A grinding device and a fault detection method for a chemical raw material

Abstract

The present invention relates to the technical field of crushing and grinding processing, and particularly relates to a grinding device and a fault detection method for chemical raw materials, including: screening out a target sequence segment according to the abnormality degree of the time sequence of motor vibration data; obtaining the interference time delay of each motor vibration data by the target sequence segment based on several data during the operation of the motor and the grinding chamber body, further calculating the probability that the motor vibration is interfered by the vibration of the grinding chamber body, and obtaining the main target sequence segment from the target sequence segment; analyzing the motor vibration data in the main target sequence segment to obtain a fault component; obtaining the severity of each fault component according to the interference of each fault component on the grinding chamber body, and further obtaining a fault detection result. The present invention obtains the probability that the vibration data of the motor is interfered by the vibration of the grinding chamber body according to the correlation between the vibration data at different positions of the grinding chamber body and the motor vibration data, thereby improving the accuracy of motor fault detection.

Description

Technical Field

[0001] The present invention relates to the technical field of crushing and grinding processing, and particularly relates to a grinding device and a fault detection method for chemical raw materials. Background Art

[0002] Chemical raw materials have extremely wide applications in the chemical industry and are the basis for manufacturing various chemical products. In the production and processing of chemical raw materials, grinding is a crucial step. The grinding device improves the utilization rate of raw materials and the quality of products by crushing, mixing, and homogenizing the raw materials. As the core component of the grinding device, the operation state of the motor directly affects the performance and production efficiency of the entire device. Fault detection of the motor of the grinding device can improve the safety and stability of the device operation and ensure the grinding efficiency.

[0003] In the prior art CN118310625B (a robot servo drive fault monitoring method and system), fault identification is carried out by analyzing the resonance phenomenon between different vibration data. This method usually obtains the vibration data of the motor by using vibration sensors and then analyzes the motor vibration data. When the fluctuation of the motor vibration data appears abnormally, it is considered that the motor has a fault. However, when the grinding device is working, the grinding chamber will have complex vibration phenomena when grinding chemical raw materials, which will interfere with the vibration data of the motor. This interference is not only caused by the resonance effect but also exists interference of other complex vibration frequencies. This vibration interference will reduce the accuracy of motor fault identification. Summary of the Invention

[0004] The present invention provides a grinding device and a fault detection method for chemical raw materials to solve the existing problems.

[0005] A grinding fault detection method for chemical raw materials of the present invention adopts the following technical solution:

[0006] An embodiment of the present invention provides a grinding fault detection method for chemical raw materials, and the method includes the following steps:

[0007] Obtain a plurality of data during the operation of the motor and the grinding chamber; the plurality of data are vibration frequency data obtained by using vibration sensors and the distances from different positions of the grinding chamber to the motor; the vibration frequency data includes the time series of the chamber vibration data and the time series of the motor vibration data;

[0008] Filter out the target sequence segments according to the abnormality degree of the time series of motor vibration data; take the position closest to the motor in the grinding bin as the reference position, and obtain the average difference between the time corresponding to the adjacent target data closest to the motor vibration data and the acquisition time of the motor vibration data as the interference time delay of each motor vibration data by the target sequence segment. The target data is the vibration data of the grinding bin when the change rate difference between the target sequence segment and the vibration data of the grinding bin at the reference position is less than the preset threshold. Further calculate the probability that the motor vibration is interfered by the vibration of the grinding bin. The probability is positively correlated with the interference time delay of each motor vibration data in the target sequence segment by the grinding bin, and negatively correlated with the vibration data of the grinding bin at the time before the interference time delay corresponding to the motor vibration data at the reference position that causes interference to the motor vibration data in the target sequence segment. Obtain the main target sequence segment from the target sequence segment; analyze the motor vibration data in the main target sequence segment to obtain the fault components;

[0009] According to the interference of each fault component to the grinding bin, obtain the severity of each fault component, and further obtain the fault detection result.

