A method and device for the health management of special vehicles

Abstract

The present invention discloses a method and device for the health management of special vehicles. The method includes: obtaining the acquisition data of components corresponding to various sensors and storing it in a data processing and analysis device, where the acquisition data stored in the data processing and analysis device follows a normal distribution; sequentially selecting the acquisition data with probability values of μ - 2σ, μ - σ, μ, μ + σ, and μ + 2σ in the normal distribution to determine a weighted correction coefficient matrix; correcting the data stored in the data processing and analysis device based on the weighted correction coefficient matrix, and using functional data analysis methods to establish a functional relationship of the corrected data with time; determining the damage conditions of each component of the special vehicle over time based on the functional relationship. This method can correct the data volatility caused by vibration, accurately depict the functional curves of the parameters of each component changing continuously over time, improve the accuracy of the health management of each device, and achieve the health management of each component under the condition of data fluctuations.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle fault diagnosis, and particularly to a method and device for the health management of special vehicles. Background Art

[0002] Special vehicles have irreplaceable advantages in operations such as engineering construction and disaster relief. However, when special vehicles start and operate, due to the operation of equipment such as engines and transmissions, the whole vehicle will generate great vibrations, resulting in large fluctuations in the data collected by some equipment sensors (such as vibration sensors, speed sensors, etc.), which further affects the accuracy of the subsequent health management and trend prediction of each equipment. Summary of the Invention

[0003] In view of this, the present invention provides a method and device for the health management of special vehicles, which can solve the technical problem of low accuracy in the health management of special vehicles.

[0004] To solve the above technical problem, the present invention is implemented as follows.

[0005] A method for the health management of special vehicles, the method comprising:

[0006] Step S1: Various sensors configured on the special vehicle itself collect data of components corresponding to the various sensors; the collected data is sent to a data bus; a data acquisition device obtains the collected data of components corresponding to the various sensors from the data bus at regular intervals and stores it in a data processing and analysis device, and the collected data stored in the data processing and analysis device follows a normal distribution N(μ,σ 2 ), where σ is the standard deviation and μ is the expectation;

[0007] Step S2: Sequentially select the collected data with probability values of μ - 2σ, μ - σ, μ, μ + σ, and μ + 2σ in the normal distribution, construct a representative matrix for the collected data corresponding to each probability value, and obtain representative matrices D μ-2σ , D μ-σ , D μ , D μ+σ , and D μ+2σ ; generate a standard matrix based on all the data in the data processing and analysis device Determine the calibration coefficient matrix for mapping each representative matrix to the standard matrix, and based on each calibration coefficient matrix, determine the weighted calibration coefficient matrix;

[0008] Step S3: Calibrate the data stored in the data processing and analysis device based on the weighted calibration coefficient matrix, and use the functional data analysis method to establish a functional relationship of the calibrated data with respect to time; determine the damage conditions of each component of the special vehicle over time based on the functional relationship.

[0009] Preferably, the calibration coefficient matrix for mapping each representative matrix to the standard matrix is solved by the partial least squares algorithm (PLS) to obtain five calibration coefficient matrices S μ-2σ , S μ-σ , S μ , S μ+σ and S μ+2σ .

[0010] Preferably, a standard matrix is generated based on all the data in the data processing and analysis device Determine the calibration coefficient matrix for mapping each representative matrix to the standard matrix, and based on each calibration coefficient matrix, determine the weighted calibration coefficient matrix, including:

[0011] Classify all the data in the data processing and analysis device according to the component type, calculate the average value of each data of the same type of components, and use it as the element in the standard matrix to generate the standard matrix Establish the functional relationship between each representative matrix and the standard matrix respectively to form five functions, and solve the five functions to obtain five calibration coefficient matrices

[0012]

[0013] The determination of the weighted calibration coefficient matrix based on each calibration coefficient matrix includes:

[0014] Weight the calibration coefficient matrices corresponding to the five representative matrices respectively to obtain the weighted calibration coefficient matrix M for the standard matrix, M = α1S μ-2σ + α2S μ-σ + α3S μ + α4S μ+σ + α5S μ+2σ , where the values of α1, α2, α3, α4, and α5 are calculated according to the probability of the standard normal distribution, and the calculation method is:

