the second basic data according to the working condition, and segment each data in the second basic data based on the matched data segmentation method, wherein the data segmentation methods for different working conditions are different;

[0038] The segmented second basic data are combined according to the second matching relationship to obtain real test data.

[0039] In this case, the data of all data units are mixed together, and the data with the same working condition are first determined. A matching pair is constructed of the data with the same working condition but different intelligent hydraulic pump states, such as normal state and fault state. The data of each matching pair are C and D. The data units from which data C and D come may be the same or different, that is, the data with the same working condition of the same data unit can be paired, or the data with the same working condition of different data units can be paired. After the pairing is completed, the data change periodicity of each working condition is inconsistent, and the data C and D are also segmented. The segmentation method is the same as the segmentation direction of data A and B. Then, the segmented data of data C and D in each matching pair are combined to obtain the amplified data in this case. The amplified data can be used as a supplementary sample for the same working condition and different intelligent hydraulic pump states.

[0040] Optionally, the process of acquiring the real test data includes: acquiring basic data from the multiple data units; the basic data includes one or more of simulation data and historical test data, and the basic data includes multiple states; processing the basic data; combining the processed basic data according to the states to obtain real test data; the combination method includes arrangement, interleaving, and mixing.

[0041] Specifically, there are multiple combinations of methods for generating real test data, and the determination of the combination depends on the type and state of the basic data. When the basic data is only a single type of data (i.e., simulation data or historical test data), the real test data can be obtained by configuring the data length and arrangement method, for example, the data of size 10s is divided into multiple sub-data of 1s, 2s, 3s, 4s, etc. in unequal parts, and then the multiple sub-data are randomly arranged and combined again to generate new real test data; when the basic data contains data of multiple states (such as normal state data and fault state data), new real test data can be generated by fusing data of multiple states of the same size, for example, data of normal state data of size 5s and fault state data of size 5s are interleaved to generate real test data with multiple state data; when the basic data contains multiple types of data, and the states of these multiple types of data are the same (such as the same type of fault data of simulation data and historical test data), the data of the same state in these multiple types of data can be mixed to obtain real test data, for example, the simulation data of a certain fault of size 5s and the historical test data of the same type of fault of size 5s are mixed to generate the real test data of the fault.

[0042] The system provided in this embodiment, first, constructs the second matching relationship of all data in each data unit according to the working condition corresponding to the data and the state of the intelligent hydraulic pump. Since the data unit combination mode of the data source in the matching pair is not exactly the same, and the two paired data of each matching pair have the same working condition and different states, it is like building a precise "pairing network" for the data. Through this unique setting, data from different sources and different characteristics can be accurately associated according to actual needs, ensuring that the real test data generated subsequently covers a rich variety of operation and failure scenarios, effectively simulating the switching of various states of the hydraulic pump under complex working conditions, and providing a comprehensive and practical sample for model verification. Secondly, the segmentation method of each data in the second basic data is determined according to the working condition, and different working conditions correspond to different segmentation methods. Fully consider the differences in the inherent laws of the data of the hydraulic pump under different working conditions in actual work, for example, under high-pressure stable operating conditions and frequent start-stop conditions, the fluctuation characteristics and key information distribution of the sensor data are completely different. Through targeted segmentation, the most representative and diagnostically valuable data fragments under each working condition can be accurately extracted to avoid blind use of data. Furthermore, the second basic data after segmentation is combined according to the established second matching relationship to obtain real test data, which not only utilizes different types of data, but also covers data of different fault states, enriching the source of verification data from multiple dimensions. On the one hand, different data source methods verify and complement each other, greatly enriching the credibility of verification data, so that the simulated real sensor signal is highly close to the real scene in terms of generation, transmission and processing, so that the data environment contacted by the model is seamlessly connected with the actual application. On the other hand, the thorny problem of insufficient fault data source has been successfully solved. In the past, due to the lack of data under specific fault conditions, some rare but extremely harmful faults could not be effectively diagnosed and verified. Now, the potential of existing data can be fully tapped, and data combinations corresponding to various fault conditions can be created through clever splitting and combination, so as to carry out diagnosis and verification of more fault conditions and improve the authenticity of data in all aspects. This series of carefully designed steps makes the final verification result highly confident, effectively improves the confidence and test efficiency of model verification, lays a solid foundation for the reliability of the intelligent hydraulic pump state diagnosis model, and effectively promotes the development process of intelligent electromechanical products.

