Environment sampling device intelligent diagnosis system and method based on artificial intelligence

Abstract

The invention discloses an environment sampling device intelligent diagnosis system and method based on artificial intelligence, and relates to the technical field of device diagnos.The diagnosis method comprises the steps that an environment sampling area is planned in advance, and a plurality of sampling points are divided; sample data acquisition is performed on the sampling device and each sampling point through a plurality of sensors, and the operation state of the sampling device and the environmental conditions of the sampling points are analyzed and evaluated based on the sample data; carrying out suitability analysis on each sampling point, and carrying out abnormal factor screening diagnosis on unmatched sampling points; performing anomaly identification on the operation state of the sampling device at each sampling point, and confirming an actual anomaly reason based on the anomaly identification condition of the sampling device; on the basis of actual abnormal reasons, performing secondary sampling on the unmatched sampling points by adopting a corresponding adjustment strategy, and performing risk reminding according to a sampling evaluation result; systematic screening of abnormal reasons is realized, and the diagnosis reliability is improved.

Description

Technical Field

[0001] This invention relates to the field of device diagnostic technology, specifically to an intelligent diagnostic system and method for environmental sampling devices based on artificial intelligence. Background Technology

[0002] Currently, the environmental monitoring field widely relies on various automatic or semi-automatic sampling devices to collect air, water, and other samples. However, the selection of sampling sites has long depended on manual experience and fixed standards, lacking intelligent analysis of real-time environmental dynamics, pollution source diffusion patterns, and geographical factors, resulting in insufficient representativeness of sampling sites or blind spots in spatial coverage. Existing diagnostic technologies mostly focus on fault alarms of the equipment itself, with very little involvement in intelligent judgment of the adaptability of the sampling points; the system cannot dynamically evaluate the applicability of the sampling device under the current environmental conditions of the sampling point, nor can it proactively recommend the optimal sampling strategy based on historical data and model predictions, thus failing to provide high-quality and timely data support for environmental diagnosis and restricting the decision-making level of precise pollution control. Summary of the Invention

[0003] The purpose of this invention is to provide an intelligent diagnostic system and method for environmental sampling devices based on artificial intelligence, so as to solve the problems raised in the prior art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: an intelligent diagnostic method for an environmental sampling device based on artificial intelligence, the diagnostic method comprising: A pre-planned environmental sampling area is established, and the area is divided into several sampling points. Periodic sample data is collected by a number of pre-positioned sensors to analyze and evaluate the operating status of the sampling device and the environmental conditions of the sampling point. Based on the evaluation results, abnormal cycles of any sampling point are identified, and the health of the sampling point is analyzed according to the occurrence of abnormalities in the abnormal cycle. The suitability of the sampling point is judged by the health of the sampling point; abnormal factors of unsuitable sampling points in the abnormal cycle are screened and diagnosed. Anomalies in the operating status of the sampling device at each sampling point are identified, and the actual causes of anomalies are confirmed based on the anomaly identification results. Based on the actual causes of the anomalies, appropriate adjustment strategies were adopted for the unsuitable sampling points to be sampled again, and risk warnings were issued based on the sampling evaluation results.

[0005] Furthermore, the environmental sampling area is divided into several sampling points, including: The topographic information of the environmental sampling area is obtained by retrieving the geographic information system. A sampling topographic database is established in advance, which stores several effective topographic information. The topographic information of the environmental sampling area is compared with several effective topographic information. If the topographic information of a sub-region in the environmental sampling area matches the effective topographic information, then the sub-region is set as the sampling sub-region. The topographic terrain for environmental sampling should avoid local extreme terrain such as isolated high points, depressions, and steep slope edges as much as possible. A relatively stable and flat foothold is required to install the device, so flat surface terrain is preferred. If water body sampling is required, a straight, stable river section with uniform water flow and no eddies or stagnant water should be selected, etc. Several sampling sub-regions are obtained in the environmental sampling area. A sampling device is randomly selected in each sampling sub-region and set as a sampling point to obtain the preliminary installation layout of the sampling device in the environmental sampling area. The remaining area in the environmental sampling area is extracted. If there are still some areas of sampling sub-regions in the remaining area, the sampling device is installed in the remaining areas until the remaining area no longer contains sampling sub-regions. Several sampling points in the environmental sampling area are obtained by summarizing.