[0010] Further, the method of filtering out the target sequence segment according to the abnormality degree of the time series of motor vibration data specifically includes:

[0011] Equally divide the time series of motor vibration data into several sequence segments, and perform fast Fourier transform on each sequence segment respectively to obtain the vibration energy values and frequency spectra corresponding to the several sequence segments; obtain the frequency component with the largest amplitude in the frequency spectrum of each sequence segment as the main frequency of the sequence segment, denoted as ; preset the frequency bandwidth range of the sequence segment according to the main frequency of each sequence segment;

[0012] Calculate the abnormality degree of the sequence segment according to the vibration energy value and frequency bandwidth range corresponding to each sequence segment;

[0013] Preset threshold , and mark the sequence segment with an abnormality degree greater than or equal to the preset threshold as the target sequence segment.

[0014] Further, the method of calculating the abnormality degree of the sequence segment according to the vibration energy value and frequency bandwidth range corresponding to each sequence segment specifically includes:

[0015]

[0016] Among them, represents the abnormality degree of the th sequence segment; represents the number of all sequence segments in the time series of motor vibration data; represents the th number of data included in the sequence segment; Indicates the vibration energy value corresponding to the -th position in the -th sequence segment; Indicates the vibration energy value corresponding to the -th position in the -th sequence segment except the -th sequence segment; Indicates the sum of all vibration energy values within the frequency bandwidth of the preset -th sequence segment; Indicates the sum of all vibration energy values within the frequency bandwidth of all sequence segments; Indicates the absolute value symbol; Indicates the exponential function with the natural constant as the base; Indicates the normalization function.

[0017] Furthermore, the specific method for obtaining the interference time delay of each motor vibration data by the target sequence segment is as follows:

[0018] Use the least squares method to fit all the motor vibration data in each target sequence segment to obtain the slope of each motor vibration data on the fitting curve; calculate the difference between the slope of the data in the target sequence segment on the fitting curve and the slope corresponding to the vibration data of the reference position grinding bin body, and a preset threshold , and record the vibration data of the grinding bin body corresponding to the case where the difference is less than the preset threshold as the target data;

[0019] Among all the target data corresponding to the motor vibration data, the difference between the time corresponding to the previous target data closest to the motor vibration data and the acquisition time of the motor vibration data is used as the reference time delay of the motor vibration data, and the average value of the reference time delays of all the motor vibration data in the -th target sequence segment is used as the interference time delay of the vibration of the grinding bin body on the motor vibration data.

[0020] Furthermore, the specific method for calculating the probability that the motor vibration is interfered by the vibration of the grinding bin body includes:

[0021]

[0022] Among them, Indicates the probability that the motor vibration in the -th target sequence segment is interfered by the vibration of the grinding bin body; Indicates the degree of abnormality of the -th target sequence segment; Indicates the number of motor vibration data in the -th target sequence segment; Indicates the -th position in the Motor vibration data; Indicates the th target sequence segment at the reference position, the th motor vibration data corresponding to the vibration data of the grinding bin body before the interference time delay ; Indicates the Pearson correlation coefficient between the motor vibration data of the th target sequence segment and the vibration data of the grinding bin body at the reference position; Indicates the interference time delay of the vibration of the grinding bin body on each motor vibration data in the th target sequence segment; Indicates the absolute value symbol; Indicates the exponential function with the natural constant as the base.

[0023] Further, the method for obtaining the main target sequence segment from the target sequence segment specifically includes: presetting a threshold . When the probability that the motor vibration data of the target sequence segment is interfered by the bin body vibration data is greater than the threshold , it is marked as a normal situation; when the probability that the motor vibration data of the target sequence segment is interfered by the bin body vibration data is less than or equal to the threshold , the target sequence segment is marked as the main target sequence segment.

[0024] Further, the method for analyzing the motor vibration data in the main target sequence segment to obtain the fault component specifically includes:

[0025] According to the difference in the vibration energy values between each vibration frequency and other vibration frequencies, combined with the probability that the motor vibration is interfered by the grinding bin body vibration, calculate the possibility that the vibration frequency in the main target sequence segment is a fault component;

[0026] Take the vibration frequency with the highest possibility of being a fault component in the th main target sequence segment as the fault component.