[0043] Furthermore, the real-time simulation module is used to call the database module based on the verification requirements of the state diagnosis model, obtain target real test data, generate sensor signals based on the obtained target real test data, and transmit the sensor signals to the state diagnosis module.

[0044] Please continue to refer to Figure 1The real-time simulation module includes a real-time simulation machine and a simulation output card. The real-time simulation machine and the simulation output card are connected to each other. The real-time simulation machine is connected to the simulation data unit and the test data unit in the database module.

[0045] Optionally, the real-time simulation machine is used to call the database module based on the lower computer Labview real-time system and preset calling parameters, and generate a real-time signal based on the acquired target real test data; the real-time signal is a real-time digital quantity; the calling parameters include at least the board output frequency; the simulation output card is used to map the real-time signal based on a mapping relationship to obtain a sensor signal, and transmit the sensor signal to the status diagnosis module based on the board output frequency.

[0046] Specifically, the real-time simulator and the simulation output card are the core components of the real-time simulation module. The real-time simulator generates digital signals based on the data obtained from the database module and the preset call parameters (such as board output frequency, working conditions, etc.). The simulation output card converts the digital signals output by the real-time simulator into output physical signals.

[0047] The calling parameters are set according to the verification requirements of the state diagnosis model. The calling parameters usually include the data source of the intelligent hydraulic pump, the board output channel and the board output frequency. Among them, the data source of the intelligent hydraulic pump represents the type of data called by the real-time simulator (for example, when the called data is simulation data, the data source is the simulation data unit; when the called data is real test data, the data source is the test data unit). The board output channel represents the output path of the signal. Each channel can output an independent signal. If the real-time simulator generates multiple signals (such as temperature and pressure), each signal can be transmitted through different board output channels. For example, channel 1 transmits temperature signals and channel 2 transmits pressure signals. The board output frequency represents the output rate of the signal.

[0048] In the specific implementation, the real-time simulation machine identifies the board and sets the relevant call parameters based on the identification results, including the data source of the intelligent hydraulic pump, the board output channel and the board output frequency. The database module is called by deploying and running the program in the lower-level Labview real-time system to generate real-time digital quantities. Then the simulation output card converts the real-time digital quantity into a sensor signal (such as temperature, pressure, and vibration signals) according to the set mapping relationship, and transmits the sensor signal to the status diagnosis module based on the set board output frequency.

[0049] It should be noted that, since the output frequency of the board is usually high, the simulation output card and the intelligent sensing unit have certain high-frequency characteristics to ensure that the transmitted sensor signal is not distorted.

[0050] Furthermore, the state diagnosis module is used to collect target sensor signals based on the verification requirements of the state diagnosis model, determine the state of the intelligent hydraulic pump based on the collected target sensor signals, and then verify the performance of the state diagnosis model.

[0051] Combined with the above description, the status diagnosis module includes the status diagnosis model. Please continue to refer to Figure 1 , the state diagnosis module also includes an intelligent perception unit and a data processing unit; wherein the intelligent perception unit is connected to the real-time simulation module, the intelligent perception unit is connected to the data processing unit, the data processing unit is connected to the state diagnosis model, the intelligent perception unit is used to determine the target sensor signal to be collected based on the verification requirements of the state diagnosis model, and collect the target sensor signal from the sensor signal, and transmit the target sensor signal to the data processing unit; the data processing unit is used to preprocess the target sensor signal and transmit the preprocessed target sensor signal to the state diagnosis model; the state diagnosis model is used to calculate the state coefficient based on the preprocessed target sensor signal, determine the state of the intelligent hydraulic pump based on the state coefficient, and verify the performance based on the intelligent hydraulic pump state.

