An Industrial Situation Awareness Multi-Sensor Fusion Data Collaborative Analysis Method and System

Abstract

The present invention provides an industrial situation awareness multi-sensor fusion data collaborative analysis method and system, which relates to the field of industrial sensor technology. The specific steps include: collecting sensor data, analyzing the correlation between sensors, longitudinally analyzing data, and outputting evaluation results. The present invention performs correlation analysis on data collected by sensors on equipment, and the sensors work together to collect sensor data, analyze the correlation between each sensor, and perform longitudinal analysis on the data collected by the sensors at the same sampling point and collection window based on the correlation, and generate a fault category determination index for determining the source of the fault, thereby assisting in determining whether the equipment fault is a sensor fault, greatly improving the maintenance efficiency of the staff.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial sensors, and specifically to a method and system for collaborative analysis of multi-sensor fusion data for industrial situation awareness. Background Art

[0002] Sensors are essential components in industry. They are responsible for collecting various physical quantity information of the environment or equipment and converting it into electrical signals for data processing and analysis. There are many types of sensors, including temperature sensors, pressure sensors, humidity sensors, flow sensors, and optical sensors, etc. Each sensor is designed for a specific application field. For example, temperature sensors are widely used in HVAC systems, food processing, and automation control to ensure precise monitoring and adjustment of temperature; pressure sensors are used in the oil, gas, and chemical industries for safety monitoring and process control. In industrial automation, sensors can provide real-time feedback by continuously monitoring the operating status of equipment. Sensors are also often used to determine the status of equipment.

[0003] In industry, a single sensor is often used to detect the data of a target device. However, when a single sensor is working, faults may occur, such as data sampling being a fixed value, 0 value, error value, etc. At this time, it is necessary for the staff to manually judge whether it is the device failure that causes the abnormal sensor sampling or the sensor itself that fails.

[0004] The above information disclosed in the background art section is only used to enhance the understanding of the background of the present disclosure. Therefore, it may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0005] The purpose of the present invention is to provide a method and system for collaborative analysis of multi-sensor fusion data for industrial situation awareness to solve the problems raised in the above background art.

[0006] To achieve the above purpose, the present invention provides the following technical solutions:

[0007] A method for collaborative analysis of multi-sensor fusion data for industrial situation awareness, the specific steps include:

[0008] S1. Set multiple sensors on the device. When the device is operating stably, collect working data through the sensors, unify the dimension of the working data, and perform standardization processing on the working data to generate standard data;

[0009] S2. Process the standard data, set a collection window, and perform correlation analysis on the standard data within the collection window to generate a window variance, where the window variance is used to reflect the variance of the standard data within the set collection window;

[0010] S3. Conduct a correlation analysis on the window variance to generate sensor outliers.

[0011] S4. When the device is operating normally, pairwise comparison and analysis are performed on the working data collected by all sensors to generate a sensor correlation coefficient.

[0012] S5. Set one of the sensors as sensor A, and set the sensor with the highest correlation coefficient with it as sensor B. Conduct a correlation analysis on the outliers of sensor A and the sampling points collected by sensor B to generate a fault category determination index.

[0013] S6. Compare the fault category determination index with the threshold and output the fault determination category.

[0014] Furthermore, the sensors include a temperature sensor, a pressure sensor, and a humidity sensor. The working data is used to reflect the data collected by the k-th sensor at the sampling point i, where i is used to index the sampling points, k is used to index the sensors, the value range of k is a positive integer from 1 to n, the value range of i is a positive integer from 1 to s, and the collection frequencies of all sensors are set to be the same, and the sampling points are synchronized.

[0015] Furthermore, the dimensions of the working data are unified, and the working data collected by s sensors are all standardized to generate standard data , and the formula is as follows:

[0016] ;

[0017] The standard data is the data obtained after the working data is standardized. The standard data is used to intuitively reflect the data fluctuation degree of the sensors.

[0018] Furthermore, set the acquisition window size to t sampling points, and conduct a correlation analysis on the standard data within the acquisition window t to generate a window variance , and the formula is as follows:

[0019] ;

[0020] Among them, the sampling window is used to intercept segments of the standard data of the k-th sensor. The intercepted segments are continuous, and the intercepted size range is t sampling points. The acquisition window is numbered, and j is used to index the numbers. The window variance The variance of the standard data reflecting the j-th acquisition window The mean value of the standard data within the j-th acquisition window, where the subscript k is used to index the sensors

[0021] Furthermore, for the window variance Perform correlation analysis to generate sensor outliers , and the abnormal formula is as follows

[0022] ;

[0023] When , it indicates that the standard data collected in the j-th acquisition window of the k-th sensor has a mutation, which is an outlier and is marked as a device failure or a sensor failure

[0024] Furthermore, when the device is working properly, pairwise comparison analysis is performed on the working data collected by all sensors to generate a sensor correlation coefficient, and the sensor correlation coefficient is used to reflect the correlation between sensors. The formula is as follows