[0027] Further, the method for calculating the possibility that the vibration frequency in the main target sequence segment is a fault component specifically includes:

[0028]

[0029] Among them, Indicates the possibility that the th vibration frequency in the th main target sequence segment is a fault component; Indicates the th main target sequence segment, the The probability that the motor vibration data in the main target sequence segment where a vibration frequency is located is interfered by the vibration data of the grinding bin body; Indicates the th vibration energy value corresponding to the vibration frequency in the main target sequence segment; Indicates the th vibration frequency in the main target sequence segment except for the th vibration energy value corresponding to the vibration frequency; Indicates the number of vibration frequencies included in the th

[0030] Further, the method for obtaining the severity of each fault component based on the interference of each fault component on the grinding bin body, and then obtaining the fault detection result includes the following specific methods:

[0031]

[0032] Among them, Indicates the severity of the fault component existing in the th main target sequence segment; Indicates the number of positions where vibration sensors are installed on the grinding bin body; Indicates the vibration energy value of the th position on the grinding bin body corresponding to the th fault component data in the th main target sequence segment; Indicates the vibration energy value of the th

[0033] preset threshold When the severity of the fault component existing in all main target sequence segments is greater than the threshold it is determined that the motor has a fault.

[0034] An embodiment of the present invention provides a grinding device for chemical raw materials, including a grinding device body, a memory, a processor included in the grinding device, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of any one of the above-mentioned grinding fault detection methods for chemical raw materials.

[0035] The beneficial effects of the technical solution of the present invention are as follows: the grinding fault detection method of chemical raw materials provided in the present embodiment aims at the defect that the traditional method relies on the experience of operators and needs to be inspected and shut down for repair only after an obvious fault occurs in the equipment. Considering the combination of several data of the motor and the grinding bin body during operation, according to the abnormality of the motor vibration data time series, the target sequence segment is screened out to understand the correlation between the motor vibration data and the bin body vibration data; then, based on several data of the motor and the grinding bin body during operation, the interference delay of the target sequence segment on each motor vibration data is obtained, and the probability that the motor vibration is interfered by the grinding bin body vibration is further calculated to obtain the main target sequence segment; the motor vibration data in the main target sequence segment is analyzed to obtain the fault component, which can help determine the specific location of the fault and facilitate targeted maintenance; at the same time, according to the interference of each fault component on the grinding bin body, the severity of each fault component is obtained, and then the fault detection result is obtained, and the fault source is accurately located. By using data analysis technology, a more objective and scientific basis for fault judgment can be obtained, and the deviation relying on manual experience can be reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. 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 paying creative work.

[0037] Figure 1 The present invention is a flow chart of the steps of a method for detecting grinding faults of chemical raw materials. DETAILED DESCRIPTION

[0038] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the grinding fault detection method of a chemical raw material proposed by the present invention, its specific implementation method, structure, characteristics and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0039] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0040] The specific scheme of the grinding fault detection method for chemical raw materials provided by the present invention is described in detail below with reference to the accompanying drawings.

[0041] See also Figure 1, which shows the step flow chart of a method for detecting grinding faults of a chemical raw material provided by an embodiment of the present invention. The method includes the following steps:

[0042] Step S001: Obtain a number of data during the operation of the motor and the grinding chamber.

[0043] It should be noted that during the grinding process, the non-uniformity of the chemical raw material will cause the force on the grinding chamber to be uneven, which will in turn affect the vibration stability of the grinding chamber. The vibration stability of the grinding chamber will interfere with the vibration data of the motor. Therefore, it is difficult to determine whether the abnormal vibration data of the motor is caused by a fault of the motor itself or by the interference of the vibration of the grinding chamber.

[0044] Specifically, in order to implement a method for detecting grinding faults of a chemical raw material proposed in this embodiment, first, a number of data during the operation of the motor and the grinding chamber need to be collected. The specific process is as follows:

[0045] Vibration sensors are installed at different positions of the grinding chamber respectively to obtain the vibration data at different positions during the grinding process, and they are sorted according to the acquisition time sequence, denoted as the time sequence of the vibration data of the chamber; an acceleration sensor is installed at the motor, and the vibration information of the motor is obtained through a vibration analyzer and sorted according to the acquisition time sequence, denoted as the time sequence of the vibration data of the motor; the distances from different positions of the grinding chamber to the motor are obtained.