[0052] Specifically, since not all sensor signals output by the real-time simulation module are necessary for the process of verifying the state diagnosis model, the intelligent sensing unit only determines the target sensor signals that need to be collected based on the verification requirements of the state diagnosis model (for example, when the verification requirements of the state diagnosis model require the collection of temperature, pressure, vibration and other signals, the target sensor signals that need to be collected are voltage, current, etc.), ensuring that all required sensor signals can be collected completely.

[0053] Optionally, the intelligent sensing unit is used to determine the target sensor signal to be collected based on the verification requirements of the state diagnosis model, and collect the target sensor signal from the sensor signal, including: the intelligent sensing unit determines the corresponding transmission mode based on relevant parameters of the target sensor signal; the transmission mode includes at least wired transmission and wireless transmission; when the transmission mode is wired transmission, the intelligent sensing unit collects the target sensor signal from the sensor signal based on the RS232 serial port; when the transmission mode is wireless transmission, the intelligent sensing unit collects the target sensor signal from the sensor signal based on a 5G wireless communication device.

[0054] Specifically, the relevant parameters determine the key characteristics of the signal transmission method, which is related to the physical characteristics of the sensor signal, environmental requirements, transmission requirements and other factors. The relevant parameters include the type of sensor signal, the frequency or bandwidth requirements of the sensor signal, the signal transmission distance, environmental factors, data real-time requirements, data transmission stability and reliability, power consumption requirements, etc. The transmission method characterizes the communication means of transmitting the sensor signal from the simulation output card to the intelligent sensing unit. Different transmission methods are suitable for different application scenarios and requirements. Transmission methods generally include wired transmission and wireless transmission. Common wired transmission methods include RS232 serial port, Ethernet, CAN bus, etc. Common wireless transmission methods include 5G wireless communication equipment, Wi-Fi, Bluetooth, etc.

[0055] In specific implementation, the intelligent perception module determines the target sensor signal to be collected based on the verification requirements of the state diagnosis model, comprehensively considers various relevant parameters of the target sensor signal, and determines the corresponding transmission method (wired transmission is superior in short distance, high stability, high real-time, low power consumption and complex environment, wireless transmission is more advantageous in long distance, flexibility, difficult wiring, and mobile device communication scenarios, especially suitable for occasions that require rapid deployment and remote monitoring. The specific choice of which transmission method needs to be weighed according to the actual signal type, transmission distance, real-time requirements, environmental complexity and power consumption requirements). When using wired transmission, the intelligent perception unit uses the RS232 serial port to transmit the target sensor signal in real time; when using wireless transmission, the intelligent perception unit uses 5G wireless communication equipment to directly transmit a large number of target sensor signals at a high rate.

[0056] The system provided in this embodiment comprehensively considers the actual signal type, transmission distance, real-time requirements, environmental complexity and power consumption requirements when determining the transmission method of the target sensor signal. Through the transmission method that combines wired transmission and wireless transmission, it can adapt to experiments with large data collection interference and inconvenient wiring in complex environments. That is, it can give full play to the advantages of wired transmission in short distances, high stability, high real-time performance, low power consumption and complex environments, and can also cover the characteristics of wireless transmission in scenarios with long distances, flexibility, difficulty in wiring, and the need for mobile device communication.

[0057] Optionally, the data processing unit is used to preprocess the target sensor signal, including: the data processing unit performs outlier processing on the target sensor signal; the data processing unit determines the target data type based on the recognition criteria of the state diagnosis model, converts the data type of the target sensor signal, and obtains the preprocessed target sensor signal; the data type of the preprocessed target sensor signal is the target data type.

[0058] Specifically, the data processing unit pre-processes the target sensor signal according to the data in the database module, eliminates or corrects errors and outliers in the target sensor signal, improves data quality, and uses statistics, machine learning and other methods to conduct in-depth analysis and mining of the target sensor signal, determine the target data type recognizable by the state diagnosis model, and convert the target sensor signal into a data type recognizable by the state diagnosis model.