[0015] According to the normal distribution function corresponding to N(μ, σ 2 ), calculate the probabilities corresponding to the probability values of μ - 2σ, μ - σ, μ, μ + σ, and μ + 2σ respectively, and denote them as P μ-2σ , P μ-σ , P μ , P μ+σ , P μ+2σ , where:

[0016]

[0017] Preferably, the step S3 includes:

[0018] Obtain the standard correction coefficient M, multiply each piece of data stored in the data processing and analysis device by the coefficient M to obtain the corrected data; store the corrected data in the database, select the corrected data within a preset time range from the database, and use the functional data analysis method to establish a functional relationship of the corrected data within the preset time range with respect to time; determine the damage conditions of each component of the special vehicle over time based on the functional relationship.

[0019] Preferably, the method for the health management of special vehicles further includes:

[0020] Step S4: Send the damage conditions of each component of the special vehicle over time to the display layer for visual display.

[0021] An apparatus for the health management of special vehicles provided by the present invention, the apparatus includes:

[0022] Acquisition module: Configured to collect data of corresponding components of various sensors configured on the special vehicle by the various sensors; send the collected data to the data bus; the data acquisition device obtains the collected data of the corresponding components of the various sensors from the data bus at regular intervals and stores it in the data processing and analysis device, and the collected data stored in the data processing and analysis device follows a normal distribution N(μ,σ 2 ), σ is the standard deviation, and μ is the expectation;

[0023] Correction module: Configured to sequentially select the collected data with probability values of μ - 2σ, μ - σ, μ, μ + σ, and μ + 2σ in the normal distribution, construct a representative matrix for the collected data corresponding to each probability value, and obtain the representative matrices D μ-2σ 、D μ-σ 、D μ 、D μ+σ and D μ+2σ ; generate a standard matrix based on all the data in the data processing and analysis device Determine the correction coefficient matrix for mapping each representative matrix to the standard matrix, and determine the weighted correction coefficient matrix based on each correction coefficient matrix;

[0024] Analysis module: Configured to correct the data stored in the data processing and analysis device based on the weighted correction coefficient matrix, and use the functional data analysis method to establish a functional relationship of the corrected data with respect to time; determine the damage conditions of each component of the special vehicle over time based on the functional relationship.

[0025] A computer-readable storage medium provided by the present invention, wherein multiple instructions are stored in the storage medium; the multiple instructions are used to be loaded and executed by a processor to perform the method as described above.

[0026] An electronic device provided by the present invention is characterized in that the electronic device includes:

[0027] A processor for executing multiple instructions;

[0028] A memory for storing multiple instructions;

[0029] Among them, the multiple instructions are used to be stored by the memory and loaded and executed by the processor to perform the method as described above.

[0030] Beneficial technical effects brought by the present invention:

[0031] (1) The present invention can correct data volatility caused by vibration and accurately depict the function curves of parameters such as the states and characteristics of each device changing continuously over time, improving the accuracy of health management of each device. The present invention realizes precise health management of each device under data fluctuation conditions.

[0032] (2) In data acquisition of the present invention, the minimum data acquisition period is 5 ms, which can more accurately characterize the continuous change trend of each device over time; in data correction, 5 representative values that can characterize data volatility are selected by means of probability, increasing the reliability of representative value selection; in function relationship establishment, the basis function can be flexibly selected according to data types to apply to periodic or non-periodic changing data.

[0033] (3) The precise health management of each device by the present invention can achieve timely detection of special vehicle equipment failures, accurate prediction of equipment operation trends, etc., thereby extending the service life of the equipment and reducing the equipment maintenance cost. Description of the Drawings

[0034] Figure 1 It is a schematic flowchart of the method for special vehicle health management of the present invention;

[0035] Figure 2 It is an equipment layout diagram of the method for special vehicle equipment health management provided by the present invention;

[0036] Figure 3 It is a schematic diagram of data correction principle in the method for special vehicle equipment health management provided by an embodiment of the present invention;

[0037] Figure 4 It is a schematic diagram of the establishment of function relationship principle in the method for special vehicle equipment health management provided by an embodiment of the present invention. Detailed Embodiments

[0038] The present invention will be described in detail below with reference to the drawings and embodiments.