[0046] So far, a number of data during the operation of the motor and the grinding chamber are obtained through the above method.

[0047] Step S002: According to the degree of abnormality of the time sequence of the motor vibration data, filter out the target sequence segments; based on a number of data during the operation of the motor and the grinding chamber, obtain the interference delay of each motor vibration data by the target sequence segments, further calculate the probability that the motor vibration is interfered by the vibration of the grinding chamber, and obtain the main target sequence segments from the target sequence segments; analyze the motor vibration data in the main target sequence segments to obtain the fault components.

[0048] It should be noted that due to the non-uniformity of the chemical raw material itself, such as uneven particle size distribution and inconsistent density, etc., the force on the grinding chamber during grinding will be uneven. This uneven force will cause the vibration of the grinding chamber to be unstable, which will in turn feedback to the motor and affect the stability of the motor vibration data; the vibration of the chamber during the grinding process causes the abnormal motor vibration data, and the abnormality of the motor will be reflected in the vibration signals at different positions of the grinding chamber, and calculate the probability that the instability of the motor vibration data is interfered by the vibration of the chamber.

[0049] Specifically, step 2.1, according to the degree of abnormality of the time sequence of the motor vibration data, filter out the target sequence segments.

[0050] It should be noted that the fast Fourier transform can convert a time-domain signal into a frequency-domain signal, thereby observing the amplitudes of different frequency components in the signal. The dominant frequency is the frequency component with the largest amplitude in the frequency spectrum, which usually reflects the dominant vibration mode of the system. The stability of the dominant frequency reflects the stability of the main vibration mode of the system. If the dominant frequency changes frequently or is unstable, then the vibration mode of the system will also become unstable.

[0051] The time series of the motor vibration data is equally divided into several sequence segments, and the fast Fourier transform is performed on each sequence segment respectively to obtain the vibration energy values and frequency spectra corresponding to the several sequence segments. The frequency component with the largest amplitude in the frequency spectrum of each sequence segment is obtained as the dominant frequency of the sequence segment, denoted as ; The frequency bandwidth range of the sequence segment is preset according to the dominant frequency of each sequence segment. According to the corresponding vibration energy value and frequency bandwidth range of the sequence segment, the degree of abnormality of the sequence segment is calculated.

[0052] It should be noted that in the embodiment of the present invention, the frequency bandwidth range of each sequence segment is preset according to experience as , which can be adjusted according to the actual situation and is not specifically limited in this embodiment.

[0053] As an embodiment, the specific method for calculating the degree of abnormality of the sequence segment is:

[0054]

[0055] Among them, represents the degree of abnormality of the th sequence segment; represents the number of all sequence segments in the time series of the motor vibration data; represents the th sequence segment; represents the th position in the th sequence segment; represents the th sequence segment; represents the th position in the th sequence segment; represents the sum of all vibration energy values within the preset frequency bandwidth range of the th sequence segment; represents the absolute value symbol; represents the exponential function with the natural constant as the base; represents the normalization function.

[0056] It should be noted that The larger the value, the more prominent the main frequency of the th sequence segment, and the frequency components are mainly concentrated near the main frequency, and the higher the frequency stability, the smaller the degree of abnormality of the sequence segment.

[0057] Preset threshold , and the sequence segments with the degree of abnormality greater than or equal to the preset threshold are recorded as target sequence segments.

[0058] It should be noted that in this embodiment, the threshold for screening target sequence segments is preset according to experience , and it can be adjusted according to the actual situation, and no specific limitation is made in this embodiment.

[0059] Step 2.2, based on a number of data during the operation of the motor and the grinding bin body, obtain the interference time delay of each target sequence segment to the vibration data of each motor, further calculate the probability that the motor vibration is interfered by the vibration of the grinding bin body, and obtain the main target sequence segment from the target sequence segments.