[0059] Optionally, determining the state of the intelligent hydraulic pump based on the target sensor signal includes: extracting features from the target sensor signal to obtain multiple features; calculating a state coefficient based on the multiple features; the state coefficient characterizes the probability of each feature; determining the probability of a label corresponding to each feature based on the state coefficient and a mapping relationship; and determining the state of the intelligent hydraulic pump based on the probability of the label corresponding to each feature.

[0060] Specifically, the state coefficient represents the importance or probability of each feature in determining the state of the intelligent hydraulic pump, that is, the correlation between each feature and the intelligent hydraulic pump in a certain state, and the size of the state coefficient reflects the contribution of the feature to a certain state. The mapping relationship represents the correlation between each feature and the corresponding intelligent hydraulic pump state label.

[0061] In specific implementation, the received target sensor signals (which may include temperature, pressure, vibration and other sensor signals) are filtered and normalized to eliminate noise, correct errors and outliers. Important features are extracted from the preprocessed target sensor signals. These features can be statistics (such as mean, variance), frequency domain features (such as frequency components) or other related indicators. Furthermore, based on the extracted features, an intelligent algorithm (such as a machine learning model) is used to calculate the state coefficient of each feature. These coefficients characterize the contribution of each feature to the state of the intelligent hydraulic pump. Based on the state coefficient and the preset mapping relationship, the probability of the corresponding state label is calculated for each feature. The label probabilities corresponding to all features are combined, and the label with the highest probability is determined as the state of the intelligent hydraulic pump (when the label probability is greater than the preset threshold).