[0039] As Figure 1 - Figure 2As shown in the figure, the present invention proposes a method for the health management of special vehicles, and the method includes:

[0040] Step S1: Various sensors configured on the special vehicle collect data of components corresponding to the various sensors; the collected data is sent to the data bus; the data acquisition device obtains the collected data of components corresponding to the various sensors from the data bus at regular intervals and stores it in the data processing and analysis device, and the collected data stored in the data processing and analysis device follows a normal distribution N(μ,σ 2 ), where σ is the standard deviation and μ is the expectation;

[0041] Step S2: Sequentially select the collected data with probability values of μ - 2σ, μ - σ, μ, μ + σ, and μ + 2σ in the normal distribution, construct a representative matrix for the collected data corresponding to each probability value, and obtain representative matrices D μ-2σ 、D μ-σ 、D μ 、D μ+σ and D μ+2σ ; generate a standard matrix based on all the data in the data processing and analysis device Determine the calibration coefficient matrix for mapping each representative matrix to the standard matrix, and based on each calibration coefficient matrix, determine the weighted calibration coefficient matrix;

[0042] Step S3: Calibrate the data stored in the data processing and analysis device based on the weighted calibration coefficient matrix, and use the functional data analysis method to establish the functional relationship of the calibrated data with time; determine the damage condition of each component of the special vehicle over time based on the functional relationship.

[0043] Further, the components corresponding to the various sensors include a special vehicle brake controller, a generator controller, an engine controller, various devices, and various device controllers.

[0044] In step S1, where: the data of the component includes the operating state and characteristic parameters.

[0045] The acquisition software in the data acquisition device collects the data of the components corresponding to the various sensors from the data bus at regular intervals and sends it to the data processing and analysis device for storage.

[0046] Further, in step S2, the calibration coefficient matrix for mapping each representative matrix to the standard matrix is solved by the partial least squares algorithm (PLS), and 5 calibration coefficient matrices S μ-2σ , S μ-σ , S μ , S μ+σ and S μ+2σ are obtained.

[0047] Generate a standard matrix based on all the data in the data processing and analysis device Determine the correction coefficient matrix for mapping each representative matrix to the standard matrix, and based on each correction coefficient matrix, determine the weighted correction coefficient matrix, including:

[0048] Classify all the data in the data processing and analysis device by component type, calculate the average value of each data for the same type of components, and use it as the element in the standard matrix to generate the standard matrix Establish the functional relationship between each representative matrix and the standard matrix respectively to form five functions, and solve the five functions to obtain five correction coefficient matrices

[0049]

[0050] The determination of the weighted correction coefficient matrix based on each correction coefficient matrix includes:

[0051] Weight the correction coefficient matrices corresponding to the five representative matrices respectively to obtain the weighted correction coefficient matrix M for the standard matrix, M = α1S μ-2σ +α2S μ-σ +α3S μ +α4S μ+σ +α5S μ+2σ , where the values of α1, α2, α3, α4, and α5 are calculated according to the probability of the standard normal distribution, and the calculation method is:

[0052] According to the normal distribution function corresponding to N(μ,σ 2 ), calculate the probabilities corresponding to μ - 2σ, μ - σ, μ, μ + σ, and μ + 2σ respectively, and denote them as P μ-2σ , P μ-σ , P μ , P μ+σ , P μ+2σ , where:

[0053]

[0054] For example, the probability at the point μ: P μ = 0.0080, the probabilities at the points μ - σ and μ + σ: P μ-σ = P μ+σ = 0.0049; the probabilities at the points μ - 2σ and μ + 2σ: P μ-2σ = P μ+2σ = 0.0011, then there are:

[0055]

[0056] The step S3 includes:

[0057] Obtain the standard correction coefficient M, multiply each piece of data stored in the data processing and analysis device by the coefficient M to obtain the corrected data; store the corrected data in the database, select the corrected data within a preset time range from the database, and use the functional data analysis method to establish a functional relationship of the corrected data within the preset time range with respect to time; determine the damage conditions of each component of the special vehicle over time based on the functional relationship.

