Water system fault prediction management platform based on sensor

Abstract

The invention relates to the technical field of data processing, in particular to a water system fault prediction management platform based on a sensor, which comprises a memory, a processor and a computer program stored in the memory and running on the processor, when the processor executes the computer program, the following steps are realized: acquiring pressure data of the water system to obtain a to-be-analyzed data sequence and positions of vacant values in the to-be-analyzed data sequence, and obtaining a reference interpolation of each vacant position according to the similarity between data in a historical period of time and the to-be-analyzed data sequence, obtaining a reference interpolation proximity according to the difference between the data sequence to be analyzed and the reference interpolation, and obtaining a position deviation degree according to the change trend of the reference interpolation; the target data is obtained according to the number of the vacant positions, the reference interpolation proximity and the position deviation degree, then the vacant positions are complemented by using the Newton interpolation method, fault early warning is performed on the water system according to the complemented to-be-analyzed data sequence, and the accuracy of fault prediction on the water system is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a water system fault prediction management platform based on sensors. Background Art

[0002] In modern industry and urban life, the fire-fighting water system is closely related to people's life safety. According to statistics, the proportion of fire fighting failures caused by unstable water systems accounts for 81%. Therefore, the fault prediction of the fire-fighting water system is particularly important.

[0003] In the prior art, pressure data of a water system in a non-working state is collected to obtain a data sequence, and then fault prediction of the water system is performed based on the data sequence. The pressure data of the water system is usually obtained by a pressure sensor. Since the pressure sensor is generally installed in a pipe well, an inspection port, and other locations in the water system, humidity interference to the sensor may be caused by water flowing through it, or the connection equipment between the sensor and the collection system may be loose, the power supply of the sensor may be unstable, and other conditions may result in single missing or continuous missing values ​​in the collected data sequence. Missing data may lead to inaccurate prediction results. Therefore, before predicting faults of the water system based on the pressure data, it is necessary to pre-process the missing values ​​in the data sequence to complete them.

[0004] The traditional method uses Newton interpolation to pre-process the missing values ​​in the data sequence. Although Newton interpolation has higher data interpolation accuracy when processing stable data sequences, the order in Newton interpolation is affected by the number of known data points. It has high accuracy when processing single vacancies, but for continuous vacancies in the data sequence, the uncertainty in the number selection will lead to uncertain interpolation results, which in turn leads to inaccurate fault prediction results for the fire water system.

[0005] Therefore, how to interpolate and complete the missing data in the data sequence according to the Newton interpolation method and improve the accuracy of fault prediction for the water system has become an urgent problem to be solved. Summary of the invention

[0006] In view of this, an embodiment of the present invention provides a sensor-based water system fault prediction management platform to solve the problem of xx.

[0007] In an embodiment of the present invention, a sensor-based water system fault prediction management platform is provided, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the following steps when executing the computer program: Acquire pressure data of the water system in a target period to obtain a data sequence to be analyzed, and acquire at least one vacant position sequence in the data sequence to be analyzed, wherein the vacant position sequence is composed of vacant positions corresponding to consecutive vacant values ​​in the data sequence to be analyzed; For any vacant position sequence, according to the similarity between the pressure data of the water system in the historical period and the data in the data sequence to be analyzed, the reference interpolation of each vacant position in the vacant position sequence is obtained; according to the difference between the data in the data sequence to be analyzed and each reference interpolation, the reference interpolation proximity corresponding to the vacant position sequence is obtained; according to the change trend between each two adjacent reference interpolations, the position deviation of all reference interpolations is obtained; Obtaining target data according to the number of vacant positions in any vacant position sequence, the reference interpolation proximity and the position deviation, and interpolating and completing the vacant positions in any vacant position sequence using Newton interpolation method according to the target data; Interpolation is performed on the vacant positions in each of the vacant position sequences to obtain a completed sequence of data to be analyzed, and fault warning is performed on the water system according to the completed sequence of data to be analyzed.

[0008] Compared with the prior art, the embodiments of the present invention have the following beneficial effects: The present invention obtains pressure data of a water system in a target period to obtain a data sequence to be analyzed, obtains at least one vacant position sequence in the data sequence to be analyzed, and the vacant position sequence is composed of vacant positions corresponding to continuous vacant values ​​in the data sequence to be analyzed; for any vacant position sequence, according to the similarity between the pressure data of the water system in a historical period and the data in the data sequence to be analyzed, obtains a reference interpolation value of each vacant position in the any vacant position sequence, and obtains a reference interpolation value of each vacant position in the any vacant position sequence according to the difference between the data in the data sequence to be analyzed and each reference interpolation value. The reference interpolation proximity corresponding to the position sequence is obtained, and the position deviation of all reference interpolations is obtained according to the change trend between each two adjacent reference interpolations; the target data is obtained according to the number of vacant positions in any vacant position sequence, the reference interpolation proximity and the position deviation, and the vacant positions in any vacant position sequence are interpolated and completed using the Newton interpolation method according to the target data; the vacant positions in each of the vacant position sequences are interpolated and completed to obtain a completed data sequence to be analyzed, and a fault warning is given to the water system according to the completed data sequence to be analyzed. Among them, according to the characteristics of the data sequence to be analyzed and historical data, the reference interpolation of each vacant position is found to make a preliminary estimate of the pressure data in each vacant position; combined with the reference interpolation proximity and position bias of the reference interpolation, and the law of Newton interpolation for interpolation and completion of vacant values, the number and distribution of target data required for Newton interpolation to interpolate and complete the vacant values ​​are determined, and then the data sequence to be analyzed is interpolated and completed using Newton interpolation, thereby improving the accuracy of interpolation and completion of continuous vacant values ​​using Newton interpolation, and predicting the fault situation of the water system according to the completed data sequence to be analyzed, thereby improving the accuracy of fault prediction for the water system. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, 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 labor.

[0010] Figure 1 is a method flow chart of a sensor-based water system fault prediction management method provided in Embodiment 1 of the present invention; Figure 2 It is a schematic diagram of an experiment using Newton interpolation method on a data sequence with continuous missing values ​​provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0011] Embodiments of the present disclosure are described in detail below, and examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present disclosure, but should not be understood as limiting the present disclosure.

[0012] It should be noted that the terms "first", "second", etc. in the specification of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure.

