A chemical pump vibration detection system

Abstract

The present invention provides a chemical pump vibration detection system, which relates to the technical field of vibration detection, and comprises: an analysis part and a detection part: the analysis part is used to realize the functions of receiving, storing and analyzing signals; the detection part is used for the sampling, storage and communication functions of the system; the analysis part comprises a signal processing unit; the detection part comprises a detection frame, a vibration sensor, a chemical pump body, a signal acquisition unit and a storage and communication unit; the vibration sensor is connected to the chemical pump body; the signal acquisition unit is used to obtain the vibration signal of the chemical pump body through the vibration sensor; the storage and communication unit is connected to the signal acquisition unit; the signal processing unit is connected to the storage and communication unit; the detection system is simultaneously arranged in three mutually perpendicular directions of the chemical pump body, namely, axial, radial vertical and radial horizontal.

Description

Technical Field

[0001] The invention relates to the technical field of vibration detection, in particular to a chemical pump vibration detection system. Background Art

[0002] Chemical pumps are widely used in petroleum, chemical, metallurgy, synthetic fiber, pharmaceutical, food, synthetic fiber and other departments to transport alkaline corrosive media. Chemical pumps mainly include: IH stainless steel chemical pump, IHF fluoroplastic chemical pump, FSB fluoroplastic chemical pump, CQB chemical magnetic pump, CQ chemical magnetic pump, FZB chemical self-priming pump, etc. Vibration refers to the periodic deformation and displacement of an object during movement. For the pump body, the important feature of the pump failure is that the machine is accompanied by abnormal vibration and noise, and its vibration signal can reflect the pump body failure in real time.

[0003] In modern chemical production, chemical pumps are one of the key equipment and are widely used to transport liquid media. However, during long-term operation, chemical pumps may experience abnormal vibrations due to mechanical vibrations, changes in medium pressure, equipment aging and other factors. These vibration abnormalities will not only affect the normal operation of the equipment, but may also cause equipment failures and even safety accidents. Therefore, it is of great significance to effectively detect and monitor the vibration of chemical pumps.

[0004] Existing chemical pump vibration detection systems usually rely on a single vibration sensor to determine the operating status of the chemical pump by monitoring the vibration signal of the chemical pump. However, single-direction vibration sensors have obvious disadvantages in practical applications. First, single-direction vibration monitoring cannot fully reflect the vibration status of the chemical pump, and it is easy to have detection blind spots; the vibration sensor system that relies on a single direction also has deficiencies in detection accuracy and reliability, and it is difficult to meet the detection needs of high precision and high reliability. Summary of the invention

[0005] In order to solve the technical problems that vibration monitoring in a single direction cannot fully reflect the vibration state of a chemical pump and is prone to detection blind spots; and that the vibration sensor system relying on a single direction also has deficiencies in detection accuracy and reliability and is difficult to meet the detection requirements of high accuracy and high reliability, the present invention provides a chemical pump vibration detection system.

[0006] The technical solution provided by the embodiment of the present invention is as follows:

[0007] First aspect:

[0008] A chemical pump vibration detection system provided by an embodiment of the present invention is characterized by comprising an analysis unit and a detection unit:

[0009] The analysis unit is used to realize the functions of receiving, storing and analyzing signals;

[0010] The detection unit is used for sampling, storage, and communication functions of the system;

[0011] The analysis unit includes a signal processing unit;

[0012] The detection unit includes a detection frame, a vibration sensor, a chemical pump body, a signal acquisition unit and a storage and communication unit;

[0013] The vibration sensor is connected to the chemical pump body;

[0014] The signal acquisition unit is used to obtain the vibration signal of the chemical pump body through the vibration sensor;

[0015] The storage communication unit is connected to the signal acquisition unit;

[0016] The signal processing unit is connected to the storage communication unit;

[0017] The detection system is simultaneously arranged in three mutually perpendicular directions: axial, radial vertical and radial horizontal of the chemical pump body;

[0018] Three groups of sliding rods are slidably arranged on the detection frame, one ends of the three groups of sliding rods are connected to a fixed plate, the fixed plate is connected to the inner side of the detection frame by a spring, a clamping sleeve is arranged on the fixed plate, three groups of vibration sensors are arranged, the three groups of vibration sensors are respectively placed in the three groups of clamping sleeves and fastened to the clamping sleeves, the chemical pump body includes a rotating mechanical base, and the detection frame is installed on the rotating mechanical base.

[0019] Furthermore, the three groups of vibration sensors are respectively located at the top and two side positions of the detection frame, and the vibration sensors located at the two side positions are symmetrically arranged with respect to the vibration sensor at the top position.