[0058] In the present invention, according to the need, data of parameters such as the states and characteristics of each device within a certain time range are taken out from the database (this data is the corrected data), and a functional data analysis algorithm is used to establish a functional relationship of its continuous change with respect to time, so as to realize the change trend of the health states of each device of the special vehicle and achieve the purpose of implementing health management for each device.

[0059] In the present invention, the basis function type of the functional data analysis method is the B-spline basis, and the number of basis functions is set to the 4th order.

[0060] Further, the method for the health management of the special vehicle further includes:

[0061] Step S4: Send the damage conditions of each component of the special vehicle over time to the display layer for visual display.

[0062] In the present invention, the health management results of each device and data such as the original parameters and states are sent to the display layer for visual display.

[0063] The present invention provides an embodiment of a method for the health management of a special vehicle.

[0064] As Figure 2 shown, the acquisition and analysis system includes a device layer, a data acquisition layer, a data processing and analysis layer, and a display layer; after each device operates, it will send parameters such as its own state and characteristics to the data bus, and the acquisition software in the data acquisition layer collects vibration, rotation speed, control signals, etc. of devices such as the engine and generator from the data bus at a cycle of 5 ms and sends them to the data processing and analysis device for storage.

[0065] As Figure 3 shown, the correction of the acquired data is realized. The data processing and analysis module first reads 1000 acquisition data of signal 1, signal 2, …, signal 10 from the database respectively. It is known that the 1000 acquisition data of each signal follow a normal distribution N(μ,σ 2 )(σ is the standard deviation and μ is the expectation). Among the 10 signals, data with probabilities of μ - 2σ, μ - σ, μ, μ + σ, and μ + 2σ are respectively selected from the 1000 data as representative values of the random acquisition data to form 5 matrices D μ-2σ, D μ-σ , D μ , D μ+σ and D μ+2σ , after averaging 1000 acquisition data of 10 signals, a standard value is formed as a matrix Function relationships representing the matrix and the standard matrix are established respectively and After solving through PLS (Partial Least Squares Algorithm), 5 coefficient matrices S are obtained μ-2σ , S μ-σ , S μ , S μ+σ and S μ+2σ ; The 5 coefficient matrices are weighted and integrated to obtain the final calibration coefficient M = α1S μ-2σ +α2S μ-σ +α3S μ +α4S μ+σ +α5S μ+2σ , where the values of α1, α2, α3, α4 and α5 can be calculated according to the probability of the standard normal distribution, and are 0.055, 0.245, 0.4, 0.245 and 0.055 respectively; After M is established, the data D collected each time is multiplied by the coefficient M to obtain the calibrated data, and the calibrated data is stored in the database

[0066] As Figure 4 shown, a function relationship changing with time is established for the calibrated data of each device. Vibration signal data of devices such as engines and generators within half a year is taken out from the database (vibration signals are used in the example, and other signal analyses can be selected according to the analysis needs in actual situations), and function relationships changing continuously with time are established respectively using the functional data analysis algorithm. For each signal where: φ n (t), n ∈ {1, 2,..., N}, c is the coefficient of the basis function is the fitting function curve, the type of the basis function is the B-spline basis, and the number of the basis functions is set to 4th order. After the function relationships of each signal are established, the change trends of the health states of each device of the special vehicle can be characterized, achieving the purpose of implementing health management for each device. Finally, the health management results of each device and the data such as the original parameters and states (other format data collected is converted into Ethernet data through the data conversion module) are stored in the database on the one hand and sent to the display layer for visual display on the other hand

[0067] Overall description of the embodiment

[0068] The devices in the device layer transmit data such as rotational speed and vibration generated by them to the data bus; the data acquisition device collects rotational speed, vibration, etc. from the data bus at a period of 5 ms and then sends them to the data processing and analysis device; the data processing and analysis device first corrects the device data using a method based on probability distribution data correction and stores it in the database, thereby reducing the data fluctuations caused by the strong vibration of the special vehicle; then the data processing and analysis module retrieves the vibration signal data of devices such as the engine and generator within half a year from the database, and respectively establishes the functional relationship of their continuous change over time using the functional data analysis algorithm; finally, the health management results of each device, as well as the original parameters, status and other data, are stored in the database on the one hand, and sent to the display layer for visual display on the other hand.