[0013] In order to illustrate the technical solution of the present invention, specific embodiments are provided below for illustration.

[0014] The specific scenario targeted by the present invention is: collecting pressure data of a water system in a non-working state to obtain a data sequence, and then predicting faults of the water system based on the data sequence. Since there may be missing values ​​in the data sequence, the Newton interpolation method is used to pre-process the missing values ​​in the data sequence. However, since the Newton interpolation method has high accuracy when processing a single vacancy, but low accuracy when processing continuous vacant data, the result of fault prediction of the fire-fighting water system is inaccurate. Therefore, the present invention interpolates and completes the missing data in the data sequence according to the characteristics of the data sequence itself and combines the law of Newton interpolation method in processing missing values, thereby improving the accuracy of fault prediction of the water system.

[0015] The embodiment of the present invention provides a sensor-based water system fault prediction management platform, including a processor and a memory, wherein the processor executes a computer program stored in the memory to implement a sensor-based water system fault prediction management method, such as Figure 1 As shown, the sensor-based water system fault prediction management method includes the following steps: Step S101, obtaining pressure data of a water system in a target period, obtaining a data sequence to be analyzed, and obtaining at least one vacant position sequence in the data sequence to be analyzed, wherein the vacant position sequence is composed of vacant positions corresponding to consecutive vacant values ​​in the data sequence to be analyzed.

[0016] In modern industry and urban life, the fire-fighting water system is closely related to people's life safety. According to statistics, the proportion of fire fighting failures caused by unstable water systems accounts for 81%. Therefore, the fault prediction of the fire-fighting water system is particularly important.

[0017] In the prior art, the pressure data of the water system in a non-working state is collected to obtain a data sequence, and then the water system fault prediction is performed based on the data sequence, wherein the pressure data of the water system is usually obtained by a pressure sensor, which is generally placed in the pipe well, inspection port and other locations of the water system, with a collection frequency of once every 2 minutes and a collection period of 6 hours, that is, when the water system is in a non-working state, the fault condition of the water system at a future moment is predicted based on the data sequence composed of the pressure data within every 6 hours. Since the pressure sensor is generally installed in the pipe well, inspection port and other locations of the water system, the sensor may be disturbed by humidity due to the passage of water, or the connection device between the sensor and the collection system may be loose, the power supply of the sensor may be unstable, etc., resulting in a single missing or continuous missing in the collected data sequence, and the missing data may lead to inaccurate prediction results. Therefore, before predicting the fault of the water system based on the pressure data, it is necessary to pre-process the missing values ​​in the data sequence.

[0018] The traditional method uses Newton interpolation to pre-process the missing values ​​in the data sequence. Although Newton interpolation has higher data interpolation accuracy when processing stable data sequences, the order in Newton interpolation is affected by the number of known data points. It has high accuracy when processing single vacancies, but for continuous vacancies in the data sequence, the uncertainty in the number selection will lead to uncertain interpolation results, which in turn leads to inaccurate fault prediction results for the fire water system.

[0019] Therefore, in the embodiment of the present invention, the collection frequency is preferably set to once every 2 minutes, and the collection period is 6 hours. There is no limitation here, and the implementer can set it according to different water system requirements. For example, a high-pressure water system can set a higher sampling frequency. The period at the current moment is recorded as the target period, and the pressure data in the target period is collected by the pressure sensor to obtain a pressure data sequence. The pressure data at the current moment is the last pressure data in the pressure data sequence.

[0020] In theory, that is, when there are no missing values, the theoretical number of data in the pressure data sequence is 180. First, the actual number of data in the pressure data sequence is obtained. If the actual number of data is equal to the theoretical number of data, it means that there are no missing values ​​in the pressure data sequence. Considering that the ARIMA algorithm can sensitively capture local changes in the time series and provide more accurate prediction results, when there are no missing values ​​in the pressure data sequence, the ARIMA algorithm is used to predict the pressure data of the water system at future times. There is no restriction here. The implementer can set the prediction algorithm according to the specific scenario to obtain the predicted value of the pressure data at future times. According to the pressure requirement range of the fire water system, the normal pressure data is between 0.10 and 0.15 MPa. If the predicted value of the pressure data is not within the pressure requirement range, the water system is warned of a fault and the staff is reminded to conduct inspections. Among them, the ARIMA algorithm is a prior art and will not be described here.

[0021] If the actual data quantity is not equal to the theoretical data quantity, that is, the actual data quantity is smaller than the theoretical data quantity, it means that there are missing values ​​in the pressure data sequence. If the ARIMA algorithm is used directly for prediction, the prediction result will be affected by incomplete data. Therefore, in an embodiment of the present invention, if there are two or more consecutive missing values ​​in the pressure data sequence, they are recorded as continuous missing values. For a single missing value in the pressure data sequence, considering that the Newton interpolation method has high accuracy when processing a single missing value and low accuracy when processing continuous missing values, the Newton interpolation method is used to interpolate and complete the single missing value in the pressure data sequence to obtain the data sequence to be analyzed. If there are no continuous missing values ​​in the data sequence to be analyzed, the ARIMA algorithm is used to predict the pressure data of the water system at future times based on the data sequence to be analyzed, and then a fault warning is given to the water system based on the prediction result. If there are continuous missing values ​​in the sequence to be analyzed, the missing positions of the continuous missing values ​​in the sequence to be analyzed are combined into a missing position sequence to obtain at least one missing position sequence, for example: the data sequence to be analyzed is (1, 2, 3, None, None, None, 7, 8, 9), where None represents a missing value, and the position in the data sequence to be analyzed starts from 0, that is, the positions of three continuous missing values ​​in the data sequence to be analyzed are (3, 4, 5), that is, the missing position sequence is (3, 4, 5), which is convenient for subsequent analysis and interpolation completion, where the Newton interpolation method is a prior art and will not be described here.

[0022] Step S102, for any vacant position sequence, based on the similarity between the pressure data of the water system in the historical period and the data in the data sequence to be analyzed, obtain the reference interpolation of each vacant position in the any vacant position sequence, based on the difference between the data in the data sequence to be analyzed and each reference interpolation, obtain the reference interpolation proximity corresponding to the any vacant position sequence, and based on the change trend between each two adjacent reference interpolations, obtain the position deviation of all reference interpolations.