[0020] Furthermore, a magnetic seat is provided at one end of the vibration sensor;

[0021] The magnetic seat and the vibration sensor are threadedly connected;

[0022] The vibration sensor comprises a cable, and the cable of the vibration sensor is connected to a vibration measuring instrument.

[0023] Furthermore, the ferrule is fixedly connected to the fixed disk via two groups of side plates, and the two groups of side plates are respectively located on both sides of the ferrule.

[0024] Furthermore, an elastic anti-skid lining is provided on the inner side of the ferrule, and the elastic anti-skid lining is bonded and fixed to the ferrule;

[0025] The installation direction of the vibration sensor is perpendicular to the axis direction of the rotating mechanical base.

[0026] Furthermore, the analysis unit also includes a remote monitoring unit, which is used to send the detection data to a remote server through a wireless communication module for real-time monitoring and analysis by a user.

[0027] Furthermore, the detection frame adopts an adjustable structure to adapt to the chemical pump bodies of different sizes and shapes.

[0028] Furthermore, the signal processing unit is specifically used for:

[0029] S1: Acquire the vibration signal sent by the storage communication unit;

[0030] S2: extracting the signal features of the vibration signal based on discrete wavelet transform and energy spectrum analysis;

[0031] S3: According to the signal characteristics of the vibration signal, a chemical pump fault detection is performed through a multi-layer perceptron network.

[0032] Furthermore, the S2 specifically includes:

[0033] S201: Determine the autospectral signal of the vibration signal:

[0034] (1)

[0035] in, P xx represents the autospectral signal, f Indicates frequency, X Indicates vibration signal, X ( f ) represents the frequency domain vibration signal, X * express X The complex conjugate of

[0036] S202: Divide the autospectral signal by the sampling frequency to obtain the power spectrum density of the vibration signal:

[0037] (2)

[0038] in, PSD represents the power spectral density, f s Indicates the sampling frequency;

[0039] S203: Extracting features of the vibration signal based on discrete wavelet transform:

[0040] (3)

[0041] in, DWT represents discrete wavelet transform, j represents the scale parameter, k represents the translation parameter, X ( t ) represents the time domain vibration signal, t Indicates time, represents the wavelet basis function, represents the complex conjugate of the wavelet basis function, d represents the differential operator;

[0042] S204: Decompose the vibration signal into a low-frequency signal and a high-frequency signal through two complementary filters, and express the vibration signal as:

[0043] (4)

[0044] in, A j Indicates j The approximate value of the low-frequency signal, D j Indicates j The detail value of the high-frequency signal, J Indicates the total number of levels of signal decomposition;

[0045] S205: Based on Parseval's energy theorem, an energy spectrum analysis is performed on the vibration signal:

[0046] (5)

[0047] in, N Indicates the total number of sampling points of the signal;

[0048] S206: By substituting formula (4) into formula (5), the low-frequency energy and high-frequency energy of the vibration signal are extracted:

[0049] (6)

[0050] in, Represents low-frequency energy, Represents high frequency energy.

[0051] Furthermore, the S3 specifically includes:

[0052] S301: In the input layer of the multilayer perceptron network, input power spectrum density features, time domain features, wavelet transform low-frequency signal features, wavelet transform high-frequency signal features, low-frequency energy, and high-frequency energy;

[0053] S302: performing nonlinear transformation in the hidden layer of the multilayer perceptron network through an activation function to extract hidden features of the vibration signal;

[0054] The hidden layer includes a first hidden layer, a second hidden layer and a third hidden layer; the first hidden layer and the second hidden layer use a ReLU activation function, and the third hidden layer uses a Sigmoid activation function;

[0055] S303: In the output layer of the multi-layer perceptron network, a fault detection result is output according to the hidden features of the vibration signal:

[0056] (7)

[0057] in, σ represents the fault detection value, and determines whether the fault detection value is greater than the preset detection value. If so, it is determined that the chemical pump has a fault. Otherwise, it is determined that the chemical pump does not have a fault. Sigmoid represents the Sigmoid activation function. w o represents the output layer weight, h represents hidden features, b o Represents the output layer bias.

[0058] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0059] (1) In the present invention, vibration sensors are simultaneously arranged in three mutually perpendicular directions, namely, axial, radial vertical and radial horizontal directions of the chemical pump body, to ensure that vibration monitoring covers all key directions in the operation of the chemical pump, thereby solving the blind spot problem caused by single-direction monitoring.