[0006] Further sample data collection and analysis evaluation include: The sampling device is equipped with several sensors, including operating condition sensors and various environmental sensors. Each of these sensors has pre-set monitoring dimensions. Whenever environmental sampling is performed at any sampling point, monitoring data for each dimension is collected at every unit time point. A unit cycle is pre-set, and all monitoring data collected in each unit cycle is summarized to generate a monitoring data set for each sampling point in each unit cycle. The purpose of arranging the operating condition sensors is to identify the operating status of the sampling device; therefore, the monitoring indicators of the operating condition sensors can include operating parameters such as device voltage, current, flow rate, pressure, and temperature. The environmental sensors collect data on the environmental conditions of the sampling environment; therefore, the monitoring indicators can include basic environmental parameters such as pollutant concentration, temperature, humidity, and wind speed. Arbitrarily select a sampling point, and arbitrarily select a monitoring data set for one unit period from the selected sampling point. Divide the selected monitoring data set into data subsets for each monitoring dimension according to the monitoring dimension. Arbitrarily select a monitoring dimension and preset a safety value for the selected monitoring dimension. Randomly select a monitoring data from the data subset. If the deviation between the selected monitoring data and the preset safety value exceeds a preset deviation threshold, the selected monitoring data is set as abnormal data. Count the number of abnormal data in the selected unit period. If the number of abnormal data exceeds a preset number threshold, the selected monitoring dimension is set as an abnormal dimension. All abnormal dimensions in any unit period are identified and summarized to obtain the abnormal dimension set and set as the corresponding evaluation result, generating the evaluation result of each sampling point in each unit period.

[0007] Further, sampling point suitability assessment and abnormal factor diagnosis include: For any sampling point, abnormal working conditions and abnormal environment dimensions are divided based on the set of abnormal dimensions within a unit period. By calculating the frequency and proportion of abnormal occurrences in each environmental dimension, the abnormal assessment value of each environmental dimension is obtained and the abnormal cycle is identified. The continuous abnormal cycle of the sampling point is identified and the expected health of the sampling point is calculated. The suitability of the sampling point is judged, and the abnormal environmental dimensions in the unsuitable sampling points are extracted to identify abnormal factors, thus completing the screening and diagnosis of abnormal factors.

[0008] Furthermore, an arbitrary sampling point is selected, and the evaluation results for each unit period of the selected sampling point are obtained; the evaluation results for each unit period are selected, the set of abnormal dimensions in the evaluation results are extracted, and each abnormal dimension in the set of abnormal dimensions is divided into an abnormal working condition dimension subset and an abnormal environment dimension subset according to the sensor source. Arbitrarily select an environment dimension. When the selected environment dimension is an abnormal environment dimension, the number of periods in each period of the selected sampling points that treat the selected environment dimension as an abnormal dimension is n, and the total number of periods of the selected sampling points is set to N. T The calculated anomaly frequency at the selected sampling points for the selected environmental dimension is p = n / N. T ; Obtain the frequency of anomalies in various environmental dimensions, based on the formula: ; Where j is a positive integer and j∈[1,a], a is the number of environment dimensions, and p j Let be the frequency of anomalies in the j-th environmental dimension; calculate the anomaly percentage β of the selected environmental dimension; because the environmental conditions of different sampling points are different, the weight percentage of each environmental dimension will also be different. By obtaining the anomaly percentage of each environmental dimension through the anomaly situation of the sampling points in different unit periods, we can accurately assess the anomaly situation and effectively distinguish different sampling points. The deviation magnitude of each abnormal environment dimension in the subset of abnormal environment dimensions is obtained, and the abnormality ratio of the i-th abnormal environment dimension is set as β. i And the deviation is f i According to the formula: ; Where c represents the number of abnormal environment dimensions; the abnormal assessment value Y for each selected unit period is calculated; and an abnormal assessment threshold Y is preset. th If Y≥Y th If so, the selected unit period will be set as the abnormal period; For each abnormal period of the selected sampling point, an abnormal period is acquired. Then, for any selected abnormal period, all preceding unit periods are extracted. If a unit period preceding and adjacent to the selected abnormal period is also an abnormal period, the extraction of preceding unit periods is continued for abnormal period judgment until the extracted unit period is no longer an abnormal period. Several adjacent and consecutive abnormal periods are set as several target abnormal periods. The abnormal period of each target abnormal period is acquired, and the abnormality evaluation value of the k-th target abnormal period is set as Y. k According to the formula: ; Among them, H start The initial health level is set as the preset value, and d is the target number of abnormal periods; the expected health level H of the selected sampling points in the selected abnormal period is calculated; a health level threshold H is preset. th If H≤H th If so, the selected sampling point is determined to be an unsuitable sampling point; When the selected sampling point is an unsuitable sampling point, several abnormal environment dimensions of the selected sampling point in each target abnormal period will be obtained to obtain the abnormal occurrence frequency of each abnormal environment dimension. If the abnormal occurrence frequency exceeds the preset occurrence frequency threshold, the corresponding abnormal environment dimension will be set as an abnormal factor to obtain the abnormal factor set of the selected sampling point.