[0023] Considering that the pressure data of the water system in a non-working state is mainly affected by environmental factors and the pressure variation range is small, the pressure data of the water system in different periods in a non-working state have strong similarity. Therefore, in an embodiment of the present invention, according to the similarity between the pressure data of the water system in a non-working state in a historical period and the data in the data sequence to be analyzed, a preliminary estimate is made of the pressure data in each vacant position to obtain a reference interpolation value for each vacant position, so as to obtain the data required for interpolation and completion of the data sequence to be analyzed using the Newton interpolation method based on the reference interpolation value, thereby improving the accuracy of interpolation and completion of continuous vacant values ​​using the Newton interpolation method and improving the accuracy of fault prediction for the water system.

[0024] Since historical data has a certain timeliness, that is, historical data farther away from the current moment may contain emergencies and have lower credibility, while historical data closer to the current moment has higher credibility. Therefore, in the embodiment of the present invention, the historical period is set to 7 days. There is no restriction here. The implementer can set it according to the specific scenario, collect the pressure data of each cycle of the water system in the non-working state before the current moment and within 7 consecutive days with the current moment, and make a preliminary estimate of the pressure data in each vacant position according to the similarity between the pressure data of each cycle and the data in the data sequence to be analyzed, so as to obtain the reference interpolation value of each vacant position.

[0025] In the embodiment of the present invention, taking the i-th vacant position sequence as an example, the method of obtaining the reference interpolation value of each vacant position in the i-th vacant position sequence is: (1) A first historical reference data sequence is obtained based on the similarity between the pressure data of the water system in a non-working state within 7 days and the data in the data sequence to be analyzed.

[0026] First, the pressure data of each cycle of the water system in the non-working state within 7 days are combined into a pressure data sequence. In order to ensure data integrity and reliability, the cycle corresponding to the pressure data sequence without missing values ​​is recorded as a historical cycle; if there are missing values ​​in each pressure data sequence, taking the ath pressure data sequence as an example, according to the position of the i-th vacant position sequence in the data sequence to be analyzed, if the data in the ath pressure data sequence that belongs to the same position as the i-th vacant position sequence is not missing, and the data adjacent to the left and right that is greater than the number of data in the i-th vacant position sequence is not missing, then the cycle corresponding to the ath pressure data sequence is recorded as a historical cycle; if the data in the same position as the i-th vacant position sequence in each pressure data sequence is still missing, the Newton interpolation method is used to interpolate and complete the missing values ​​in each pressure data sequence, and the cycle corresponding to each pressure data sequence is recorded as a historical cycle.

[0027] Considering that the credibility of pressure data in the historical period that is the same as the time period of the target period (the period in which the current moment is located) is higher, therefore, in an embodiment of the present invention, the pressure data sequence corresponding to the historical period that is the same as the time period of the target period is recorded as the first historical data sequence. If there is no historical period that is the same as the time period of the target period, the pressure data sequences corresponding to all historical periods are recorded as the first historical data sequence. Then, based on the similarity between each first historical data sequence and the data sequence to be analyzed, a reference index of each first historical data sequence is obtained, which is used to screen out the first historical data sequence that is most similar to the data sequence to be analyzed and recorded as the first historical reference data sequence, so as to facilitate the subsequent reference interpolation of each vacant position in the i-th vacant position sequence based on the first historical reference data sequence, so that the credibility of the reference interpolation of each vacant position in the i-th vacant position sequence is higher. Taking the t-th first historical data sequence as an example, the specific method of obtaining the reference index of the t-th first historical data sequence is: In the data sequence to be analyzed, the data before the i-th vacant position sequence is recorded as the left data sequence, and the data after the i-th vacant position sequence is recorded as the right data sequence; According to the position of the left data sequence in the data sequence to be analyzed, a historical left data sequence is obtained in the t-th first historical data sequence, and according to the position of the right data sequence in the data sequence to be analyzed, a historical right data sequence is obtained in the t-th first historical data sequence; Calculate the inverse of the left DTW distance between the left data sequence and the historical left data sequence to obtain the left similarity, which is recorded as , calculate the inverse of the right DTW distance between the right data sequence and the historical right data sequence, and obtain the right similarity, recorded as , calculate the average value between the left similarity and the right similarity to obtain the overall similarity between the t-th first historical data sequence and the data sequence to be analyzed, wherein the DTW distance is a prior art and will not be described in detail here; Obtain the interval period between the historical period where the t-th first historical data sequence is located and the target period, perform linear normalization on the reciprocal of the interval period, and obtain the credibility of the t-th first historical data sequence, wherein the linear normalization is a prior art and will not be described in detail here; The overall similarity degree and the credibility are weightedly summed to obtain a reference index of the t-th first historical data sequence.

[0028] In one embodiment, the calculation formula of the reference index of the t-th first historical data sequence is:

[0029] in, represents the reference index of the t-th first historical data sequence, represents the first weight, represents the second weight, Indicates the degree of left similarity, Indicates the degree of right similarity, represents the interval period between the historical period where the t-th first historical data sequence is located and the target period, represents the linear normalization function.

[0030] It should be noted that That is, the overall similarity. The greater the overall similarity, the more similar the t-th first historical data sequence is to the data sequence to be analyzed. The larger it is, the higher the credibility of the reference interpolation value of each vacant position in the i-th vacant position sequence obtained based on the t-th first historical data sequence; That is, the credibility of the t-th first historical data sequence, The larger the value is, the closer the historical period corresponding to the tth first historical data sequence is to the target period. The larger the value, the more valuable the data in the t-th first historical data sequence is for reference, and the more reliable the reference interpolation of each vacant position in the i-th vacant position sequence obtained based on the t-th first historical data sequence is. , There is no restriction here, and implementers can set it according to specific scenarios.