[0060] (2) In the present invention, the detection accuracy and reliability of the system are improved through multi-directional and multi-sensor collaborative detection. The system no longer relies on a sensor in a single direction, but conducts comprehensive analysis through vibration data in multiple directions, thereby improving the accuracy and reliability of detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0062] Figure 1 A logic block diagram of a chemical pump vibration detection system provided by an embodiment of the present invention;

[0063] Figure 2 A logic block diagram of a detection unit provided by an embodiment of the present invention;

[0064] Figure 3 A schematic diagram of the structure of a detection frame provided by an embodiment of the present invention;

[0065] Figure 4 A schematic diagram of the assembly of a vibration sensor and a ferrule provided in an embodiment of the present invention;

[0066] Figure 5 A schematic diagram of installing a vibration sensor provided by an embodiment of the present invention;

[0067] Figure 6 A schematic diagram of use provided by an embodiment of the present invention;

[0068] Figure 7 This is a partial enlarged view of area A provided in an embodiment of the present invention.

[0069] Reference numerals:

[0070] 1. Detection frame; 11. Through slot; 2. Slide rod; 3. Fixed plate; 4. Spring; 5. Card sleeve; 51. Side plate; 52. Elastic anti-slip lining; 6. Vibration sensor; 61. Magnetic seat; 62. Cable; 7. Chemical pump body; 71. Rotating machinery base. DETAILED DESCRIPTION

[0071] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0072] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "example" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of the word "example" is intended to present the concept in a specific way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either of the two.

[0073] In the embodiments of the present invention, "image" and "picture" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same. "of", "corresponding, relevant" and "corresponding" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same.

[0074] In the embodiments of the present invention, sometimes the subscripts such as W 1 It may be mistakenly written as a non-subscript form such as W1. When the difference is not emphasized, the meanings they express are the same.

[0075] In order to make the technical problems, technical solutions and advantages to be solved by the present invention more clear, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0076] The embodiment of the present invention provides a chemical pump vibration detection system, including an analysis unit and a detection unit: the analysis unit is used to realize the signal receiving, storage and analysis functions. The detection unit is used for the sampling, storage and communication functions of the system. The analysis unit includes a signal processing unit. The detection unit includes a detection frame, a vibration sensor, a chemical pump body, a signal acquisition unit and a storage and communication unit. The vibration sensor is connected to the chemical pump body. The signal acquisition unit is used to obtain the vibration signal of the chemical pump body through the vibration sensor. The storage and communication unit is connected to the signal acquisition unit. The signal processing unit is connected to the storage and communication unit. The detection system is simultaneously arranged in three mutually perpendicular directions of the chemical pump body, namely, axial, radial vertical and radial horizontal.

[0077] In the present invention, by arranging vibration sensors in three mutually perpendicular directions, namely, axial, radial vertical and radial horizontal, the vibration conditions of the chemical pump in multiple directions can be fully monitored, thus avoiding the blind spot problem caused by single-direction monitoring.

[0078] Three groups of slide bars 2 are slidably arranged on the detection frame 1, and one ends of the three groups of slide bars 2 are connected to the fixed plate 3. The fixed plate 3 is connected to the inner side of the detection frame 1 through a spring 4. A clamping sleeve 5 is arranged on the fixed plate 3. Three groups of vibration sensors 6 are arranged. The three groups of vibration sensors 6 are respectively placed in the three groups of clamping sleeves 5 and fastened to the clamping sleeves 5. The chemical pump body 7 includes a rotating mechanical base 71, and the detection frame 1 is installed on the rotating mechanical base 71.

[0079] It should be noted that the core principle of the chemical pump vibration detection system is to detect the vibration signal generated by the chemical pump during operation through a vibration sensor. The vibration sensor is installed in different positions of the chemical pump body, and the sliding adjustment of the detection frame ensures that the sensor can accurately collect vibration data in the axial, radial vertical and radial horizontal directions. The vibration signal obtained by the signal acquisition unit is transmitted to the signal processing unit through the storage communication unit, and the signal processing unit analyzes and stores the received vibration signal. The spring and ferrule design on the detection frame ensures close contact between the vibration sensor and the chemical pump body, thereby improving the accuracy and stability of data acquisition.

[0080] In a possible implementation, three groups of vibration sensors 6 are respectively located at the top and two side positions of the detection frame 1 , and the vibration sensors 6 at the two side positions are symmetrically arranged with respect to the vibration sensor 6 at the top position.

[0081] Among them, three groups of vibration sensors 6 can be installed on the detection frame 1 at one time, and comprehensive vibration signals can be obtained by increasing detection points, and comprehensive measurement data can be output, so as to achieve comprehensive vibration detection of the chemical pump body 7.

[0082] It should be noted that the detection frame 1 is a semicircular arc structure, and the shape of the detection frame 1 adapts to the outer contour of the rotating mechanical base 71. The rotating mechanical base 71 has built-in mechanical mechanisms such as an output shaft and a coupling. The rotating mechanical base 71 is selected as the vibration detection position of the chemical pump body 7, and the rotating mechanical base 71 is used as the main measuring point.