[0009] Further, the identification of sampling device anomalies and confirmation of the actual cause of the anomaly include: Select an abnormal dimension set for any sampling point within any unit period, extract the abnormal operating condition dimension subset from the abnormal dimension set, and extract the deviation amplitude of each abnormal operating condition dimension in the abnormal operating condition dimension subset. Let the deviation amplitude of the u-th abnormal operating condition dimension be f1. u ; In the selected sampling points, several unit periods that are located before the selected unit period and are adjacent to and continuous with the selected unit period are extracted. The deviation amplitude of the u-th abnormal operating condition dimension in each of the several unit periods is obtained. If the u-th abnormal operating condition dimension is a normal operating condition dimension in a certain unit period, then the deviation amplitude of the u-th abnormal operating condition dimension in that unit period is 0. The average deviation amplitude is calculated by averaging the deviation amplitude of the u-th abnormal operating condition dimension in the several unit periods. If the deviation amplitude of the selected unit period exceeds the average deviation amplitude, then the u-th abnormal operating condition dimension is set as the target abnormal dimension, thus obtaining several target abnormal dimensions in the selected unit period. Each unit period of the u-th abnormal operating condition dimension is set as the target abnormal dimension, and the abnormal environment dimension in each target unit period is obtained. When the selected sampling point is an unsuitable sampling point, the abnormal factor set of the selected sampling point is compared with the abnormal environment dimension of each unit period. If each abnormal factor in the abnormal factor set has an abnormal environment dimension that matches it, the u-th abnormal operating condition dimension is set as a valid abnormal dimension. When the selected sampling point is not an unsuitable sampling point, all target abnormal dimensions are set as valid abnormal dimensions. The system identifies valid anomaly dimensions among several target anomaly dimensions. If all target anomaly dimensions are valid anomaly dimensions, the actual anomaly is determined to be caused by a sampling device malfunction; otherwise, the actual anomaly is determined to be caused by a sampling point malfunction.

[0010] Furthermore, adjustments will be made based on the actual causes of the anomalies, including: If a real-time unit period of any sampling point is arbitrarily selected, and the selected sampling point is an unsuitable sampling point, then the actual cause of the anomaly of the selected sampling point is identified. If the actual cause of the selected sampling point's anomaly is a sampling device malfunction, a risk alert will be sent to the sampling device at the selected sampling point. If the actual cause of the selected sampling point's anomaly is a sampling point malfunction, the sampling devices in the sampling sub-area where the selected sampling point is located will be rearranged, and environmental sampling will be performed again. The newly generated sampling points will be sampled again. If there are still sampling point anomalies, a risk alert will be sent to the sampling sub-area where the sampling point is located.

[0011] To better implement the above methods, an intelligent diagnostic system for environmental sampling devices is also proposed. The diagnostic system includes a sampling node division module, a sampling data analysis module, an abnormal sampling analysis module, an abnormal cause identification module, and a sampling strategy adjustment module. The sampling node division module is used to pre-plan an environmental sampling area and divide the environmental sampling area into several sampling points; The sampling data analysis module is used to periodically collect sample data through several pre-arranged sensors, and to analyze and evaluate the operating status of the sampling device and the environmental conditions of the sampling point. The abnormal sampling analysis module is used to identify abnormal cycles of any sampling point based on the evaluation results, analyze the health of the sampling point according to the occurrence of abnormalities in the abnormal cycle, judge the suitability of the sampling point through the health of the sampling point, and screen and diagnose abnormal factors of unsuitable sampling points in the abnormal cycle. The anomaly cause identification module is used to identify anomalies in the operating status of the sampling device at each sampling point, and to confirm the actual cause of the anomaly based on the anomaly identification results of the sampling device. The sampling strategy adjustment module is used to perform secondary sampling on unsuitable sampling points based on the actual cause of the anomaly, and to issue risk warnings based on the sampling evaluation results.

[0012] Furthermore, the abnormal sampling analysis module includes a sampling adaptation judgment unit and an abnormal diagnosis screening unit; The sampling adaptation judgment unit is used to identify abnormal cycles of any sampling point based on the evaluation results, and analyze the health of the sampling point according to the occurrence of abnormalities in the abnormal cycle, and judge the adaptability of the sampling point through the health of the sampling point; the abnormal diagnosis screening unit is used to screen and diagnose abnormal factors of unsuitable sampling points in the abnormal cycle.

[0013] Furthermore, the anomaly cause identification module includes a device status identification unit and an actual cause judgment unit; The device status identification unit is used to identify anomalies in the operating status of the sampling device at each sampling point; the actual cause judgment unit is used to confirm the actual cause of the anomaly based on the anomaly identification of the sampling device.

[0014] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention automatically identifies and lays out suitable sub-regions for sampling, overcoming the subjectivity and blind spots of manual sampling compared to the traditional method of selecting sampling points based on human experience or fixed standards, thus providing a more scientific and comprehensive data foundation for environmental monitoring. 2. This invention differs from existing technologies that only alarm for equipment malfunctions. Through multi-dimensional anomaly identification and health assessment, it can accurately distinguish between "device anomalies" and "environmental anomalies," thereby achieving systematic screening and precise attribution of the causes of anomalies and improving the comprehensiveness and reliability of diagnosis. 3. This invention achieves continuous quantitative evaluation of the suitability of sampling points through comprehensive calculation of historical and real-time data. Compared with the traditional static judgment method, it can identify unsuitable points more promptly and enhance the system's adaptability in changing environments. Attached Figure Description