[0031] Similarly, the reference index of each first historical data sequence is obtained. Further, the first historical reference data sequence is screened according to the reference index of each first historical data sequence. Specifically, according to experimental statistics, the first preset range is set to , set the second preset range to There is no restriction here, and the implementer can set it according to the specific scenario. The first historical data sequence corresponding to the maximum value among all reference indices is selected as the first historical reference data sequence. If all reference indices do not belong to , but there exists The reference index of The first historical data sequence corresponding to the reference index is used as the target historical data sequence, and the average value of the data in the same position in all target data sequences is calculated to obtain the first historical reference data sequence; If all reference indices do not belong to and , then obtain the second historical data sequences of all historical periods except the historical period where the first historical data sequence is located within the historical period (within 7 days in the embodiment of the present invention), and obtain the reference index of each second historical data sequence according to the similarity between each second historical data sequence and the data sequence to be analyzed according to the method for obtaining the reference index of the first historical data sequence; If there is a reference index of all second historical data series that belongs to or The reference index is obtained by obtaining the first historical reference data sequence in the first historical data sequence, and obtaining the first historical reference data sequence in all the second historical data sequences; If all reference indices of the second historical data series do not belong to and , indicating that the data in the target period (the period where the data sequence to be analyzed is located) is not similar to the data in the historical period, and the data in the data sequence to be analyzed is abnormal. Therefore, if all the reference indexes of the second historical data sequence do not belong to and , a fault warning is given to the water system.

[0032] (2) Combining the characteristics of the i-th vacant position sequence and the first historical reference sequence, a reference interpolation value for each vacant position in the i-th vacant position sequence is obtained.

[0033] Since it is impossible to ensure that the data in the historical period has the same change pattern as the data in the target period, and at the same time, the data in the first historical reference data sequence is most similar to the data in the data sequence to be analyzed except for the i-th vacant position sequence, in the embodiment of the invention, the predicted value of each vacant position in the i-th vacant position sequence is introduced, and then the reference interpolation of each vacant position in the i-th vacant position sequence is obtained according to the predicted value of each vacant position in the i-th vacant position sequence and the similarity between the pressure data at the same position in the first historical reference data sequence and the i-th vacant position sequence, so that the reference interpolation of each vacant position in the i-th vacant position sequence has a higher credibility. Specifically: According to the data in the data sequence to be analyzed, the ARIMA algorithm is used to obtain the predicted value corresponding to each vacant position in the i-th vacant position sequence to form a predicted value sequence. There is no restriction here, and the implementer can set the prediction algorithm according to the specific scenario; According to the position of the i-th missing position sequence in the data sequence to be analyzed, pressure data at the same position as any missing position sequence is acquired in the first historical reference data sequence to form a first target reference sequence corresponding to any missing position sequence; Calculate the absolute value of the difference between two elements at each same position between the predicted value sequence and the first target reference sequence to obtain a mean of the absolute values ​​of the differences, and record the reciprocal of the mean of the absolute values ​​of the differences as the degree of numerical proximity between the predicted value sequence and the first target reference sequence; Calculate the DTW distance between the predicted value sequence and the first target reference sequence, and use the reciprocal of the DTW distance as the similarity index between the predicted value sequence and the first target reference sequence, denoted as ; Calculate the average value between the numerical proximity and the similarity index to obtain the change trend consistency index between the predicted value sequence and the first target reference sequence, denoted as ,Right now

[0034] in, Represents the consistency index of the change trend between the predicted value sequence and the first target reference sequence, represents the number of elements in the predicted value sequence, represents the jth predicted value in the predicted value sequence, represents the jth pressure data in the first target reference sequence, Represents the similarity index between the predicted value sequence and the first target reference sequence, Represents the absolute value symbol.

[0035] It should be noted that The smaller the value is, the closer the two data at the same position between the predicted value sequence and the first target reference sequence are. The bigger, The larger it is, the more similar the predicted value sequence is to the first target reference sequence, and the closer the data in the first target reference sequence is to the possible true value of each vacant position in the i-th vacant value sequence; This indicates that the predicted value sequence is more similar to the change trend of the data in the first target reference sequence. The larger it is, the more similar the predicted value sequence is to the first target reference sequence, and the closer the data in the first target reference sequence is to the possible true value of each missing position in the i-th missing value sequence.

[0036] Further, according to experimental statistics, the preset change trend consistency index threshold is set to 0.7, which is not limited here. The implementer can set it according to the specific scenario. If the change trend consistency index is greater than or equal to 0.7, the pressure data in the first target reference sequence is used as the reference interpolation value of each vacant position in any vacant position sequence; If the change trend consistency index is less than 0.7, then obtain the third historical data series of all historical cycles in the historical period, obtain the reference index of each third historical data series, and set the reference index to The third historical data sequence is used as the second historical reference data sequence, the second target reference sequence in each of the second historical reference data sequences is obtained, the change trend consistency index between the predicted value sequence and each of the second target reference sequences is calculated, the second target reference sequence corresponding to the maximum value among all the change consistency indexes is selected as the final target reference sequence, and the pressure data in the final target reference sequence is used as the reference interpolation value for each vacant position in any vacant position sequence.

[0037] At this point, the reference interpolation value of each vacant position in the i-th vacant position sequence is obtained, which is used to represent the possible numerical conditions of each vacant position. Further, the data required for interpolation and completion of the data sequence to be analyzed using the Newton interpolation method is obtained based on the reference interpolation value.

[0038] Through the experiment of using Newton interpolation method on the data sequence with continuous vacancy values, it can be known that for a data sequence with continuous vacancy values, the interpolation results obtained by using Newton interpolation method are different if the number of data selected around the vacancy value and the number of left and right data are different. Figure 2 , which is a schematic diagram of an experiment using Newton interpolation method for a data sequence with continuous missing values. Figure 2In the figure, the four black boxes correspond to four sections of code. The first section of code selects 3 data before the first missing value and 4 data after the last missing value, a total of 7 data, that is, selects all data in the data sequence except the missing values, and uses Newton interpolation to obtain the filling result, that is, the interpolation result; the second section of code selects 2 data before the first missing value and 1 data after the last missing value, a total of 3 data, and uses Newton interpolation to obtain the filling result, that is, the interpolation result; the third section of code selects 1 data before the first missing value and 4 data after the last missing value, a total of 5 data, and uses Newton interpolation to obtain the filling result, that is, the interpolation result; the fourth section of code selects 3 data before the first missing value and 2 data after the last missing value, a total of 5 data, and uses Newton interpolation to obtain the filling result, that is, the interpolation result. According to Figure 2 From the filling results of each code segment in , we can see that Newton interpolation method roughly satisfies the following rules: the more data selected around the vacant value, that is, the larger the order in Newton interpolation method (order = number of selected data - 1), the further the final interpolation result is from the real data of its nearest neighbor; conversely, the smaller the order, the closer the final interpolation result is to the real data of its nearest neighbor; the larger the difference in the number of selected data on the left and right sides, the more biased the final interpolation result will be towards the data on the side with larger number.