[0060] It should be noted that due to the interference caused by the vibration data of the grinding bin body, the vibration data of the motor will be abnormal, but this abnormality caused by interference will not occur immediately and there is a certain time delay. When the vibration data of the grinding bin body is more similar to the vibration data of the motor after a certain time delay, the credibility of the abnormality degree of the vibration data of this sequence segment of the motor is lower, that is, at this time, the abnormality of the vibration data of the motor is caused by the interference of the normal vibration of the grinding bin body during the grinding of chemical raw materials. Therefore, according to the correlation between the vibration data at different positions of the grinding bin body and the vibration data of the motor, the interference probability of the vibration data of the motor in different sequence segments by the vibration data of the grinding bin body is obtained.

[0061] Take the position of the grinding bin body closest to the motor as the reference position, use the least squares method to fit all the motor vibration data in each target sequence segment, and obtain the slope of each motor vibration data on the fitting curve; calculate the difference between the slope of the data in the target sequence segment on the fitting curve and the slope corresponding to the vibration data of the grinding bin body at the reference position, preset threshold , and record the vibration data of the grinding bin body corresponding to the difference less than the preset threshold as the target data.

[0062] It should be noted that in this embodiment, the threshold for screening target data is preset according to experience , and it can be adjusted according to the actual situation, and no specific limitation is made in this embodiment.

[0063] Among all the target data corresponding to the motor vibration data, take the difference between the time corresponding to the previous target data closest to the motor vibration data and the acquisition time of the motor vibration data as the reference time delay of the motor vibration data. Take the The mean of the reference time delays of all motor vibration data in a target sequence segment is used as the interference time delay of the vibration of the grinding bin on the motor vibration data.

[0064] As an embodiment, the specific method for calculating the interference time delay of a target sequence segment on each motor vibration data is as follows:

[0065]

[0066] Wherein, represents the interference time delay of the vibration of the grinding bin on each motor vibration data in the th target sequence segment, represents the number of motor vibration data in the th target sequence segment, represents the reference time delay of the th motor vibration data in the th target sequence segment.

[0067] As an embodiment, the specific method for calculating the probability that the motor vibration of a target sequence segment is interfered by the vibration of the grinding bin is as follows:

[0068]

[0069] Wherein, represents the probability that the motor vibration of the th target sequence segment is interfered by the vibration of the grinding bin; represents the degree of abnormality of the th target sequence segment; represents the number of motor vibration data in the th target sequence segment; represents the th motor vibration data in the th target sequence segment; represents the vibration data of the grinding bin corresponding to the th motor vibration data in the th target sequence segment at the reference position before the interference time delay ; represents the Pearson correlation coefficient between the motor vibration data of the th target sequence segment and the vibration data of the grinding bin at the reference position; represents the interference time delay of the vibration of the grinding bin on each motor vibration data in the th target sequence segment; represents the absolute value symbol; represents the exponential function with the natural constant as the base.

[0070] It should be noted that represents the similarity between two vibration data. The closer the ratio is to 1, the greater the probability that the motor vibration data on the th target sequence segment is interfered by the silo vibration data. represents the probability that the instability of the motor vibration on the th target sequence segment is affected by the motor itself. The smaller the interference time delay, the greater the Pearson correlation coefficient between the motor vibration data on the th target sequence segment and the vibration data of the grinding silo at the reference position. It indicates that the probability that the motor vibration data on the th target sequence segment is affected by the motor itself is greater, and the corresponding probability that the motor vibration on the th target sequence segment is interfered by the vibration of the grinding silo is smaller.

[0071] Preset threshold , when the probability that the motor vibration data of the target sequence segment is interfered by the silo vibration data is greater than the threshold , it is marked as a normal situation; when the probability that the motor vibration data of the target sequence segment is interfered by the silo vibration data is less than or equal to the threshold , the target sequence segment is marked as the main target sequence segment.

[0072] It should be noted that when the probability that the motor vibration data of the target sequence segment is interfered by the silo vibration data is greater than the threshold , it indicates that the instability of the motor vibration data at this time is a normal phenomenon; when the probability that the motor vibration data of the target sequence segment is interfered by the silo vibration data is less than or equal to the threshold , it indicates that the instability of the motor vibration data at this time is caused by a fault in the motor itself.