[0069] The present invention provides a device for the health management of special vehicles, and the device includes:

[0070] Acquisition module: configured to collect data of corresponding components of various sensors configured by the special vehicle itself by various sensors; send the collected data to the data bus; the data acquisition device obtains the collected data of corresponding components of various sensors from the data bus at a period and stores it in the data processing and analysis device, and the collected data stored in the data processing and analysis device follows a normal distribution N(μ,σ 2 ), σ is the standard deviation, and μ is the expectation;

[0071] Correction module: configured to sequentially select the collected data with probability values of μ - 2σ, μ - σ, μ, μ + σ, and μ + 2σ in the normal distribution, construct a representative matrix for the collected data corresponding to each probability value, and obtain the representative matrices D μ-2σ , D μ-σ , D μ , D μ+σ , and D μ+2σ ; generate a standard matrix based on all the data in the data processing and analysis device Determine the correction coefficient matrix for mapping each representative matrix to the standard matrix, and based on each correction coefficient matrix, determine the weighted correction coefficient matrix;

[0072] Analysis module: configured to correct the data stored in the data processing and analysis device based on the weighted correction coefficient matrix, and establish the functional relationship of the corrected data changing over time using the functional data analysis method; determine the damage condition of each component of the special vehicle over time based on the functional relationship.

[0073] A computer-readable storage medium provided by the present invention, in which multiple instructions are stored; the multiple instructions are used to be loaded and executed by a processor to perform the method as described above.

[0074] An electronic device provided by the present invention is characterized in that the electronic device includes:

[0075] A processor for executing multiple instructions;

[0076] A memory for storing multiple instructions;

[0077] Among them, the multiple instructions are used to be stored by the memory and loaded and executed by the processor for the method as described above.

[0078] The above specific embodiments only describe the design principle of the present invention. The shapes and names of the components in this description can be different and are not limited. Therefore, those skilled in the art of the present invention can modify or equivalently replace the technical solutions recorded in the foregoing embodiments; and these modifications and replacements do not depart from the spirit and technical solutions of the present invention, and shall all fall within the protection scope of the present invention.

Claims

1. A method for the health management of special vehicles, characterized in that, Including: Step S1: Various sensors configured on the special vehicle collect data of components corresponding to the various sensors; Send the collected data to the data bus; The data acquisition device obtains the acquisition data of the corresponding components of the various sensors from the data bus at regular intervals and stores it in the data processing and analysis device. The acquisition data stored in the data processing and analysis device follows a normal distribution N(μ,σ 2 ), where σ is the standard deviation and μ is the expectation; Step S2: Sequentially select the acquisition data with probability values of μ - 2σ, μ - σ, μ, μ + σ, and μ + 2σ in the normal distribution, construct a representative matrix for the acquisition data corresponding to each probability value, and obtain the representative matrices D μ-2σ , D μ-σ , D μ , D μ+σ and D μ+2σ ; Generate a standard matrix based on all the data in the data processing and analysis device Determine the correction coefficient matrix for the mapping of each representative matrix to the standard matrix, and determine the weighted correction coefficient matrix based on each correction coefficient matrix; Step S3: Correct the data stored in the data processing and analysis device based on the weighted correction coefficient matrix, and establish a function relationship of the corrected data changing with time using the functional data analysis method; Determine the damage conditions of each component of the special vehicle over time based on the function relationship; In step S2, the correction coefficient matrices for mapping each representative matrix to the standard matrix are solved by the partial least squares algorithm, and five correction coefficient matrices S μ-2σ , S μ-σ , S μ , S μ+σ and S μ+2σ ; Generating a standard matrix based on all the data in the data processing and analysis device Determining a correction coefficient matrix for mapping each representative matrix to the standard matrix, and determining a weighted correction coefficient matrix based on each correction coefficient matrix, including: Classify all the data in the data processing and analysis device by component type, calculate the mean value of each data of the same type of components, and use it as the element in the standard matrix to generate the standard matrix Establish the functional relationships between each representative matrix and the standard matrix respectively to form five functions, and solve the five functions to obtain five correction coefficient matrices Determining the weighted correction coefficient matrix based on each correction coefficient matrix includes: The weighted correction coefficient matrix M corresponding to the standard matrix is obtained by weighting the correction coefficient matrices corresponding to the five representative matrices, M = α1S μ-2σ + α2S μ-σ + α3S μ + α4S μ+σ + α5S μ+2σ , where the values of α1, α2, α3, α4, and α5 are calculated according to the probabilities of the standard normal distribution.