[0039] Based on the above rules, in the embodiment of the present invention, the difference between the reference interpolation corresponding to the i-th vacant position sequence in the data sequence to be analyzed and its adjacent real data (pressure data in the data sequence to be analyzed) is calculated to obtain the reference interpolation proximity corresponding to the i-th vacant position sequence, which is used to determine the order in the Newton interpolation method: the greater the reference interpolation proximity, the closer the reference interpolation is to its adjacent real data, and when the Newton interpolation method is used to interpolate and complete the vacant positions in the i-th vacant position sequence, the smaller the selected order is, that is, the fewer the number of data selected around the vacant position; conversely, the more the number of data selected around the vacant position. Among them, the specific method of obtaining the reference interpolation proximity corresponding to the i-th vacant position sequence is: Obtain the reference interpolation value of the first vacant position in the i-th vacant position sequence, recorded as the left reference interpolation value, and obtain the reference interpolation value of the last vacant position in the i-th vacant position sequence, recorded as the right reference interpolation value; In the data sequence to be analyzed, the pressure data of the left adjacent position of the first vacant position in the i-th vacant position sequence is obtained, recorded as the left pressure data, and the pressure data of the right adjacent position of the last vacant position in the i-th vacant position sequence is obtained, recorded as the right pressure data; The absolute value of the difference between the left pressure data and the left reference interpolation is calculated, recorded as the left difference absolute value, the absolute value of the difference between the right pressure data and the right reference interpolation is calculated, recorded as the right difference absolute value, and the reciprocal of the average value between the left difference absolute value and the right difference absolute value is calculated to obtain the reference interpolation proximity corresponding to the i-th vacant position sequence.

[0040] In one embodiment, the calculation formula for the reference interpolation proximity corresponding to the i-th vacant position sequence is:

[0041] in, represents the reference interpolation proximity corresponding to the i-th vacant position sequence, represents the left reference interpolation, Represents the left pressure data, represents the right reference interpolation, Indicates right pressure data, Represents the absolute value symbol.

[0042] It should be noted that The smaller it is, the smaller the difference between the reference interpolation value corresponding to the first vacant position in the i-th vacant position sequence and its left adjacent pressure data is in the data sequence to be analyzed. The smaller it is, the smaller the difference between the reference interpolation value corresponding to the last vacant position in the i-th vacant position sequence and its right adjacent pressure data is in the data sequence to be analyzed. The larger the value is, the closer the reference interpolation value corresponding to the i-th vacant position sequence is to its adjacent real data (pressure data in the data sequence to be analyzed), and the smaller the order is selected when Newton interpolation is used to interpolate and complete the i-th vacant position sequence.

[0043] In addition to the order affecting the interpolation results of the Newton interpolation method, the number of data selected on the left and right sides of the vacant position will also affect the interpolation results of the Newton interpolation method: the greater the difference in the number of data selected on the left and right sides, the final interpolation result will be biased towards the data on the side with a larger number. At the same time, the value of a vacant position in the i-th vacant position sequence may be a turning point in the data change trend in the data sequence to be analyzed. If the turning point is on the left side of the i-th vacant position sequence, then when using the Newton interpolation method, the amount of data selected on the left side of the i-th vacant position sequence is more than that on the right side. Conversely, if the turning point is on the right side of the i-th vacant position sequence, then when using the Newton interpolation method, the amount of data selected on the right side of the i-th vacant position sequence is more than that on the left side, so as to improve the accuracy of interpolation and completion of the i-th vacant position sequence using the Newton interpolation method. Therefore, in the embodiment of the present invention, the possible position of the turning point is found according to the change trend between the reference interpolation values ​​corresponding to each two adjacent vacant positions in the i-th vacant position sequence, and the position bias of all reference interpolation values ​​is obtained according to the position of the turning point, and then the number of data selected on the left and right sides of the vacant position is determined according to the position bias. In the embodiment of the present invention, firstly, according to the change trend between the reference interpolation values ​​corresponding to each two adjacent vacant positions in the i-th vacant position sequence, the turning reference interpolation, that is, the turning point, is found, specifically: Mapping all reference interpolation values ​​to the same scatter plot, wherein the abscissa of the scatter plot represents the vacant position and the ordinate represents the reference interpolation value; For any reference interpolation in the scatter plot, if the reference interpolation is the first reference interpolation in the scatter plot, the slope between the reference interpolation and its right adjacent reference interpolation is obtained, which is recorded as , subtract the inverse of the slope from the constant 1 to obtain the turning rate of any reference interpolation, recorded as ,Right now ; If any of the reference interpolation values ​​is the last reference interpolation value in the scatter plot, the slope between the any of the reference interpolation values ​​and its left adjacent reference interpolation value is obtained, which is recorded as , subtract the inverse of the slope from the constant 1 to obtain the turning rate of any reference interpolation, recorded as ,Right now ; If any of the reference interpolation values ​​is not a reference interpolation value other than the first reference interpolation value and the last reference interpolation value in the scatter plot, the left slope between the reference interpolation value and its left adjacent reference interpolation value is calculated, which is recorded as , and the right slope between any reference interpolation and its right adjacent reference interpolation, denoted as , subtract the inverse of the absolute value of the difference between the left slope and the right slope from the constant 1 to obtain the turning rate of any reference interpolation, which is recorded as ,Right now ; The larger the value, the more obvious the difference in the trend of any reference interpolation before and after. The larger the value is, the more likely it is that any reference interpolation is a turning point, that is, the more likely it is that any reference interpolation is a turning reference interpolation. Therefore, the turning rates of all reference interpolations are obtained, and the reference interpolation corresponding to the maximum value of all turning rates is recorded as the turning reference interpolation.