[0015] Figure 1 This is a flowchart of an intelligent diagnostic method for an environment sampling device based on artificial intelligence. Figure 2 This is a schematic diagram of an intelligent diagnostic system for an environment sampling device based on artificial intelligence. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Example: Figures 1 to 2 As shown, the present invention provides an intelligent diagnostic method for an environmental sampling device based on artificial intelligence. The diagnostic method includes: A pre-planned environmental sampling area is established, and the area is divided into several sampling points. The environmental sampling area is divided into several sampling points, including: The topographic information of the environmental sampling area is obtained by retrieving the geographic information system. A sampling topographic database is established in advance, which stores several effective topographic information. The topographic information of the environmental sampling area is compared with several effective topographic information. If the topographic information of a certain sub-region in the environmental sampling area matches a certain effective topographic information, then the certain sub-region is set as the sampling sub-region. Several sampling sub-regions are obtained in the environmental sampling area. A sampling device is randomly selected in each sampling sub-region and set as a sampling point to obtain the preliminary installation layout of the sampling device in the environmental sampling area. The remaining area in the environmental sampling area is extracted. If there are still some areas of sampling sub-regions in the remaining area, the sampling device is installed in the remaining areas until the remaining area no longer contains sampling sub-regions. Several sampling points in the environmental sampling area are obtained by summarizing.

[0018] Periodic sample data is collected by a number of pre-positioned sensors to analyze and evaluate the operating status of the sampling device and the environmental conditions of the sampling point. The sample data collection and analysis evaluation includes: The sampling device is equipped with several sensors, including operating condition sensors and various environmental sensors. Several monitoring dimensions are preset for the operating condition sensors and various environmental sensors respectively. Whenever environmental sampling is carried out at any sampling point, the monitoring data of each monitoring dimension is collected at every unit time point. A unit cycle is preset, and all the monitoring data collected in each unit cycle are summarized to generate a set of monitoring data for each sampling point in each unit cycle. Arbitrarily select a sampling point, and arbitrarily select a monitoring data set for one unit period from the selected sampling point. Divide the selected monitoring data set into data subsets for each monitoring dimension according to the monitoring dimension. Arbitrarily select a monitoring dimension and preset a safety value for the selected monitoring dimension. Randomly select a monitoring data from the data subset. If the deviation between the selected monitoring data and the preset safety value exceeds a preset deviation threshold, the selected monitoring data is set as abnormal data. Count the number of abnormal data in the selected unit period. If the number of abnormal data exceeds a preset number threshold, the selected monitoring dimension is set as an abnormal dimension. All abnormal dimensions in any unit period are identified and summarized to obtain the abnormal dimension set and set as the corresponding evaluation result, generating the evaluation result of each sampling point in each unit period.

[0019] Based on the evaluation results, abnormal cycles of any sampling point are identified, and the health of the sampling point is analyzed according to the occurrence of abnormalities in the abnormal cycle. The suitability of the sampling point is judged by the health of the sampling point; abnormal factors of unsuitable sampling points in the abnormal cycle are screened and diagnosed. Among them, the sampling point suitability assessment and abnormal factor diagnosis include: For any sampling point, abnormal working conditions and abnormal environment dimensions are divided based on the set of abnormal dimensions within a unit period. By calculating the frequency and proportion of abnormal occurrences in each environmental dimension, the abnormal assessment value of each environmental dimension is obtained and the abnormal cycle is identified. The continuous abnormal cycle of the sampling point is identified and the expected health of the sampling point is calculated. The suitability of the sampling point is judged, and the abnormal environmental dimensions in the unsuitable sampling points are extracted to identify abnormal factors, thus completing the screening and diagnosis of abnormal factors.