[0062] The verification system of the intelligent hydraulic pump state diagnosis model provided in this embodiment, firstly, through the combination of software and hardware, at the software level, the real-time simulator can call the database module as needed, use the simulation data to quickly simulate the sensor digital signals under various working conditions, and convert them into analog signals with the simulation output card, which not only saves the cost and time of building a complex physical experimental environment, but also provides accurate input for model verification with highly repeatable and controllable simulation sensor signals, meeting the needs of various test scenarios. At the hardware level, the physical chip equipped with the state diagnosis model is responsible for collecting analog signals and diagnosing in real time. Its high-speed calculation and stability ensure the timeliness and accuracy of diagnosis, close to the real industrial scene, and make the verification results more credible. The database module contains a variety of data, which can generate a large amount of test data of different working conditions after splitting and combining. The real-time simulator can also be flexibly configured to simulate extremely complex working conditions. At the same time, the state diagnosis module feeds back the diagnosis results in real time, which is convenient for rapid adjustment and optimization, improves the adaptability of the model to the state diagnosis of intelligent hydraulic pumps under various working conditions, and has the advantages of convenient simulation verification and reliable physical verification, which effectively improves the convenience of state diagnosis and data authenticity, enhances the confidence of model verification and test efficiency, and lays a solid foundation for the development of intelligent electromechanical products. This application uses both simulation data and test data to verify the intelligent hydraulic pump state diagnosis model. In the simulation stage of the intelligent hydraulic pump state diagnosis model, simulation verification is performed through simulation data. In the physical verification stage of the intelligent hydraulic pump state diagnosis model, physical verification is performed through real test data. Compared with the previous model verification only through a single simulation data or test data, this application not only has the convenience and speed of simulation verification, but also has the advantages of high credibility of physical verification. It can quickly and reliably perform the state diagnosis of the intelligent hydraulic pump, improve the convenience of the intelligent hydraulic pump state diagnosis and the authenticity of the data, and further improve the confidence of model verification and test efficiency, thereby providing a solid foundation for the development of intelligent electromechanical products. On the second aspect, first, the second matching relationship of all data in each data unit is constructed according to the working conditions corresponding to the data and the state of the intelligent hydraulic pump. Since the data unit combination method of the data source in the matching pair is not exactly the same, and the two paired data of each matching pair have the same working conditions but different states, this is like building a precise "pairing network" for the data. Through this unique setting, data from different sources and with different characteristics can be accurately linked according to actual needs, ensuring that the subsequent generated real test data covers a wide variety of operation and failure scenarios, effectively simulating the switching of various states of the hydraulic pump under complex working conditions, and providing comprehensive and practical samples for model verification. Secondly, the segmentation method of each data in the second basic data is determined according to the working conditions, and different segmentation methods correspond to different working conditions. Fully consider the differences in the inherent laws of the data of the hydraulic pump under different working conditions in actual work. For example, the fluctuation characteristics and key information distribution of the sensor data are completely different under high-pressure stable operating conditions and frequent start-stop conditions.Through targeted segmentation, the most representative and diagnostically valuable data fragments under each working condition can be accurately extracted to avoid blind use of data. Furthermore, the second basic data after segmentation is combined according to the established second matching relationship to obtain real test data, which not only utilizes different types of data, but also covers data of different fault states, enriching the source of verification data from multiple dimensions. On the one hand, different data source methods verify and complement each other, greatly enriching the credibility of verification data, so that the simulated real sensor signal is highly close to the real scene in terms of generation, transmission and processing, so that the data environment contacted by the model is seamlessly connected with actual application. On the other hand, the thorny problem of insufficient fault data source has been successfully solved. In the past, due to the lack of data under specific fault conditions, some rare but extremely harmful faults could not be effectively diagnosed and verified. Now, the potential of existing data can be fully tapped, and data combinations corresponding to various fault conditions can be created through clever splitting and combination, so as to carry out diagnosis and verification of more fault conditions and improve the authenticity of data in all aspects. This series of carefully designed steps makes the final verification result highly confident, effectively improves the confidence and test efficiency of model verification, lays a solid foundation for the reliability of the intelligent hydraulic pump state diagnosis model, and effectively promotes the development process of intelligent electromechanical products. On the third hand, in this application, the real-time simulation module has strong expansibility, can be expanded according to the needs of different state diagnosis of intelligent hydraulic pumps, and supports various types of sensors and different signal types (such as temperature, pressure, vibration, etc.). For example, when a new sensor is introduced or different working states need to be monitored, only the corresponding module configuration needs to be adjusted or increased without reconstructing the entire system. In addition, the real-time simulation module can quickly generate real-time signal data, which can be used to simulate the state of the intelligent hydraulic pump in a real working environment. This real-time performance can help to find problems in time, respond quickly, and improve the response speed of the system. By generating real-time data that matches the real conditions, the confidence of model verification is increased. For example, in the state diagnosis process, real-time data can provide more accurate feature information, thereby improving the accuracy of state recognition. Fourthly, when determining the transmission method of the target sensor signal, the present application comprehensively considers the actual signal type, transmission distance, real-time requirements, environmental complexity and power consumption requirements. Through the transmission method that combines wired transmission and wireless transmission, it can adapt to experiments with large data collection interference and inconvenient wiring in complex environments. It can give full play to the advantages of wired transmission in short distances, high stability, high real-time performance, low power consumption and complex environments, and can also cover the characteristics of wireless transmission in scenarios with long distances, flexibility, difficulty in wiring, and the need for mobile device communication.Fifthly, the simulation output card and intelligent sensing unit in the present application both have a high-rate sampling frequency. When the output frequency of the board is high, it can overcome the problem of signal distortion of voltage and current signals at high frequencies, ensure that the transmitted sensor signal is not distorted, and increase the confidence of model verification.

[0063] Corresponding to the aforementioned embodiment of a verification system for an intelligent hydraulic pump state diagnosis model, the present application also provides an embodiment of a verification method for an intelligent hydraulic pump state diagnosis model.