[0044] After obtaining the turning reference interpolation, the position deviation of all reference interpolations is obtained according to the vacant position corresponding to the turning reference interpolation, and then the number of data selected on the left and right sides of the vacant position is determined according to the position deviation. Specifically: According to the first vacancy position and the last vacancy position in the i-th vacancy position sequence, the vacancy center position of the i-th vacancy position sequence is obtained, the absolute value of the difference between the vacancy center position and the vacancy position corresponding to the turning reference interpolation is calculated to obtain the position difference degree, the reciprocal of the sum of the position difference degree and a constant 1 is calculated, the reciprocal is subtracted from the constant 1 to obtain the position deviation degree of all reference interpolations.

[0045] In one embodiment, the calculation formula of the position deviation is:

[0046] in, Indicates the position deviation, represents the first vacant position in the i-th vacant position sequence, represents the last vacant position in the i-th vacant position sequence, Indicates the vacant position corresponding to the turning reference interpolation, 1 indicates a constant, Represents the absolute value symbol.

[0047] It should be noted that The smaller the value, the closer the distance between the center of the vacancy and the turning reference interpolation value is. The smaller it is, the closer the number of data selected on the left and right sides of the vacant position is.

[0048] At this point, based on the data characteristics of the data sequence to be analyzed and the rules of Newton interpolation in processing continuous missing values, the reference interpolation proximity and position bias of the reference interpolation corresponding to the i-th missing position sequence are obtained, which are used to subsequently determine the known data required in the Newton interpolation method, thereby improving the accuracy of interpolation completion of the i-th missing position sequence.

[0049] Step S103, obtaining target data according to the number of vacant positions in any vacant position sequence, the reference interpolation proximity and the position deviation, and interpolating and completing the vacant positions in any vacant position sequence using Newton interpolation method according to the target data.

[0050] Through step S102, the reference interpolation proximity and position deviation corresponding to the i-th vacant position sequence are obtained. Further, the number of data required by the Newton interpolation algorithm is determined according to the reference interpolation proximity, which is recorded as the target data number. Specifically: Calculate the difference between constant 1 and the reference interpolation proximity, add constant 1 to the difference to obtain an addition result, round up the product of the addition result and the number of vacant positions in the i-th vacant position sequence to obtain the target data quantity.

[0051] In one embodiment, the target data quantity is calculated as follows:

[0052] in, Indicates the number of target data. represents the number of vacant positions in the i-th vacant position sequence, represents the reference interpolation proximity, 1 represents a constant, Indicates the round-up symbol.

[0053] It should be noted that The larger it is, the closer the reference interpolation corresponding to the i-th vacant position sequence is to its adjacent real data (pressure data in the data sequence to be analyzed). When using Newton interpolation to interpolate and complete the i-th vacant position sequence, the smaller the selected order is, and thus the smaller A is, and the smaller the number of target data is.

[0054] Furthermore, according to the positional bias, the difference in the number of selected data on the left and right sides of the i-th vacant position sequence is obtained, which is recorded as the quantity distribution difference index. Specifically: The product of the target data quantity and the position deviation degree is rounded up to obtain a quantity distribution difference index.

[0055] In one embodiment, the calculation formula of the quantity distribution difference index is:

[0056] in, represents the quantity distribution difference index, Indicates the position deviation, Indicates the number of target data. Indicates the round-up symbol.

[0057] It should be noted that The smaller it is, the closer the number of selected data on the left and right sides of the vacant position is. The smaller it is, the closer the number of data selected on the left and right sides of the i-th vacancy position sequence is.

[0058] Furthermore, according to the target data quantity and quantity distribution difference index, in the data sequence to be analyzed, the target data required for interpolating and completing the vacant positions in the i-th vacant value sequence using the Newton interpolation method is obtained, specifically: In the data sequence to be analyzed, a first preset number of data before the i-th vacant position sequence and a second preset number of data after the i-th vacant position sequence are respectively obtained, the sum of the first preset number and the second preset number is equal to the target data quantity, the positions of the first preset number of data in the data sequence to be analyzed are continuous with the first vacant position in the i-th vacant position sequence, and the positions of the second preset number of data in the data sequence to be analyzed are continuous with the last vacant position in the i-th vacant position sequence; In the i-th vacancy position sequence, if the vacancy position corresponding to the turning reference interpolation is greater than the vacancy center position, the difference between the second preset number and the first preset number is the quantity distribution difference index; if the vacancy position corresponding to the turning reference interpolation is less than the vacancy center position, the difference between the first preset number and the second preset number is the quantity distribution difference index; if the vacancy position corresponding to the turning reference interpolation is equal to the vacancy center position, the first preset number is equal to the second preset number; A first preset number of data before the i-th vacant position sequence and a second preset number of data after any vacant position sequence are recorded as target data.

[0059] Specifically, if the number of data before the i-th vacant position sequence in the data sequence to be analyzed is less than the first preset number, or the number of data after the i-th vacant position sequence in the data sequence to be analyzed is less than the second preset number, the data that can be actually obtained in the data sequence to be analyzed will be used as the target data.

[0060] After the target data is obtained, each vacant position in the i-th vacant position sequence is interpolated and completed using the Newton interpolation method according to the target data.

[0061] Step S104, interpolating and completing the vacant positions in each of the vacant position sequences to obtain a completed sequence of data to be analyzed, and performing fault warning on the water system according to the completed sequence of data to be analyzed.

[0062] According to step S103, each vacant position in the data sequence to be analyzed is interpolated and completed to obtain a completed data sequence to be analyzed. Based on the completed data sequence to be analyzed, the ARIMA algorithm is used to predict the pressure data of the water system at a future moment to obtain a predicted value of the pressure data at a future moment. According to the pressure requirement range of the fire water system, the normal pressure data is between 0.10 and 0.15 MPa. If the predicted value of the pressure data is not within the pressure requirement range, a fault warning is issued for the water system to remind the staff to conduct inspections.