[0020] Specifically, a sampling point is randomly selected, and the evaluation results for each unit period of the selected sampling point are obtained; the evaluation results for each unit period are randomly selected, the set of abnormal dimensions in the evaluation results are extracted, and each abnormal dimension in the set of abnormal dimensions is divided into an abnormal working condition dimension subset and an abnormal environment dimension subset according to the sensor source. Arbitrarily select an environment dimension. When the selected environment dimension is an abnormal environment dimension, the number of periods in each period of the selected sampling points that treat the selected environment dimension as an abnormal dimension is n, and the total number of periods of the selected sampling points is set to N. T The calculated anomaly frequency at the selected sampling points for the selected environmental dimension is p = n / N. T ; Obtain the frequency of anomalies in various environmental dimensions, based on the formula: ; Where j is a positive integer and j∈[1,a], a is the number of environment dimensions, and p j Let be the frequency of anomalies in the j-th environmental dimension; calculate the proportion β of anomalies in the selected environmental dimension. The specific implementation details of the above methods are as follows: The selected environmental dimension is PM2.5. Data was collected continuously for 20 cycles at a single sampling point, and anomalies were detected in 8 of those cycles. The anomaly frequency for this selected environmental dimension is p = 8 / 20 = 40%. Other environmental dimensions such as CO, temperature, and humidity were also selected, and their anomaly frequencies were found to be 30%, 20%, and 30%, respectively. Therefore, the anomaly percentage for the selected environmental dimension PM2.5 is calculated to be β = 40% / 120% = 33.3%. The deviation magnitude of each abnormal environment dimension in the subset of abnormal environment dimensions is obtained, and the abnormality ratio of the i-th abnormal environment dimension is set as β. i And the deviation is f i According to the formula: ; Where c represents the number of abnormal environment dimensions; the abnormal assessment value Y for each selected unit period is calculated; and an abnormal assessment threshold Y is preset. th If Y≥Y th If so, the selected unit period will be set as the abnormal period; The specific implementation details of the above methods are as follows: Several abnormal environmental dimensions were set as PM2.5 concentration, CO concentration, temperature, and humidity, and the abnormal percentages of each abnormal environmental dimension were obtained as 33.3%, 25%, 16.7%, and 25%, respectively. The deviation amplitudes of each abnormal environmental dimension in the selected unit period were 40%, 30%, 50%, and 40%, respectively. The abnormal assessment value of the selected unit period was calculated as Y = 33.3%×40% + 25%×30% + 16.7%×50% + 25%×40% = 13.32% + 7.5% + 8.35% + 10% = 39.17%. The preset abnormal assessment threshold was 35%. Since 39.17% > 35%, the selected unit period was set as the abnormal period. For each abnormal period of the selected sampling point, an abnormal period is acquired. Then, for any selected abnormal period, all preceding unit periods are extracted. If a unit period preceding and adjacent to the selected abnormal period is also an abnormal period, the extraction of preceding unit periods is continued for abnormal period judgment until the extracted unit period is no longer an abnormal period. Several adjacent and consecutive abnormal periods are set as several target abnormal periods. The abnormal period of each target abnormal period is acquired, and the abnormality evaluation value of the k-th target abnormal period is set as Y. k According to the formula: ; Among them, H start The initial health level is set as the preset value, and d is the target number of abnormal periods; the expected health level H of the selected sampling points in the selected abnormal period is calculated; a health level threshold H is preset. th If H≤H th If so, the selected sampling point is determined to be an unsuitable sampling point; The specific implementation details of the above methods are as follows: Five consecutive abnormal cycles were set, and the abnormal assessment values ​​of the five abnormal cycles were obtained as 5%, 10%, 13%, 20% and 39.17%, respectively. The initial health level was preset to 100. The expected health level of the selected sampling point was calculated as H = 100 × (1 - 48% - 39.17%) = 100 × 12.83% = 12.83. When the selected sampling point is an unsuitable sampling point, several abnormal environment dimensions of the selected sampling point in each target abnormal period will be obtained to obtain the abnormal occurrence frequency of each abnormal environment dimension. If the abnormal occurrence frequency exceeds the preset occurrence frequency threshold, the corresponding abnormal environment dimension will be set as an abnormal factor to obtain the abnormal factor set of the selected sampling point.

[0021] Anomalies in the operating status of the sampling device at each sampling point are identified, and the actual causes of anomalies are confirmed based on the anomaly identification results. The identification of sampling device anomalies and confirmation of the actual cause of the anomalies include: Select an abnormal dimension set for any sampling point within any unit period, extract the abnormal operating condition dimension subset from the abnormal dimension set, and extract the deviation amplitude of each abnormal operating condition dimension in the abnormal operating condition dimension subset. Let the deviation amplitude of the u-th abnormal operating condition dimension be f1. u ; In the selected sampling points, several unit periods that are located before the selected unit period and are adjacent to and continuous with the selected unit period are extracted. The deviation amplitude of the u-th abnormal operating condition dimension in each of the several unit periods is obtained. If the u-th abnormal operating condition dimension is a normal operating condition dimension in a certain unit period, then the deviation amplitude of the u-th abnormal operating condition dimension in that unit period is 0. The average deviation amplitude is calculated by averaging the deviation amplitude of the u-th abnormal operating condition dimension in the several unit periods. If the deviation amplitude of the selected unit period exceeds the average deviation amplitude, then the u-th abnormal operating condition dimension is set as the target abnormal dimension, thus obtaining several target abnormal dimensions in the selected unit period. Each unit period of the u-th abnormal operating condition dimension is set as the target abnormal dimension, and the abnormal environment dimension in each target unit period is obtained. When the selected sampling point is an unsuitable sampling point, the abnormal factor set of the selected sampling point is compared with the abnormal environment dimension of each unit period. If each abnormal factor in the abnormal factor set has an abnormal environment dimension that matches it, the u-th abnormal operating condition dimension is set as a valid abnormal dimension. When the selected sampling point is not an unsuitable sampling point, all target abnormal dimensions are set as valid abnormal dimensions. The system identifies valid anomaly dimensions among several target anomaly dimensions. If all target anomaly dimensions are valid anomaly dimensions, the actual anomaly is determined to be caused by a sampling device malfunction; otherwise, the actual anomaly is determined to be caused by a sampling point malfunction.