[0083] In a possible implementation, a magnetic seat 61 is disposed at one end of the vibration sensor 6. The magnetic seat 61 is threadedly connected to the vibration sensor 6. The vibration sensor 6 includes a cable 62, and the cable 62 of the vibration sensor 6 is connected to the vibration measuring instrument.

[0084] It should be noted that the three groups of vibration sensors 6 are evenly spaced to perform multi-point detection on the rotating mechanical base 71 to obtain comprehensive and accurate vibration data. By installing the three groups of vibration sensors at the top and both sides of the detection frame, all-round monitoring of the vibration signal of the chemical pump can be achieved.

[0085] In the present invention, the vibration sensors on both sides are symmetrically arranged with respect to the top vibration sensor, ensuring that accurate vibration data can be obtained in three mutually perpendicular directions. This layout can fully capture various vibration signals generated by the chemical pump during operation, ensuring the comprehensiveness and accuracy of the data. At the same time, through the reasonable layout of the vibration sensors, all-round monitoring of the vibration signals of the chemical pump is achieved, ensuring the comprehensiveness and accuracy of the detection data.

[0086] Furthermore, the symmetrically arranged vibration sensors can effectively avoid misjudgment caused by data deviation in a single direction, thus improving the monitoring accuracy and reliability of the system. At the same time, the sensor layout at the top and both sides can capture the vibration characteristics of the chemical pump in different directions, providing more comprehensive data support for fault diagnosis.

[0087] Optionally, sensors are added to the bottom of the detection frame to obtain vibration data in more directions.

[0088] In a possible implementation manner, the ferrule 5 and the fixed disk 3 are fixedly connected via two sets of side plates 51 , and the two sets of side plates 51 are respectively located on two sides of the ferrule 5 .

[0089] In a possible implementation, an elastic anti-skid lining 52 is provided on the inner side of the ferrule 5, and the elastic anti-skid lining 52 is bonded and fixed to the ferrule 5. A through slot 11 is provided on the detection frame 1. The installation direction of the vibration sensor 6 is perpendicular to the axial direction of the rotating mechanical base 71.

[0090] It should be noted that, when in use, three groups of vibration sensors 6 are installed on the detection frame 1, and then the detection frame 1 is installed on the rotating mechanical base 71 of the chemical pump body 7. The spring 4 is compressed under the action of external force, and the slide bar 2 slides, so that the three groups of vibration sensors 6 can be tightly attached to the outer surface of the rotating mechanical base 71, thereby realizing the rapid vertical installation of the three groups of vibration sensors 6. The three groups of vibration sensors 6 are respectively located at the top and both sides of the rotating mechanical base 71. The cables 62 of the vibration sensors 6 are connected to the vibration measuring instrument and the data recorder. The three groups of vibration sensors 6 are evenly spaced, and multi-point detection is performed on the rotating mechanical base 71 to obtain comprehensive and accurate vibration data. The vibration measuring instrument is used to receive the vibration sensor 6 signal and convert it into readable data, including vibration frequency, amplitude and phase, etc. The data recorder is used to record and store vibration data.

[0091] In a possible implementation, after the detection system is initialized, if the signal acquisition unit does not receive a command from the analysis unit, the detection system enters the signal acquisition operation by default. When entering the acquisition operation, one minute of data is first collected, and the amplitude change is analyzed to determine whether the unit is in operation. When the status is displayed as operation, the system starts sampling, otherwise it enters the sleep state.

[0092] Each group of collected three-dimensional vibration signals is stored in sequence, and the corresponding sampling time and the last storage position of each time are recorded. Before each use of the system, the analysis unit sends a time calibration command to the detection unit to calibrate the time and update the current time. When the analysis unit sends a communication instruction to the detection unit, communication starts to realize the data transmission function. After each transmission, the analysis unit sends a clear instruction to clear the detection unit storage for the next system sampling and storage.

[0093] In the present invention, the system performs sampling after detecting the operating status of the chemical pump, thereby avoiding unnecessary energy consumption, and ensuring the accuracy and continuity of the data through the time calibration and data transmission functions. At the same time, the setting of the zeroing instruction enables the system to be reinitialized before each sampling, thereby ensuring the effective use of storage space and the convenience of data management.

[0094] In a possible implementation, the analysis unit further includes a remote monitoring unit, which is used to send the detection data to a remote server via the wireless communication module for real-time monitoring and analysis by a user.

[0095] It should be noted that the analysis department has added a remote monitoring unit, which sends the detection data to the remote server through the wireless communication module. Users can monitor and analyze the vibration data of the chemical pump in real time through the remote server, and promptly discover and deal with potential faults. The remote monitoring unit can transmit data to the user's monitoring terminal through the Internet to realize the functions of remote monitoring and management.