[0073] It should be noted that in this embodiment, the threshold for screening the main target sequence segment is preset according to experience , which can be adjusted according to the actual situation, and this embodiment does not make specific limitations.

[0074] Step 2.3, analyze the motor vibration data in the main target sequence segment to obtain the fault component.

[0075] It should be noted that when a motor fails, a specific frequency of vibration will be generated at the fault location, causing a significant change in the vibration energy value of this vibration frequency, which is different from the numerical distribution in the normal state. According to the difference between the vibration energy value of each vibration frequency and the vibration energy values of other vibration frequencies, the possibility of each vibration frequency being a fault component is obtained.

[0076] According to the difference between the vibration energy value of each vibration frequency and the vibration energy values of other vibration frequencies, combined with the probability that the motor vibration is interfered by the vibration of the grinding silo, calculate the possibility that the vibration frequency in the main target sequence segment is a fault component.

[0077] As an embodiment, the specific method for calculating the possibility that the vibration frequency in the main target sequence segment is a fault component is as follows:

[0078]

[0079] Wherein, represents the possibility that the th vibration frequency in the th main target sequence segment is a fault component; represents the probability that the motor vibration data of the th vibration frequency in the th main target sequence segment is interfered by the vibration data of the grinding bin body; represents the vibration energy value corresponding to the th vibration frequency in the th main target sequence segment; represents the vibration energy value corresponding to the th vibration frequency except the th vibration frequency in the th main target sequence segment; represents the number of vibration frequencies included in the th main target sequence segment; represents the absolute value symbol.

[0080] It should be noted that when a fault occurs in the motor itself, the abnormality of the vibration data will be persistent, that is, an abnormality will also occur in the next time period. When the th vibration frequency in the th main target sequence segment has a greater degree of abnormality and a greater correlation with the same vibration frequency in the adjacent next time period, it indicates that the th vibration frequency in the th main target sequence segment is more likely to be a fault component; the smaller the difference, the smaller the possibility that the component is a fault component, and the greater the difference, the greater the possibility that the vibration frequency is a fault component.

[0081] The vibration frequency with the greatest possibility of being a fault component in the th main target sequence segment is recorded as the fault component.

[0082] Thus far, the fault component is obtained through the above method.

[0083] Step S003: According to the interference of each fault component on the grinding bin body, obtain the severity of each fault component, and further obtain the fault detection result.

[0084] As an embodiment, the specific calculation method for calculating the severity of each fault component is as follows:

[0085]

[0086] Among them, represents the severity of the fault component existing in the th main target sequence segment; represents the number of positions where vibration sensors are installed on the grinding bin body; represents the th main target sequence segment and the th fault component data, corresponding to the vibration energy value at the th position on the grinding bin body; represents the vibration energy value of the th main target sequence segment and the th fault component data; represents the number of fault component data included in the th main target sequence segment; represents the absolute value symbol.

[0087] It should be noted that represents the difference between the vibration energy value of the fault component in the th main target sequence segment and the vibration energy value of the grinding bin body at the th position. The smaller the difference, the more consistent the fault is with the vibration data of the bin body, and the greater the severity of the motor fault.

[0088] Preset threshold , when the severity of the fault component in all main target sequence segments is greater than the threshold , it is determined that the motor has a fault.

[0089] It should be noted that in this embodiment, the threshold for screening the main target sequence segment is preset according to experience and can be adjusted according to actual situations. No specific limitation is made in this embodiment.

[0090] Through the above steps, a method for detecting grinding faults of chemical raw materials is completed.

[0091] On the other hand, an embodiment of the present invention provides a grinding device for chemical raw materials, including a grinding device body, a memory, a processor included in the grinding device, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the method for detecting grinding faults of chemical raw materials described in steps S001 to S003 are implemented.