2. The method according to claim 1, wherein The calculation methods of α1, α2, α3, α4, and α5 are: According to the normal distribution function corresponding to N(μ,σ 2 ), calculate the probabilities corresponding to μ - 2σ, μ - σ, μ, μ + σ, and μ + 2σ with probability values respectively, denoted as P μ-2σ , P μ-σ , P μ , P μ+σ , P μ+2σ , where:

3. The method according to claim 2, characterized in that The step S3 includes: Obtain the weighted correction coefficient matrix M corresponding to the standard matrix, multiply each piece of data stored in the data processing and analysis device by the coefficient M to obtain the corrected data; Store the corrected data in the database, select the corrected data within a preset time range from the database, and establish a function relationship of the corrected data within the preset time range changing with time using the functional data analysis method; Determine the damage conditions of each component of the special vehicle over time based on the function relationship.

4. The method according to any one of claims 1 to 3, characterized in that, The method for the health management of the special vehicle further includes: Step S4: Send the damage conditions of each component of the special vehicle over time to the display layer for visual display.

5. An apparatus for the health management of special vehicles, which is used to execute the method described in any one of claims 1-4, characterized in that, The device includes: Acquisition module: configured to collect data of corresponding components of various sensors configured for the special vehicle itself; send the collected data to the data bus; the data acquisition device periodically obtains the collected data of the corresponding components of various sensors from the data bus and stores it in the data processing and analysis device, and the collected data stored in the data processing and analysis device follows a normal distribution N(μ,σ 2 ), where σ is the standard deviation and μ is the expectation; Calibration module: Configured to sequentially select the acquisition data with probability values of μ - 2σ, μ - σ, μ, μ + σ, and μ + 2σ in the normal distribution, construct a representative matrix for the acquisition data corresponding to each probability value, and obtain the representative matrices D μ-2σ , D μ-σ , D μ , D μ+σ and D μ+2σ ; Generate a standard matrix based on all the data in the data processing and analysis device Determine the calibration coefficient matrix for the mapping of each representative matrix to the standard matrix, and based on each calibration coefficient matrix, determine the weighted calibration coefficient matrix; Analysis module: Configured to correct the data stored in the data processing and analysis device based on the weighted correction coefficient matrix, and establish a function relationship of the corrected data changing with time using the functional data analysis method; Determine the damage conditions of each component of the special vehicle over time based on the function relationship; Among them, the correction coefficient matrices for mapping each representative matrix to the standard matrix are solved by the partial least squares algorithm, and five correction coefficient matrices S μ-2σ , S μ-σ , S μ , S μ+σ and S μ+2σ are obtained; Generate a standard matrix based on all the data in the data processing and analysis device Determine the correction coefficient matrix for mapping each representative matrix to the standard matrix, and based on each correction coefficient matrix, determine the weighted correction coefficient matrix, including: Classify all the data in the data processing and analysis device by component type, calculate the mean value of each data of the same type of components, and use it as the element in the standard matrix to generate the standard matrix Establish the functional relationships between each representative matrix and the standard matrix respectively to form five functions, and solve the five functions to obtain five correction coefficient matrices Determining the weighted correction coefficient matrix based on each correction coefficient matrix includes: The weighted correction coefficient matrix M corresponding to the standard matrix is obtained by weighting the correction coefficient matrices corresponding to the five representative matrices, M = α1S μ-2σ + α2S μ-σ + α3S μ + α4S μ+σ + α5S μ+2σ , where the values of α1, α2, α3, α4 and α5 are calculated according to the probabilities of the standard normal distribution.

6. A computer-readable storage medium, in which multiple instructions are stored; the multiple instructions are used to be loaded and executed by a processor to perform the method according to any one of claims 1-4.

7. An electronic device, characterized in that, The electronic device includes: A processor for executing multiple instructions; A memory for storing multiple instructions; Wherein, the multiple instructions are used to be stored by the memory and loaded and executed by the processor to perform the method according to any one of claims 1-4.

Images