[0096] Optionally, users can choose different wireless communication technologies according to actual needs to adapt to different usage environments and data transmission requirements. The functions of the remote monitoring unit can also be expanded, such as adding data storage and backup functions to improve data security and reliability. In addition, a variety of monitoring terminals can be designed, such as PCs, tablets, mobile phones, etc., to meet the monitoring needs and usage habits of different users.

[0097] In the present invention, the introduction of the remote monitoring unit enables the user to monitor the operating status of the chemical pump anytime and anywhere, thereby improving the management efficiency and fault handling capability of the system. At the same time, data is transmitted through the wireless communication module, which reduces the complexity and maintenance cost of the wired connection and enhances the flexibility and convenience of the system.

[0098] Furthermore, the real-time monitoring and analysis functions enable users to detect abnormal situations in a timely manner and take appropriate measures to ensure the safe operation of chemical pumps.

[0099] In a possible implementation, the vibration sensor is an accelerometer, which can perform accurate measurement in a high frequency range.

[0100] It should be noted that the accelerometer measures the acceleration changes generated by the chemical pump during operation, obtains the vibration signal, and transmits the signal to the signal processing unit for analysis and processing. The use of accelerometers as vibration sensors improves the measurement accuracy and sensitivity of the system in the high-frequency range, and can accurately capture the tiny vibration changes generated by the chemical pump during operation, which helps to detect potential faults and abnormalities at an early stage. At the same time, the high-frequency measurement capability of the accelerometer enables the system to maintain high detection accuracy under various working conditions, enhancing the reliability and adaptability of the system.

[0101] In a possible implementation manner, the detection unit further includes a sleep unit; the sleep unit is used to reduce the power consumption of the system.

[0102] In a possible implementation, the detection frame 1 adopts an adjustable structure to adapt to chemical pump bodies 7 of different sizes and shapes.

[0103] In the present invention, the structure of the detection frame adopts an adjustable structure, which can adapt to chemical pump bodies of different sizes and shapes, thereby enhancing the versatility of the system.

[0104] In a possible implementation manner, the signal processing unit is specifically configured to:

[0105] S1: Acquire the vibration signal sent by the storage communication unit.

[0106] Optionally, after acquiring the vibration signal, the time domain characteristics of the vibration signal are extracted by calculating the root mean square value, variance or signal waveform factor.

[0107] S2: Extract the signal features of the vibration signal based on discrete wavelet transform and energy spectrum analysis.

[0108] Among them, Discrete Wavelet Transform (DWT) is a mathematical transformation used for signal analysis, especially suitable for processing non-stationary signals. It realizes the time-frequency analysis of signals by decomposing them into different scales (frequencies) and time domain positions.

[0109] Among them, Energy Spectral Analysis is a technology used to analyze the energy distribution of signals, mainly used to study the energy distribution of signals in different frequency components. Energy spectrum analysis is widely used in signal processing, vibration analysis, fault diagnosis and other fields to determine in which frequency range the main energy of the signal is concentrated, thereby helping to identify signal characteristics or equipment status.

[0110] In a possible implementation, S2 specifically includes:

[0111] S201: Determine the autospectral signal of the vibration signal:

[0112] (1)

[0113] in, P xx represents the autospectral signal, f Indicates frequency, X Indicates vibration signal, X ( f ) represents the frequency domain vibration signal, X * express X The complex conjugate of .

[0114] Among them, Auto-Spectral Signal (Auto-Spectrum) is a frequency domain analysis method in signal processing, which is used to describe the frequency components of the signal itself and its energy distribution. It is obtained by calculating the power spectral density (PSD) or auto-power spectrum after Fourier transform of the signal, and is used to analyze the energy intensity of the signal at different frequencies.

[0115] S202: Divide the autospectral signal by the sampling frequency to obtain the power spectrum density of the vibration signal:

[0116] (2)

[0117] in, PSD represents the power spectral density, f s Indicates the sampling frequency.

[0118] Among them, Power Spectral Density (PSD) is an important concept in signal processing, which is used to describe the power distribution of a signal in the frequency domain. Specifically, power spectral density represents the power contained in a signal per unit frequency. It reveals the distribution of the signal's energy at different frequencies and helps analyze the frequency characteristics of the signal.

[0119] S203: Extract features of the vibration signal based on discrete wavelet transform:

[0120] (3)

[0121] in, DWT represents discrete wavelet transform, j represents the scale parameter, k represents the translation parameter, X ( t ) represents the time domain vibration signal, t Indicates time, represents the wavelet basis function, represents the complex conjugate of the wavelet basis function, d Represents the differentiation operator.