[0092] The present embodiment provides a method for detecting grinding faults of chemical raw materials. In view of the defects that the traditional method relies on the experience of operators and needs to be shut down for inspection and repair only after an obvious fault occurs in the equipment, the method takes into account a number of data of the motor and the grinding bin body during operation, and according to the abnormality of the motor vibration data time series, a target sequence segment is screened out to understand the correlation between the motor vibration data and the bin body vibration data; then, based on a number of data of the motor and the grinding bin body during operation, the interference delay of the target sequence segment on each motor vibration data is obtained, and the probability that the motor vibration is interfered by the grinding bin body vibration is further calculated to obtain the main target sequence segment; the motor vibration data in the main target sequence segment is analyzed to obtain the fault component, which can help determine the specific location of the fault and facilitate targeted maintenance; at the same time, according to the interference of each fault component on the grinding bin body, the severity of each fault component is obtained, and then the fault detection result is obtained, and the fault source is accurately located. By using data analysis technology, a more objective and scientific basis for fault judgment can be obtained, and the deviation relying on manual experience can be reduced.

[0093] It should be noted that the The model is only used to represent negative correlation and constrain the output of the model to be in In the specific implementation, it can be replaced by other models with the same purpose. This embodiment is only based on The model is described as an example without any specific limitation. is the input to the model.

[0094] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for detecting grinding faults of chemical raw materials, characterized in that The method includes the following steps: Obtain a number of data during the operation of the motor and the grinding chamber body; the number of data is vibration frequency data obtained by using vibration sensors, and the distances from different positions on the grinding chamber body where vibration sensors are arranged to the motor; the vibration frequency data includes the time series of the vibration data of the chamber body and the time series of the vibration data of the motor; According to the degree of abnormality of the time series of the motor vibration data, filter out the target sequence segment; take the position on the grinding chamber body closest to the motor as the reference position, and obtain the average difference between the time corresponding to the adjacent target data closest to the motor vibration data and the acquisition time of the motor vibration data as the interference time delay of each motor vibration data by the target sequence segment. The target data is the vibration data of the grinding chamber body when the change rate difference between the target sequence segment and the vibration data of the grinding chamber body at the reference position is less than a preset threshold. Further calculate the probability that the motor vibration is interfered by the vibration of the grinding chamber body. The probability is positively correlated with the interference time delay of each motor vibration data in the target sequence segment by the grinding chamber body, and negatively correlated with the vibration data of the grinding chamber body at the time before the interference time delay corresponding to the motor vibration data at the reference position that causes interference to the motor vibration data in the target sequence segment. Obtain the main target sequence segment from the target sequence segment; analyze the motor vibration data in the main target sequence segment to obtain the fault component; According to the interference of each fault component to the grinding chamber body, obtain the severity of each fault component, and further obtain the fault detection result; The specific method for obtaining the interference time delay of each motor vibration data by the target sequence segment is as follows: Use the least squares method to fit all the motor vibration data in each target sequence segment to obtain the slope of each motor vibration data on the fitting curve; calculate the difference between the slope of the data in the target sequence segment on the fitting curve and the slope corresponding to the vibration data of the reference position grinding bin body, and a preset threshold , and record the vibration data of the grinding bin body corresponding to the case where the difference is less than the preset threshold as the target data; Among all the target data corresponding to the motor vibration data, the difference between the time corresponding to the previous target data closest to the motor vibration data and the acquisition time of the motor vibration data is used as the reference time delay of the motor vibration data. The average value of the reference time delays of all the motor vibration data in the th target sequence segment is used as the interference time delay of the vibration of the grinding bin body on the motor vibration data.

2. The grinding fault detection method for a chemical raw material according to claim 1, wherein, The method of filtering out the target sequence segment according to the degree of abnormality of the time series of the motor vibration data includes the following specific method: The time series of the motor vibration data is equally divided into several sequence segments, and the fast Fourier transform is performed on each sequence segment respectively to obtain the vibration energy values and frequency spectra corresponding to the several sequence segments; the frequency component with the largest amplitude in the frequency spectrum of each sequence segment is obtained as the main frequency of the sequence segment, denoted as ; the frequency bandwidth range of the sequence segment is preset according to the main frequency of each sequence segment; Calculate the degree of abnormality of the sequence segment according to the vibration energy value and the frequency bandwidth range corresponding to each sequence segment; Preset threshold Sequences with an abnormality degree greater than or equal to the preset threshold are recorded as target sequences.