[0063] In summary, the present invention obtains pressure data of a water system in a target period to obtain a data sequence to be analyzed, obtains at least one vacant position sequence in the data sequence to be analyzed, and the vacant position sequence is composed of vacant positions corresponding to continuous vacant values ​​in the data sequence to be analyzed; for any vacant position sequence, according to the similarity between the pressure data of the water system in a historical period and the data in the data sequence to be analyzed, obtains a reference interpolation value for each vacant position in the any vacant position sequence, and obtains the reference interpolation value for each vacant position in the any vacant position sequence according to the difference between the data in the data sequence to be analyzed and each reference interpolation value. The reference interpolation proximity corresponding to the missing position sequence is used to obtain the position deviation of all reference interpolations according to the change trend between each two adjacent reference interpolations; the target data is obtained according to the number of vacant positions in any vacant position sequence, the reference interpolation proximity and the position deviation, and the vacant positions in any vacant position sequence are interpolated and completed using the Newton interpolation method according to the target data; the vacant positions in each of the vacant position sequences are interpolated and completed to obtain a completed data sequence to be analyzed, and a fault warning is given to the water system according to the completed data sequence to be analyzed. Among them, according to the characteristics of the data sequence to be analyzed and historical data, the reference interpolation of each vacant position is found to make a preliminary estimate of the pressure data in each vacant position; combined with the reference interpolation proximity and position bias of the reference interpolation, and the law of Newton interpolation for interpolation and completion of vacant values, the number and distribution of target data required for Newton interpolation to interpolate and complete the vacant values ​​are determined, and then the data sequence to be analyzed is interpolated and completed using Newton interpolation, thereby improving the accuracy of interpolation and completion of continuous vacant values ​​using Newton interpolation, and predicting the fault situation of the water system according to the completed data sequence to be analyzed, thereby improving the accuracy of fault prediction for the water system.

[0064] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.

Claims

1. A sensor-based water system fault prediction management platform, characterized by: The method comprises a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the following steps when executing the computer program: Acquire pressure data of the water system in a target period to obtain a data sequence to be analyzed, and acquire at least one vacancy position sequence in the data sequence to be analyzed, wherein the vacancy position sequence is composed of vacancy positions corresponding to consecutive vacancy values ​​in the data sequence to be analyzed; For any vacant position sequence, according to the similarity between the pressure data of the water system in the historical period and the data in the data sequence to be analyzed, the reference interpolation of each vacant position in the vacant position sequence is obtained; according to the difference between the data in the data sequence to be analyzed and each reference interpolation, the reference interpolation proximity corresponding to the vacant position sequence is obtained; according to the change trend between each two adjacent reference interpolations, the position deviation of all reference interpolations is obtained; Obtaining target data according to the number of vacant positions in any vacant position sequence, the reference interpolation proximity and the position deviation, and interpolating and completing the vacant positions in any vacant position sequence using Newton interpolation method according to the target data; Interpolation is performed on the vacant positions in each of the vacant position sequences to obtain a completed sequence of data to be analyzed, and fault warning is performed on the water system according to the completed sequence of data to be analyzed.

2. The sensor-based water system fault prediction management platform according to claim 1 is characterized in that: The obtaining of a reference interpolation value for each vacant position in any vacant position sequence based on similarity between the pressure data of the water system in the historical period and the data in the data sequence to be analyzed comprises: In the historical period, obtaining a first historical data sequence corresponding to at least one historical period that is the same as the time period of the target period, and obtaining a reference index for each of the first historical data sequences according to the similarity between each of the first historical data sequences and the data sequence to be analyzed; If there is a reference index belonging to the first preset reference range among all the reference indexes, then the first historical data sequence corresponding to the maximum value is selected from all the reference indexes as the first historical reference data sequence; if all the reference indexes do not belong to the first preset reference range, but there is a reference index belonging to the second preset reference range, then the first historical data sequences corresponding to all the reference indexes belonging to the second preset reference range are taken as the target historical data sequence, and the average value of the data at the same position in all the target data sequences is calculated to obtain the first historical reference data sequence, and the first preset reference range is greater than the second preset reference range; According to the data in the data sequence to be analyzed, using a preset prediction algorithm, obtaining the prediction value corresponding to each vacant position in any vacant position sequence to form a prediction value sequence; According to the position of any vacant position sequence in the to-be-analyzed data sequence, pressure data at the same position as any vacant position sequence is acquired in the first historical reference data sequence to form a first target reference sequence corresponding to any vacant position sequence; According to the similarity between the predicted value sequence and the first target reference sequence, a change trend consistency index between the predicted value sequence and the first target reference sequence is obtained; if the change trend consistency index is greater than or equal to a preset change trend consistency index threshold, the pressure data in the first target reference sequence is used as a reference interpolation for each vacant position in any vacant position sequence.

3. The sensor-based water system fault prediction management platform according to claim 2 is characterized in that: The obtaining, according to the similarity between each of the first historical data sequences and the data sequence to be analyzed, a reference index of each of the first historical data sequences, comprises: In the data sequence to be analyzed, the data before any vacant position sequence is recorded as a left data sequence, and the data after any vacant position sequence is recorded as a right data sequence; For any first historical data sequence, according to the position of the left data sequence in the data sequence to be analyzed, a historical left data sequence is obtained in the any first historical data sequence, and according to the position of the right data sequence in the data sequence to be analyzed, a historical right data sequence is obtained in the any first historical data sequence; Calculate the reciprocal of the left DTW distance between the left data sequence and the historical left data sequence to obtain the left similarity, calculate the reciprocal of the right DTW distance between the right data sequence and the historical right data sequence to obtain the right similarity, calculate the average value between the left similarity and the right similarity, and obtain the overall similarity between any first historical data sequence and the data sequence to be analyzed; Obtaining an interval period between a historical period in which any first historical data sequence is located and the target period, performing linear normalization on the reciprocal of the interval period, and obtaining the credibility of any first historical data sequence; The overall similarity and the credibility are weightedly summed to obtain a reference index of any first historical data sequence.

4. The sensor-based water system fault prediction management platform according to claim 2 is characterized in that: The obtaining, according to the similarity between the predicted value sequence and the first target reference sequence, a change trend consistency index between the predicted value sequence and the first target reference sequence includes: Calculate the absolute value of the difference between two elements at each same position between the predicted value sequence and the first target reference sequence to obtain a mean of the absolute values ​​of the differences, and record the reciprocal of the mean of the absolute values ​​of the differences as the degree of numerical proximity between the predicted value sequence and the first target reference sequence; Calculating a DTW distance between the predicted value sequence and the first target reference sequence, and using the reciprocal of the DTW distance as a similarity index between the predicted value sequence and the first target reference sequence; The average value between the numerical proximity degree and the similarity index is calculated to obtain a change trend consistency index between the predicted value sequence and the first target reference sequence.