[0025] ;

[0026] Where is the data of the q-th sampling point of one of the sensors is the average value of the data sampled by this sensor is the data of the q-th sampling point of another sensor is the average value of the data sampled by this sensor is the correlation coefficient between the two selected sensors. The closer it is to 1, the higher the degree of correlation

[0027] Furthermore, the sensor with an outlier is set as the A sensor, and the corresponding sensor with the maximum sensor correlation coefficient is set as the B sensor. Perform correlation analysis on the standard data collected by the A sensor and the B sensor to generate a fault category determination index GP; among them, the outlier of the A sensor is , and the sampling point collected by the corresponding B sensor is . Perform correlation analysis on the outlier of the A sensor and the sampling point collected by the corresponding B sensor to generate a fault category determination index GP. The formula is as follows

[0028] ;

[0029] The fault category determination index GP is used to determine the fault determination of the A sensor

[0030] Furthermore, the threshold Set to 0.9r - 1.1r. When occurs, the output fault determination category is that the device has a fault. At this time, the device needs to be repaired. When is not within the threshold range, the output fault determination category is that the A sensor has a fault. At this time, the A sensor needs to be repaired.

[0031] The present invention also provides an industrial situation awareness multi - sensor fusion data collaborative analysis system for implementing an industrial situation awareness multi - sensor fusion data collaborative analysis method, including:

[0032] A sensor collaboration system for setting multiple sensors on the device. When the device is operating stably, it collects working data through the sensors, unifies the dimension of the working data, and performs standardization processing on the working data to generate standard data;

[0033] A data processing system for processing the standard data, setting a collection window, and performing correlation analysis on the standard data within the collection window to generate a window variance, where the window variance is used to reflect the variance of the standard data within the set collection window;

[0034] A window analysis system for performing correlation analysis on the window variance to generate sensor anomaly points;

[0035] A correlation coefficient analysis system for pairwise comparing and analyzing the working data collected by all sensors when the device is operating normally to generate a sensor correlation coefficient;

[0036] A fault category determination index module for setting one of the sensors as the A sensor and the sensor with the highest correlation coefficient with it as the B sensor, performing correlation analysis on the anomaly points of the A sensor and the sampling points collected by the B sensor to generate a fault category determination index;

[0037] A fault category output module for comparing the fault category determination index with the threshold and outputting a fault determination category.

[0038] Compared with the prior art, the beneficial effects of the present invention are:

[0039] By performing correlation analysis on the data collected by the sensors on the device, the sensors work together collaboratively. By collecting sensor data, analyzing the correlation between each sensor, and based on the correlation, longitudinally analyzing the data collected by the sensors for the same sampling points and collection windows to generate a fault category determination index for determining the source of the fault, thereby assisting in judging whether the device has a fault or the sensor has a fault, greatly improving the maintenance efficiency of the staff. Description of the Drawings

[0040] Figure 1 This is a schematic diagram of the overall method flow of the present invention;

[0041] Figure 2 This is a schematic diagram of the overall system flow of the present invention. Detailed implementation manners

[0042] To make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below in conjunction with specific embodiments.

[0043] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should have the ordinary meanings understood by those with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar terms used in the present invention do not represent any order, quantity or importance, but are only used to distinguish different components. The terms such as "including" or "comprising" mean that the elements or objects appearing before this word cover the elements or objects listed after this word and their equivalents, without excluding other elements or objects. The terms such as "connected" or "coupled" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The terms such as "upper", "lower", "left" and "right" are only used to represent relative positional relationships, and when the absolute position of the object being described changes, the relative positional relationship may also change accordingly.

[0044] Embodiment:

[0045] Please refer to Figure 1 , the present invention provides a technical solution:

[0046] An industrial situation awareness multi-sensor fusion data collaborative analysis method. In industry, a single sensor is often used to detect the data of target devices. However, when a single sensor is working, faults may occur, such as data sampling being a fixed value, 0 value, error value, etc. In this case, it is necessary for the staff to manually judge whether it is the device that fails and causes the sensor sampling to be abnormal, or whether the sensor itself fails. The present invention collaborates with other sensors to work together, collects sensor data, analyzes the correlation between each sensor, and uses the sensors with high correlation to associate the abnormal sensors, thereby assisting in judging whether it is the device or the sensor that fails, greatly improving the maintenance efficiency of the staff. The specific steps include:

[0047] Step 1: Set multiple sensors on the device. When the device is working stably, collect working data through the sensors, unify the dimension of the working data, and perform standardization processing on the working data to generate standard data;

[0048] The sensors include a temperature sensor, a pressure sensor, and a humidity sensor, and may also be other relevant sensors for detecting other data. The sampling frequencies of the sensors are set to the same value. At specific sampling points, all sensors will collect data at the same moment. The working data is used to reflect the data collected by the k-th sensor at the sampling point i, where i is used to index the sampling points, k is used to index the sensors, the value range of k is a positive integer from 1 to n, and the value range of i is a positive integer from 1 to s.