[0064] Figure 2 This is a flow chart of the verification method of the intelligent hydraulic pump status diagnosis model provided in Example 2 of this application. Figure 2 The method provided in this embodiment, the verification method of the intelligent hydraulic pump state diagnosis model is applied to the verification system of the intelligent hydraulic pump state diagnosis model provided in any one of the first aspects of the present application; the method comprises:

[0065] S201 . A state diagnosis module collects a target sensor signal from among the sensor signals based on a verification requirement of the state diagnosis model.

[0066] S202 , the real-time simulation module calls the database module based on the verification requirements of the state diagnosis model, simulates and generates a target sensor signal based on the acquired target real test data, and transmits the target sensor signal to the state diagnosis module.

[0067] S203 . The state diagnosis module determines the state of the intelligent hydraulic pump based on the target sensor signal, and verifies the performance of the state diagnosis model based on the state of the intelligent hydraulic pump.

[0068] In the specific implementation, the specific implementation process and implementation principle of steps S201 to S203 can be found in the description of embodiment 1, and will not be repeated here.

[0069] The method of this embodiment can be used to execute Figure 1 The steps, specific implementation principles and implementation processes of the system embodiment shown are similar and will not be repeated here.

[0070] The implementation process of the functions and effects of each unit in the above-mentioned device is specifically described in the implementation process of the corresponding steps in the above-mentioned method, and will not be repeated here.

[0071] For the device embodiment, since it basically corresponds to the method embodiment, the relevant parts can refer to the partial description of the method embodiment. The device embodiment described above is only schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present application scheme. A person of ordinary skill in the art can understand and implement it without paying any creative work.

[0072] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

Claims

1. A verification system for an intelligent hydraulic pump status diagnosis model, characterized in that: The verification system comprises a database module, a real-time simulation module and a state diagnosis module mounted on a physical chip connected in sequence; the state diagnosis module comprises a state diagnosis model; The database module includes a plurality of data units, each of which contains data of different types, which at least include simulation data and test data; the simulation data is collected during the intelligent hydraulic pump modeling process, and the simulation data is used to train the state diagnosis model; the test data includes historical test data and real test data, and the historical test data is collected during the historical test of the intelligent hydraulic pump; the real test data is obtained by combining the split results of the simulation data and the split results of the historical test data under the same working condition and state, and the splitting methods of different data types are different, and the real test data is used to verify the performance of the state diagnosis model; The real-time simulation module includes a real-time simulation machine and a simulation output card connected to each other, wherein the real-time simulation machine is used to call the database module based on the verification requirements of the state diagnosis model, and generate a sensor digital signal based on the acquired target real test data, and the simulation output card is used to convert the sensor digital signal into a sensor analog signal and transmit it to the state diagnosis module; The state diagnosis module is used to collect the target sensor signal in the sensor simulation signal based on the verification requirements of the state diagnosis model, determine the state of the intelligent hydraulic pump based on the target sensor signal, and verify the performance of the state diagnosis model based on the intelligent hydraulic pump state.

2. The system according to claim 1, characterized in that The process of obtaining the real test data includes: A first matching relationship between the simulation data and the historical verification data is established according to the working conditions and intelligent hydraulic pump states corresponding to the data, wherein each matching pair of the first matching relationship includes the simulation data and the historical verification data, and the working conditions and intelligent hydraulic pump states of the simulation data and the historical verification data in each matching pair are the same, and the set of matching pairs under the first matching relationship is used as the first basic data; Splitting each data in the first basic data according to the periodicity of the data in each group of matching pairs, wherein the splitting units of data with different change periods are different; The processed first basic data are combined according to the first matching relationship to obtain real test data.

3. The system according to claim 1, characterized in that The process of obtaining the real test data includes: A second matching relationship of all data in each data unit is established according to the working condition and the intelligent hydraulic pump state corresponding to the data, the combination mode of the data units of the data sources in each matching pair in the second matching relationship is not completely the same, the working condition of the two paired data of each matching pair in the second matching relationship is the same but the state is different, and the set of matching pairs is used as the second basic data; Determine a segmentation method for each data in the second basic data according to the working condition, and segment each data in the second basic data based on the matched data segmentation method, wherein the data segmentation methods for different working conditions are different; The segmented second basic data are combined according to the second matching relationship to obtain real test data.