5. The sensor-based water system fault prediction management platform according to claim 2, characterized in that: After obtaining the reference index of each of the first historical data sequences, the method further includes: If all reference indexes do not belong to the first preset reference range and the second preset reference range, obtaining second historical data sequences of all historical periods within the historical period except the historical period where the first historical data sequence is located, and obtaining a reference index for each second historical data sequence based on the similarity between each second historical data sequence and the data sequence to be analyzed; If there is a reference index belonging to the first preset reference range or the second preset reference range among the reference indexes of all second historical data sequences, a first historical reference data sequence is obtained from all second historical data sequences; If the reference indexes of all second historical data sequences do not belong to the first preset reference range and the second preset reference range, a fault warning is issued for the water system.

6. The sensor-based water system fault prediction management platform according to claim 2, characterized in that: After obtaining the change trend consistency index between the predicted value sequence and the first target reference sequence, the method further includes: If the change trend consistency index is less than the preset change trend consistency index threshold, then obtain the third historical data sequence of all historical periods in the historical period, obtain the reference index of each of the third historical data sequences, take the third historical data sequence whose reference index belongs to the first preset range as the second historical reference data sequence, obtain the second target reference sequence in each of the second historical reference data sequences, calculate the change trend consistency index between the predicted value sequence and each of the second target reference sequences, select the second target reference sequence corresponding to the maximum value among all the change consistency indicators as the final target reference sequence, and use the pressure data in the final target reference sequence as the reference interpolation for each vacant position in any vacant position sequence.

7. The sensor-based water system fault prediction management platform according to claim 1, characterized in that: The obtaining, according to the difference between the data in the data sequence to be analyzed and each reference interpolation, the reference interpolation proximity corresponding to any vacant position sequence comprises: Obtaining a reference interpolation value of a first vacant position in any vacant position sequence, recorded as a left reference interpolation value, and obtaining a reference interpolation value of a last vacant position in any vacant position sequence, recorded as a right reference interpolation value; In the data sequence to be analyzed, the pressure data of the position adjacent to the left of the first vacant position in any vacant position sequence is obtained, recorded as the left pressure data, and the pressure data of the position adjacent to the right of the last vacant position in any vacant position sequence is obtained, recorded as the right pressure data; The absolute value of the difference between the left pressure data and the left reference interpolation is calculated, recorded as the left difference absolute value, the absolute value of the difference between the right pressure data and the right reference interpolation is calculated, recorded as the right difference absolute value, and the reciprocal of the average value between the left difference absolute value and the right difference absolute value is calculated to obtain the reference interpolation proximity corresponding to any vacant position sequence.

8. The sensor-based water system fault prediction management platform according to claim 1, characterized in that: The position deviation of all reference interpolation values ​​is obtained according to the change trend between each two adjacent reference interpolation values, including: Mapping all reference interpolation values ​​to the same scatter plot, wherein the abscissa of the scatter plot represents the vacant position and the ordinate represents the reference interpolation value; For any reference interpolation value in the scatter plot, if the any reference interpolation value is the first reference interpolation value in the scatter plot, obtain the slope between the any reference interpolation value and the reference interpolation value adjacent to its right, and subtract the reciprocal of the slope from a constant 1 to obtain a turning rate of the any reference interpolation value; If the any reference interpolation value is the last reference interpolation value in the scatter plot, obtaining the slope between the any reference interpolation value and the reference interpolation value adjacent to its left, and subtracting the reciprocal of the slope from a constant 1 to obtain a turning rate of the any reference interpolation value; If any of the reference interpolation values ​​is not a value other than the first reference interpolation value and the last reference interpolation value in the scatter plot, the left slope between the any of the reference interpolation values ​​and the reference interpolation value adjacent to its left, and the right slope between the any of the reference interpolation values ​​and the reference interpolation value adjacent to its right are calculated, and the inflection rate of the any of the reference interpolation values ​​is obtained by subtracting the reciprocal of the absolute value of the difference between the left slope and the right slope from a constant 1; Obtain the turning rates of all reference interpolations, and record the reference interpolation corresponding to the maximum value of all turning rates as the turning reference interpolation; According to the first vacancy position and the last vacancy position in any vacancy position sequence, the vacancy center position of any vacancy position sequence is obtained, the absolute value of the difference between the vacancy center position and the vacancy position corresponding to the turning reference interpolation is calculated to obtain the position difference degree, the reciprocal of the sum of the position difference degree and a constant 1 is calculated, the reciprocal is subtracted from the constant 1 to obtain the position deviation degree of all reference interpolations.

9. The sensor-based water system fault prediction management platform according to claim 8, characterized in that: The step of obtaining target data according to the number of vacant positions in any vacant position sequence, the reference interpolation proximity and the position deviation degree comprises: Calculate the difference between the constant 1 and the reference interpolation proximity, add the constant 1 and the difference to obtain an addition result, and round up the product of the addition result and the number of vacant positions in any vacant position sequence to obtain the target data quantity; The product of the target data quantity and the position deviation degree is rounded up to obtain a quantity distribution difference index, and the target data is obtained according to the target data quantity and the quantity distribution difference index.

10. The sensor-based water system fault prediction management platform according to claim 9, characterized in that: The obtaining of target data according to the target data quantity and the quantity distribution difference indicator comprises: In the data sequence to be analyzed, a first preset number of data before any vacant position sequence and a second preset number of data after any vacant position sequence are respectively obtained, the sum of the first preset number and the second preset number is equal to the target data quantity, the positions of the first preset number of data in the data sequence to be analyzed are continuous with the first vacant position in any vacant position sequence, and the positions of the second preset number of data in the data sequence to be analyzed are continuous with the last vacant position in any vacant position sequence; In any of the vacancy position sequences, if the vacancy position corresponding to the turning reference interpolation is larger than the vacancy center position, the difference between the second preset number and the first preset number is the quantity distribution difference index; if the vacancy position corresponding to the turning reference interpolation is smaller than the vacancy center position, the difference between the first preset number and the second preset number is the quantity distribution difference index; if the vacancy position corresponding to the turning reference interpolation is equal to the vacancy center position, the first preset number is equal to the second preset number; A first preset number of data before any of the vacant position sequences and a second preset number of data after any of the vacant position sequences are recorded as target data.

Images