[0122] S204: Decompose the vibration signal into a low-frequency signal and a high-frequency signal through two complementary filters, and express the vibration signal as:

[0123] (4)

[0124] in, A j Indicates j The approximate value of the low-frequency signal, D j Indicates j The detail value of the high-frequency signal, J Represents the total number of levels of signal decomposition.

[0125] S205: Based on Parseval's energy theorem, perform energy spectrum analysis on vibration signals:

[0126] (5)

[0127] in, N Indicates the total number of sampling points of the signal.

[0128] Among them, Parseval's Theorem, also known as Parseval's Theorem, is an important theorem in signal processing and Fourier analysis. It reveals the energy conservation relationship between the time domain and the frequency domain, that is, the energy of the signal in the time domain is equal to its energy in the frequency domain.

[0129] S206: By substituting formula (4) into formula (5), the low-frequency energy and high-frequency energy of the vibration signal are extracted:

[0130] (6)

[0131] in, Represents low-frequency energy, Represents high frequency energy.

[0132] In the present invention, discrete wavelet transform can perform time-frequency domain analysis on non-stationary signals, analyze signal changes from multiple scales, and accurately identify early fault signs. At the same time, through the scale parameters and translation parameters in wavelet transform, the signal can be decomposed into low-frequency approximations and high-frequency details of different frequency components, which improves the comprehensiveness of signal analysis.

[0133] Furthermore, by calculating the autospectral signal and power spectrum density of the vibration signal, the frequency information of the signal can be quickly extracted, and the vibration signal can be processed in real time or near real time to meet the needs of online monitoring and fault diagnosis of industrial equipment. At the same time, through the detailed analysis of low-frequency and high-frequency energy, combined with the time-frequency characteristics of the vibration signal, the system can identify early and relatively weak fault signals. The extraction and comparison of multiple signal features can help detect anomalies in the early stages of fault evolution and avoid major equipment damage.

[0134] In summary, by combining discrete wavelet transform and energy spectrum analysis, efficient and comprehensive feature extraction of vibration signals can be achieved, thereby improving the accuracy and reliability of equipment vibration monitoring and fault diagnosis. This method is not only suitable for complex non-stationary signals, but also can help identify early faults, provide more accurate energy distribution analysis, and improve the overall detection performance of the system.

[0135] S3: According to the signal characteristics of the vibration signal, the chemical pump fault detection is performed through a multi-layer perceptron network.

[0136] Among them, the Multilayer Perceptron (MLP) network is a classic feedforward artificial neural network, which is usually used for machine learning tasks such as classification and regression. It is one of the basic structures of neural networks. It can handle complex nonlinear problems by calculating through multiple layers of neurons (also called perceptrons).

[0137] In a possible implementation, S3 specifically includes:

[0138] S301: In the input layer of the multi-layer perceptron network, power spectrum density features, time domain features, wavelet transform low-frequency signal features, wavelet transform high-frequency signal features, low-frequency energy, and high-frequency energy are input.

[0139] The time domain features are extracted from the original vibration signal, and specifically may be root mean square value, variance or signal waveform factor, etc.

[0140] Among them, the low-frequency signal characteristics of wavelet transform refer to the low-frequency components (also called approximation coefficients) decomposed in the wavelet transform process, which mainly contain the overall outline or trend information of the signal. The approximation coefficients of wavelet transform can be further processed by subsequent feature extraction methods, such as calculating statistical characteristics such as mean, variance, skewness, kurtosis, etc., to represent the characteristics of low-frequency signals.

[0141] Among them, the high-frequency signal characteristics of wavelet transform refer to the high-frequency components (also called detail coefficients) decomposed in the wavelet transform process, which mainly contain the rapid changes or detail information of the signal. The detail coefficients of wavelet transform can be further processed by subsequent feature extraction methods, such as calculating statistical features such as mean, variance, skewness, kurtosis, etc., to represent the characteristics of high-frequency signals.

[0142] In the present invention, at the input layer, the system simultaneously inputs multiple signal features into the network, which cover the time domain and frequency domain information of the vibration signal, provide a comprehensive description of the signal, and help to accurately distinguish between normal and faulty states of the equipment.

[0143] S302: In the hidden layer of the multi-layer perceptron network, a nonlinear transformation is performed through an activation function to extract hidden features of the vibration signal.

[0144] The hidden layer includes a first hidden layer, a second hidden layer and a third hidden layer. The first hidden layer and the second hidden layer use the ReLU activation function, and the third hidden layer uses the Sigmoid activation function.

[0145] In the present invention, by using ReLU and Sigmoid activation functions in the hidden layer, the network can perform nonlinear mapping on the input multi-dimensional features, extract complex hidden features, and improve the accuracy of detection.