[0022] Based on the actual causes of the anomalies, corresponding adjustment strategies were adopted for the unsuitable sampling points to be sampled again, and risk warnings were issued based on the sampling evaluation results. Among these, adjustments based on the actual causes of the anomalies include: If a real-time unit period of any sampling point is arbitrarily selected, and the selected sampling point is an unsuitable sampling point, then the actual cause of the anomaly of the selected sampling point is identified. If the actual cause of the selected sampling point's anomaly is a sampling device malfunction, a risk alert will be sent to the sampling device at the selected sampling point. If the actual cause of the selected sampling point's anomaly is a sampling point malfunction, the sampling devices in the sampling sub-area where the selected sampling point is located will be rearranged, and environmental sampling will be performed again. The newly generated sampling points will be sampled again. If there are still sampling point anomalies, a risk alert will be sent to the sampling sub-area where the sampling point is located.

[0023] An intelligent diagnostic system for an environmental sampling device includes a sampling node division module, a sampling data analysis module, an abnormal sampling analysis module, an abnormal cause identification module, and a sampling strategy adjustment module. The sampling node division module is used to pre-plan an environmental sampling area and divide the environmental sampling area into several sampling points; The sampling data analysis module is used to periodically collect sample data through several pre-arranged sensors, and to analyze and evaluate the operating status of the sampling device and the environmental conditions of the sampling point. The abnormal sampling analysis module is used to identify abnormal cycles of any sampling point based on the evaluation results, analyze the health of the sampling point according to the occurrence of abnormalities in the abnormal cycle, judge the suitability of the sampling point through the health of the sampling point, and screen and diagnose abnormal factors of unsuitable sampling points in the abnormal cycle. The anomaly cause identification module is used to identify anomalies in the operating status of the sampling device at each sampling point, and to confirm the actual cause of the anomaly based on the anomaly identification results of the sampling device. The sampling strategy adjustment module is used to perform secondary sampling on unsuitable sampling points based on the actual cause of the anomaly, and to issue risk warnings based on the sampling evaluation results.

[0024] The abnormal sampling analysis module includes a sampling adaptation judgment unit and an abnormal diagnosis screening unit. The sampling adaptation judgment unit is used to identify abnormal cycles of any sampling point based on the evaluation results, and analyze the health of the sampling point according to the occurrence of abnormalities in the abnormal cycle, and judge the adaptability of the sampling point through the health of the sampling point; the abnormal diagnosis screening unit is used to screen and diagnose abnormal factors of unsuitable sampling points in the abnormal cycle.

[0025] The abnormal cause identification module includes a device status identification unit and an actual cause judgment unit; The device status identification unit is used to identify anomalies in the operating status of the sampling device at each sampling point; the actual cause judgment unit is used to confirm the actual cause of the anomaly based on the anomaly identification of the sampling device.

[0026] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. An artificial intelligence-based intelligent diagnosis method for an environmental sampling device, characterized in that: The diagnostic method comprises: pre-planning an environmental sampling area, and dividing the environmental sampling area into a plurality of sampling points; periodic sample data collection is performed through a plurality of sensors arranged in advance, and the running state of the sampling device and the environmental condition of the sampling point are analyzed and evaluated; based on the evaluation result, the abnormal period of any sampling point is identified, and the health degree of the sampling point is analyzed according to the abnormal occurrence in the abnormal period, the adaptability of the sampling point is judged through the health degree of the sampling point, and the abnormal factors in the abnormal period of the non-adaptive sampling point are screened and diagnosed; the running state of the sampling device at each sampling point is identified, and the actual abnormal reason is confirmed based on the abnormal identification of the sampling device; based on the actual abnormal reason, corresponding adjustment strategy is taken for the secondary sampling of the non-adaptive sampling point, and risk warning is given according to the sampling evaluation result. 2.The intelligent diagnosis method of an environment sampling device based on artificial intelligence according to claim 1, characterized in that, A plurality of sampling points are divided in the environmental sampling area, which comprises: the terrain information of the environmental sampling area is obtained by calling the geographic information system, a sampling terrain database is pre-established, and a plurality of effective terrain information is stored in the sampling terrain database; the terrain information of the environmental sampling area is compared with the plurality of effective terrain information, if the terrain information of a certain sub-region in the environmental sampling area meets a certain effective terrain information, the certain sub-region is set as a sampling sub-region; a plurality of sampling sub-regions in the environmental sampling area are obtained, a sampling device is installed at a random position in each sampling sub-region and set as a sampling point, the preliminary installation layout of the sampling device in the environmental sampling area is obtained, the remaining area in the environmental sampling area is extracted, if there is still part of the area of the sampling sub-region in the remaining area, the sampling device is continuously installed in the part of the area, until the remaining area no longer contains the sampling sub-region, and the plurality of sampling points of the environmental sampling area are obtained. 3.The intelligent diagnosis method of the environment sampling device based on artificial intelligence according to claim 1, characterized in that: Sample data collection and analysis evaluation comprises: a plurality of sensors arranged on the sampling device include working condition sensors and various environmental sensors, a plurality of monitoring dimensions are pre-set for each working condition sensor and each environmental sensor, monitoring data of each monitoring dimension is collected every unit time point when environmental sampling is performed at any sampling point, a unit period is pre-set, all monitoring data collected in each unit period is summarized to generate a monitoring data set of each sampling point in each unit period; an arbitrary sampling point is selected, and a monitoring data set of an arbitrary unit period is selected from the selected sampling point, the selected monitoring data set is divided into data subsets of each monitoring dimension according to the monitoring dimension; an arbitrary monitoring dimension is selected, a safety value is pre-set for the selected monitoring dimension, an arbitrary monitoring data is selected from the data subset, if the deviation amplitude between the selected monitoring data and the pre-set safety value exceeds the pre-set deviation amplitude threshold, the selected monitoring data is set as abnormal data; the number of abnormal data in the selected unit period is counted, if the number of abnormal data exceeds the pre-set number threshold, the selected monitoring dimension is set as an abnormal dimension; All abnormal dimensions in any unit cycle are identified and summarized to obtain an abnormal dimension set and set as a corresponding evaluation result, and the evaluation result of each sampling point in each unit cycle is generated. 4.The intelligent diagnosis method of an environment sampling device based on artificial intelligence according to claim 1, characterized in that: The sampling point adaptability judgment and abnormal factor diagnosis include: For any sampling point, the abnormal condition and abnormal environment dimension are divided according to the abnormal dimension set in the unit cycle, the abnormal occurrence frequency and abnormal proportion of each environment dimension are calculated, the abnormal evaluation value of each environment dimension is analyzed and obtained, and the abnormal cycle is identified; the continuous abnormal cycle of the sampling point is identified and the expected health degree of the sampling point is calculated, the adaptability of the sampling point is judged, the abnormal environment dimension in the inadaptable sampling point is extracted for abnormal factor identification, and the abnormal factor screening diagnosis is completed.