3. The grinding fault detection method for a chemical raw material according to claim 2, characterized in that, The specific method included in calculating the degree of abnormality of the sequence segment according to the vibration energy value and the frequency bandwidth range corresponding to each sequence segment is as follows: Among them, represents the abnormality degree of the th sequence segment; represents the number of all sequence segments in the time series of motor vibration data; represents the th sequence segment, the number of data it contains; represents the th sequence segment, the vibration energy value corresponding to the th position; represents the vibration energy value corresponding to the th sequence segment, except for the th sequence segment, at the th position; represents the sum of all vibration energy values within the frequency bandwidth range of the th preset sequence segment; represents the sum of all vibration energy values within the frequency bandwidth range of all sequence segments; represents the absolute value symbol; represents the exponential function with the natural constant as the base; represents the normalization function.

4. The grinding fault detection method for a chemical raw material according to claim 1, wherein, The specific method included in calculating the probability that the motor vibration is interfered by the vibration of the grinding chamber body is as follows: Among them, represents the probability that the motor vibration of the th target sequence segment is interfered by the vibration of the grinding bin body; represents the degree of abnormality of the th target sequence segment; represents the number of motor vibration data in the th target sequence segment; represents the th motor vibration data in the th target sequence segment; represents the vibration data of the grinding bin body corresponding to the th motor vibration data in the th target sequence segment at the reference position before the interference time delay ; represents the Pearson correlation coefficient between the motor vibration data of the th target sequence segment and the vibration data of the grinding bin body at the reference position; represents the interference time delay of the vibration of the grinding bin body to each motor vibration data in the th target sequence segment; represents the absolute value symbol; represents the exponential function with the natural constant as the base.

5. The grinding fault detection method for a chemical raw material according to claim 1, characterized in that, The method for obtaining the main target sequence segment from the target sequence segment specifically includes: presetting a threshold , when the probability that the motor vibration data of the target sequence segment is interfered by the silo vibration data is greater than the threshold , it is marked as a normal situation; when the probability that the motor vibration data of the target sequence segment is interfered by the silo vibration data is less than or equal to the threshold , the target sequence segment is marked as the main target sequence segment.

6. The grinding fault detection method for a chemical raw material according to claim 1, wherein The specific method included in analyzing the motor vibration data in the main target sequence segment to obtain the fault component is as follows: According to the difference in the vibration energy values between each vibration frequency and other vibration frequencies, and in combination with the probability that the motor vibration is interfered by the vibration of the grinding chamber body, calculate the possibility that the vibration frequency in the main target sequence segment is a fault component; Denote the vibration frequency with the highest probability of being a fault component in the th main target sequence segment as the fault component.

7. The grinding fault detection method for a chemical raw material according to claim 6, wherein The specific method included in calculating the possibility that the vibration frequency in the main target sequence segment is a fault component is as follows: Among them, represents the probability that the -th vibration frequency in the -th main target sequence segment is a fault component; represents the probability that the motor vibration data of the main target sequence segment where the -th vibration frequency is located in the -th main target sequence segment is interfered by the vibration data of the grinding chamber body; represents the vibration energy value corresponding to the -th vibration frequency in the -th main target sequence segment; represents the vibration energy value corresponding to the -th vibration frequency in the -th main target sequence segment, excluding the -th vibration frequency; represents the number of vibration frequencies included in the -th main target sequence segment; represents the absolute value symbol.

8. The grinding fault detection method for a chemical raw material according to claim 1, wherein The specific method included in obtaining the severity of each fault component according to the interference of each fault component to the grinding chamber body, and further obtaining the fault detection result is as follows: Among them, represents the severity of the fault component in the th main target sequence segment; represents the number of positions where vibration sensors are installed on the grinding bin body; represents the th main target sequence segment and the th fault component data, corresponding to the vibration energy value at the th position on the grinding bin body; represents the th main target sequence segment and the th fault component data of the vibration energy value; represents the th main target sequence segment containing the number of fault component data; represents the absolute value symbol; Preset threshold When the severity of the fault component in all main target sequence segments is greater than the threshold it is determined that the motor has a fault.

9. A grinding device for chemical raw materials, comprising a grinding device body, a memory, a processor included in the grinding device, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method described in any one of claims 1 to 8.

Images