4. The system according to claim 1, characterized in that The real-time simulation machine is used to call the database module based on the Labview real-time system of the lower computer and the preset calling parameters, and generate a real-time signal based on the acquired target real test data; the real-time signal is a real-time digital quantity; the calling parameters at least include the board output frequency; The simulation output card is used to map the real-time signal based on the mapping relationship to obtain the sensor signal, and transmit the sensor signal to the status diagnosis module based on the output frequency of the board card.

5. The system according to claim 1, characterized in that The state diagnosis module also includes an intelligent perception unit and a data processing unit; wherein the intelligent perception unit is connected to the real-time simulation module, the intelligent perception unit is connected to the data processing unit, and the data processing unit is connected to the state diagnosis model. The intelligent sensing unit is used to determine the target sensor signal to be collected based on the verification requirements of the state diagnosis model, collect the target sensor signal from the sensor signal, and transmit the target sensor signal to the data processing unit; The data processing unit is used to preprocess the target sensor signal and transmit the preprocessed target sensor signal to the state diagnosis model; The state diagnosis model is used to calculate a state coefficient based on the preprocessed target sensor signal, determine a smart hydraulic pump state based on the state coefficient, and verify performance based on the smart hydraulic pump state.

6. The system according to claim 5, characterized in that The data processing unit is used to pre-process the target sensor signal, including: The data processing unit performs abnormal value processing on the target sensor signal; The data processing unit determines the target data type based on the recognition standard of the state diagnosis model, converts the data type of the target sensor signal, and obtains a preprocessed target sensor signal; the data type of the preprocessed target sensor signal is the target data type.

7. The system according to claim 5, characterized in that The intelligent sensing unit is used to determine the target sensor signal to be collected based on the verification requirements of the state diagnosis model, and collect the target sensor signal from the sensor signal, including: The intelligent sensing unit determines a corresponding transmission mode based on relevant parameters of the target sensor signal; the transmission mode includes at least wired transmission and wireless transmission; When the transmission mode is wired transmission, the intelligent sensing unit collects the target sensor signal from the sensor signal based on the RS232 serial port; When the transmission mode is wireless transmission, the intelligent sensing unit collects the target sensor signal from the sensor signal based on a 5G wireless communication device.

8. The system according to claim 1, characterized in that During the training phase of the state diagnosis model, the real-time simulation module calls the simulation data in the database module, and the simulation data is input into the state diagnosis model after being processed by the real-time simulation module and the state diagnosis module to train the state diagnosis model; after training the state diagnosis model, the real-time simulation module calls the real test data in the database module, and the real test data is input into the state diagnosis model after being processed by the real-time simulation module and the state diagnosis module to verify the performance of the state diagnosis model.

9. The system according to claim 1, characterized in that The determining the state of the intelligent hydraulic pump based on the target sensor signal comprises: Extracting features from the target sensor signal to obtain multiple features; Calculate a state coefficient based on the features; the state coefficient represents the probability of each feature; Based on the state coefficient and the mapping relationship, determine the probability of the label corresponding to each feature; The state of the intelligent hydraulic pump is determined based on the probability of the label corresponding to each feature.

10. A verification method for an intelligent hydraulic pump state diagnosis model, characterized in that: The verification method of the intelligent hydraulic pump state diagnosis model is applied to the verification system of the intelligent hydraulic pump state diagnosis model according to any one of claims 1 to 9; the method comprises: The state diagnosis module collects a target sensor signal from the sensor signals based on the verification requirement of the state diagnosis model; The real-time simulation module calls the database module based on the verification requirements of the state diagnosis model, simulates and generates target sensor signals based on the acquired target real test data, and transmits the target sensor signals to the state diagnosis module; The state diagnosis module determines a state of the intelligent hydraulic pump based on the target sensor signal, and verifies performance of the state diagnosis model based on the intelligent hydraulic pump state.

Images