5. The intelligent diagnosis method of the environment sampling device based on artificial intelligence according to claim 4, characterized in that: An arbitrary sampling point is selected, and the evaluation results of each unit cycle in the selected sampling point are obtained; an evaluation result of an arbitrary unit cycle is selected, the abnormal dimension set in the evaluation result is extracted, and each abnormal dimension in the abnormal dimension set is divided into an abnormal condition dimension sub-set and an abnormal environment dimension sub-set according to the sensor source; Optionally, one environmental dimension is selected. When the selected environmental dimension is an abnormal environmental dimension, the number of periods in which the selected environmental dimension is regarded as abnormal at the selected sampling point is n, and the total number of periods of the selected sampling point is N T . The abnormal occurrence frequency of the selected environmental dimension at the selected sampling point is calculated as p = n / N T . The abnormal occurrence frequencies of the various environmental dimensions are obtained, and the formula is: ； wherein j is a positive integer and j∈[1,a], a is the number of environmental dimensions, p j is the abnormality occurrence frequency of the jth environmental dimension; the abnormality proportion β of the selected environmental dimension is calculated; The deviation amplitudes of each abnormal environment dimension in the abnormal environment dimension subset are obtained, and the abnormal proportion of the i-th abnormal environment dimension is set as β i , and the deviation amplitude is f i , according to the formula: ； Wherein, c is the number of abnormal environment dimensions; the abnormal evaluation value Y of the selected unit period is calculated; an abnormal evaluation threshold Y is preset th If Y≥Y th , the selected unit period is set as an abnormal period; The selected sampling points are acquired for each abnormal period. An arbitrary abnormal period is selected. Each unit period before the selected abnormal period is extracted. If a unit period located before the selected abnormal period and adjacent to the selected abnormal period is also an abnormal period, the previous unit period is continuously extracted for abnormal period judgment until the extracted unit period is not an abnormal period. The extracted several adjacent and continuous abnormal periods are set as several target abnormal periods. The abnormal period of each target abnormal period is acquired. The abnormal evaluation value of the kth target abnormal period is set as Y k , according to the formula: ； H start is a preset initial health degree, d is a target number of abnormal periods; the expected health degree H of the selected sampling point in the selected abnormal period is calculated; a health degree threshold H th is preset, and if H≤H th , the selected sampling point is determined as an unsuitable sampling point. When the selected sampling point is an inadaptive sampling point, a plurality of abnormal environment dimensions in each target abnormal cycle of the selected sampling point are obtained, the abnormal occurrence frequency of each abnormal environment dimension is obtained, and if the abnormal occurrence frequency exceeds a preset occurrence frequency threshold, the corresponding abnormal environment dimension is set as an abnormal factor to obtain an abnormal factor set of the selected sampling point.