[0146] S303: In the output layer of the multi-layer perceptron network, the fault detection result is output according to the hidden features of the vibration signal:

[0147] (7)

[0148] in, σ Represents the fault detection value, and determines whether the fault detection value is greater than the preset detection value. If so, it is determined that the chemical pump has a fault. Otherwise, it is determined that the chemical pump does not have a fault. Sigmoid represents the Sigmoid activation function. w o represents the output layer weight, h represents hidden features, b o Represents the output layer bias.

[0149] In the present invention, the multi-layer perceptron network has fast forward propagation computing capability, especially the computing of the output layer is very efficient, and can quickly output the fault detection result after the vibration signal is input. By judging the fault detection value in real time, the system can immediately respond to the detected abnormal situation. At the same time, by setting the preset detection value, the judgment standard can be adjusted according to different working conditions and application requirements, so as to meet different detection accuracy and fault tolerance requirements.

[0150] It should be noted that those skilled in the art can set the size of the preset detection value according to actual needs, and the present invention is not limited here.

[0151] Optionally, the training of the multilayer perceptron network adopts the back propagation algorithm, and the weights and biases between neurons in each layer are updated as shown in formulas (8) and (9):

[0152] (8)

[0153] (9)

[0154] in, w represents the weight, b Indicates bias, u represents the number of iterations, Indicates L Layer i neurons and L +1st floor j The weights between neurons, α u Indicates uThe learning rate at the iteration, represents the partial derivative, J ( ) represents the loss function, Indicates L Tier i The bias of a neuron, λ represents the regularization coefficient.

[0155] Among them, the backpropagation algorithm is a key algorithm in neural network training, which is used to calculate the gradient of the neural network and update its weights. The algorithm adjusts the weights and biases in the network layer by layer by means of error backpropagation, so that the predicted value of the network output gradually approaches the true value.

[0156] Optionally, the loss function may adopt a mean square error loss function.

[0157] Optionally, the learning rate is specified as:

[0158] (10)

[0159] in, α min represents the minimum learning rate, α max represents the maximum learning rate, U max represents the maximum number of iterations, and cos represents the cosine function.

[0160] In the present invention, as the training progresses, the learning rate gradually decreases from the maximum value to the minimum value. This means that in the early stages of training, the learning rate is large, which helps to quickly explore the solution space and avoid falling into the local optimum; in the later stages of training, the learning rate gradually decreases, which helps to refine the parameter adjustment and reduce the risk of overfitting.

[0161] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0162] (1) In the present invention, vibration sensors are simultaneously arranged in three mutually perpendicular directions, namely, axial, radial vertical and radial horizontal directions of the chemical pump body, to ensure that vibration monitoring covers all key directions in the operation of the chemical pump, thereby solving the blind spot problem caused by single-direction monitoring.

[0163] (2) In the present invention, the detection accuracy and reliability of the system are improved through multi-directional and multi-sensor collaborative detection. The system no longer relies on a sensor in a single direction, but conducts comprehensive analysis through vibration data in multiple directions, thereby improving the accuracy and reliability of detection.

[0164] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

[0165] There are a few points to note:

[0166] (1) The drawings of the embodiments of the present invention only involve structures related to the embodiments of the present invention. Other structures may refer to conventional designs.

[0167] (2) For the sake of clarity, in the drawings used to describe the embodiments of the present invention, the thickness of layers or regions is exaggerated or reduced, that is, these drawings are not drawn according to the actual scale. It is understood that when an element such as a layer, film, region or substrate is referred to as being "on" or "under" another element, the element may be "directly" "on" or "under" the other element or there may be intermediate elements.

[0168] (3) In the absence of conflict, the embodiments of the present invention and the features therein may be combined with each other to obtain new embodiments.

[0169] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. The protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A chemical pump vibration detection system, characterized in that: Including analysis department and detection department: The analysis unit is used to realize the functions of receiving, storing and analyzing signals; The detection unit is used for sampling, storage, and communication functions of the system; The analysis unit includes a signal processing unit; The detection unit includes a detection frame, a vibration sensor, a chemical pump body, a signal acquisition unit and a storage and communication unit; The vibration sensor is connected to the chemical pump body; The signal acquisition unit is used to obtain the vibration signal of the chemical pump body through the vibration sensor; The storage communication unit is connected to the signal acquisition unit; The signal processing unit is connected to the storage communication unit; the signal processing unit is specifically used for: S1: Acquire the vibration signal sent by the storage communication unit; S2: extracting the signal features of the vibration signal based on discrete wavelet transform and energy spectrum analysis; S3: According to the signal characteristics of the vibration signal, the chemical pump is fault detected through a multi-layer perceptron network; The S2 specifically includes: S201: Determine the autospectral signal of the vibration signal: （1） in, P xx represents the autospectral signal, f Indicates frequency, X Indicates vibration signal, X ( f ) represents the frequency domain vibration signal, X * express X The complex conjugate of ; S202: Divide the autospectral signal by the sampling frequency to obtain the power spectrum density of the vibration signal: （2） in, PSD represents the power spectral density, f s Indicates the sampling frequency; S203: Extracting features of the vibration signal based on discrete wavelet transform: （3） in, DWT represents discrete wavelet transform, j represents the scale parameter, k represents the translation parameter, X ( t ) represents the time domain vibration signal, t Indicates time, represents the wavelet basis function, represents the complex conjugate of the wavelet basis function, d represents the differential operator; S204: Decompose the vibration signal into a low-frequency signal and a high-frequency signal through two complementary filters, and express the vibration signal as: （4） in, A j Indicates j The approximate value of the low-frequency signal, D j Indicates j The detail value of the high-frequency signal, J Indicates the total number of levels of signal decomposition; S205: Based on Parseval's energy theorem, an energy spectrum analysis is performed on the vibration signal: （5） in, N Indicates the total number of sampling points of the signal; S206: By substituting formula (4) into formula (5), the low-frequency energy and high-frequency energy of the vibration signal are extracted: （6） in, Represents low-frequency energy, Indicates high frequency energy; The S3 specifically includes: S301: In the input layer of the multilayer perceptron network, input power spectrum density features, time domain features, wavelet transform low-frequency signal features, wavelet transform high-frequency signal features, low-frequency energy, and high-frequency energy; S302: performing nonlinear transformation in the hidden layer of the multi-layer perceptron network through an activation function to extract hidden features of the vibration signal; The hidden layer includes a first hidden layer, a second hidden layer and a third hidden layer; the first hidden layer and the second hidden layer use a ReLU activation function, and the third hidden layer uses a Sigmoid activation function; S303: In the output layer, the fault detection result is output according to the hidden features of the vibration signal: （7） in, σ represents the fault detection value, and determines whether the fault detection value is greater than the preset detection value. If so, it is determined that the chemical pump has a fault. Otherwise, it is determined that the chemical pump does not have a fault. Sigmoid represents the Sigmoid activation function. w o represents the output layer weight, h represents hidden features, b o represents the output layer bias; The detection system is simultaneously arranged in three mutually perpendicular directions: axial, radial vertical and radial horizontal of the chemical pump body; Three groups of slide bars (2) are slidably arranged on the detection frame (1), one end of each of the three groups of slide bars (2) is connected to a fixed plate (3), the fixed plate (3) is connected to the inner side surface of the detection frame (1) via a spring (4), a clamping sleeve (5) is arranged on the fixed plate (3), three groups of vibration sensors (6) are arranged, the three groups of vibration sensors (6) are respectively placed in the three groups of clamping sleeves (5) and are fastened to the clamping sleeves (5), the chemical pump body (7) includes a rotating mechanical base (71), and the detection frame (1) is installed on the rotating mechanical base (71).

2. The chemical pump vibration detection system according to claim 1, characterized in that: The three groups of vibration sensors (6) are respectively located at the top position and two side positions of the detection frame (1), and the vibration sensors (6) located at the two side positions are symmetrically arranged with respect to the vibration sensors (6) at the top position.

3. The chemical pump vibration detection system according to claim 1, characterized in that: A magnetic seat (61) is provided at one end of the vibration sensor (6); The magnetic seat (61) is threadedly connected to the vibration sensor (6); The vibration sensor (6) comprises a cable (62), and the cable (62) of the vibration sensor (6) is connected to a vibration measuring instrument.

4. The chemical pump vibration detection system according to claim 1, characterized in that: The ferrule (5) and the fixed plate (3) are fixedly connected via two sets of side plates (51), and the two sets of side plates (51) are respectively located on two sides of the ferrule (5).

5. The chemical pump vibration detection system according to claim 1, characterized in that: An elastic anti-slip lining (52) is provided on the inner side surface of the ferrule (5), and the elastic anti-slip lining (52) is bonded and fixed to the ferrule (5); The installation direction of the vibration sensor (6) is perpendicular to the axial direction of the rotating machine base (71).

6. The chemical pump vibration detection system according to claim 1, characterized in that: The analysis unit also includes a remote monitoring unit, which is used to send the detection data to a remote server through a wireless communication module for real-time monitoring and analysis by a user.

7. The chemical pump vibration detection system according to claim 1, characterized in that: The detection frame (1) adopts an adjustable structure and is adapted to chemical pump bodies (7) of different sizes and shapes.

Images