6. The intelligent diagnosis method for an environment sampling device based on artificial intelligence according to claim 5, characterized in that: Sampling device abnormality identification and actual abnormal reason confirmation include: An arbitrary sampling point is selected in an arbitrary unit cycle, and an abnormal dimension set is selected, and an abnormal working condition dimension subset in the abnormal dimension set is extracted, and the deviation amplitude of each abnormal working condition dimension in the abnormal working condition dimension subset is extracted, and the deviation amplitude of the u-th abnormal working condition dimension is set as f1 u ; In the selected sampling point, a plurality of unit cycles located before the selected unit cycle and adjacent and continuous to the selected unit cycle are extracted, the deviation amplitude of the u-th abnormal condition dimension in each of the plurality of unit cycles is obtained, and if the u-th abnormal condition dimension in a certain unit cycle is a normal condition dimension, the deviation amplitude of the u-th abnormal condition dimension in the certain unit cycle is obtained as 0; the average deviation amplitude is calculated by averaging the deviation amplitudes of the u-th abnormal condition dimension in the plurality of unit cycles, and if the deviation amplitude of the selected unit cycle exceeds the average deviation amplitude, the u-th abnormal condition dimension is set as a target abnormal dimension to obtain a plurality of target abnormal dimensions of the selected unit cycle; Each unit cycle in which the u-th abnormal condition dimension is the target abnormal dimension is set as a target unit cycle, and the abnormal environment dimension in each target unit cycle is obtained; when the selected sampling point is an inadaptive sampling point, the abnormal factor set of the selected sampling point is compared with the abnormal environment dimension of each unit cycle, and if each abnormal factor in the abnormal factor set matches the abnormal environment dimension, the u-th abnormal condition dimension is set as an effective abnormal dimension; when the selected sampling point is not an inadaptive sampling point, all target abnormal dimensions are set as effective abnormal dimensions; Identify the effective abnormal dimension in several target abnormal dimensions, when all target abnormal dimensions are effective abnormal dimensions, then determine that the actual abnormal reason is sampling device abnormality, otherwise, determine that the actual abnormal reason is sampling point abnormality.

7. The intelligent diagnosis method for an environment sampling device based on artificial intelligence according to claim 6, characterized in that: Based on the actual abnormal reason, corresponding adjustment is carried out, including: Select an arbitrary sampling point real-time unit cycle, when the selected sampling point is an unsuitable sampling point, identify the actual abnormal reason of the selected sampling point; If the actual abnormal reason of the selected sampling point is sampling device abnormality, send a risk warning to the sampling device of the selected sampling point, if the actual abnormal reason of the selected sampling point is sampling point abnormality, re-arrange the sampling devices in the sampling sub-area where the selected sampling point is located, and re-sample the new sampling points generated by re-sampling, if there is still sampling point abnormality, send a risk warning to the sampling sub-area.

8. An environmental sampling device intelligent diagnosis system for performing an artificial intelligence-based environmental sampling device intelligent diagnosis method according to any one of claims 1-7, characterized in that: The diagnostic system comprises a sampling node division module, a sampling data analysis module, an abnormal sampling analysis module, an abnormal reason identification module and a sampling strategy adjustment module; The sampling node division module is used for pre-planning an environmental sampling area, and dividing the environmental sampling area into several sampling points; The sampling data analysis module is used for periodically collecting sample data through the pre-arranged sensors, and analyzing and evaluating the sampling device running state and the sampling point environment condition; The abnormal sampling analysis module is used for identifying the abnormal period of any sampling point based on the evaluation result, and analyzing the sampling point health degree according to the abnormal occurrence in the abnormal period, and judging the adaptability of the sampling point through the sampling point health degree; The abnormal sampling analysis module is used for identifying the abnormal period of any sampling point based on the evaluation result, and analyzing the sampling point health degree according to the abnormal occurrence in the abnormal period, and judging the adaptability of the sampling point through the sampling point health degree; The abnormal reason identification module is used for identifying the abnormal state of the sampling device at each sampling point, and confirming the actual abnormal reason based on the abnormal identification of the sampling device; The sampling strategy adjustment module is used for taking corresponding adjustment strategy for the unsuitable sampling point based on the actual abnormal reason, and giving a risk warning according to the sampling evaluation result.

9. The intelligent diagnostic system for an environmental sampling device of claim 8, wherein: The abnormal sampling analysis module comprises a sampling adaptation judgment unit and an abnormal diagnosis screening unit; The sampling adaptation judgment unit is used for identifying the abnormal period of any sampling point based on the evaluation result, and analyzing the sampling point health degree according to the abnormal occurrence in the abnormal period, and judging the adaptability of the sampling point through the sampling point health degree; The abnormal diagnosis screening unit is used for screening and diagnosing the abnormal factors of the unsuitable sampling point in the abnormal period.

10. The intelligent diagnostic system for an environmental sampling device of claim 8, wherein: The abnormal reason identification module comprises a device state identification unit and an actual reason judgment unit; The device state identification unit is used for identifying the abnormal state of the sampling device at each sampling point, and the actual reason judgment unit is used for confirming the actual abnormal reason based on the abnormal identification of the sampling device.

Images