[0049] To facilitate the common processing of all data, it is necessary to unify the dimension of the working data and reduce its value proportionally. For the working data collected by s sensors all are standardized to generate standard data , and the formula is as follows:

[0050] ;

[0051] The standard data is the data obtained after the working data is standardized. is the variance of the data collected by the k-th sensor, is the average value of the data collected by the k-th sensor. The standard data is used to intuitively reflect the degree of data fluctuation obtained by the sensors.

[0052] Step 2: To perform partitioning processing on the standard data, a collection window is set, and correlation analysis is performed on the standard data within the collection window to generate a window variance, which is used to reflect the variance of the standard data within the set collection window;

[0053] Set the size of the collection window to t sampling points. Perform correlation analysis on the standard data within the collection window t to generate a window variance , and the formula is as follows:

[0054] ;

[0055] Among them, the sampling window is used to intercept segments of the standard data of the k-th sensor. The intercepted segments are continuous, and the intercepted size range is t sampling points. For example is a sampling window. The collection windows are numbered, and j is used to index the numbers. The window variance is used to reflect the variance of the standard data in the j-th collection window, used to reflect the mean value of the standard data within the j-th acquisition window, where the subscript k is used to index the sensors, and the window variance is used to reflect the variance of the standard data within the set acquisition window, which can reflect the degree of change in the local data collected, The larger it is, the greater the degree of data change.

[0056] Step 3: Conduct a correlation analysis on the window variance to generate sensor outliers;

[0057] For the window variance Conduct a correlation analysis to generate sensor outliers , and the abnormal formula is:

[0058] ;

[0059] When , it indicates that the standard data collected in the j-th acquisition window of the k-th sensor has a mutation, which is an outlier, and mark the sensor outlier , the sensor outlier is used to reflect the degree of change in variance between two adjacent acquisition windows. The larger the value, the more unstable the data collected by the sensor, that is, the sensor fails or the device fails, and it is marked as the device fails or the sensor fails.

[0060] Step 4: When the device is working normally, pairwise comparison and analysis are performed on the working data collected by all sensors to generate a sensor correlation coefficient;

[0061] When the device is working normally, pairwise comparison and analysis are performed on the working data collected by all sensors to generate a sensor correlation coefficient. The sensor correlation coefficient is used to reflect the correlation between sensors, and the formula is:

[0062] ;

[0063] Among them, is the data of the q-th sampling point of one of the sensors, is the average value of the data sampled by this sensor, is the data of the q-th sampling point of another sensor, is the average value of the data sampled by this sensor, is the correlation coefficient between the two selected sensors. The closer the value of r is to 1, the higher the degree of correlation, that is, the higher the degree of correlation between the data collected by the two selected sensors.

[0064] Step 5: Set one of the sensors as sensor A, and set the sensor with the highest correlation coefficient with it as sensor B. Analyze the correlation between the abnormal points of sensor A and the sampling points collected by sensor B to generate a fault category determination index; compare and analyze the data of the sensor with abnormal data with the data of the other sensor with the highest correlation with it to determine the source of the abnormal data.

[0065] The sensor with abnormal points is set as sensor A, and the correlation coefficient between sensors with the maximum value is set as sensor B. Analyze the correlation between the standard data collected by sensor A and sensor B to generate a fault category determination index GP; among them, the abnormal points of sensor A are: , and the sampling points collected by the corresponding sensor B are: . Analyze the correlation between the abnormal points of sensor A and the sampling points collected by the corresponding sensor B to generate a fault category determination index GP. The formula is as follows:

[0066] ;

[0067] The fault category determination index GP reflects the difference value of the window variance between the two sensors with the highest correlation at the same collection point, and is used to determine the fault determination of sensor A. The farther the value of GP is from 1, the greater the data difference between the two sensors, and the higher the probability of sensor failure. The closer the value of GP is to 1, the higher the probability of equipment failure.

[0068] Step 6: Compare the fault category determination index with the threshold and output the fault determination category.

[0069] The threshold is set to 0.9r - 1.1r. When , output the fault determination category as equipment failure, and at this time, the equipment needs to be repaired. When is not within the threshold range, output the fault determination category as sensor A failure, and at this time, sensor A needs to be repaired.

[0070] Refer to Figure 2 , the present invention also provides an industrial situation awareness multi-sensor fusion data collaborative analysis system for executing an industrial situation awareness multi-sensor fusion data collaborative analysis method, including:

[0071] A sensor cooperation system for setting multiple sensors on the equipment. When the equipment is operating stably, collect working data through the sensors, unify the dimension of the working data, and perform standardization processing on the working data to generate standard data;

[0072] A data processing system is used to process standard data. An acquisition window is set, and correlation analysis is performed on the standard data within the acquisition window to generate a window variance, which is used to reflect the variance of the standard data within the set acquisition window.

[0073] A window analysis system is used to perform correlation analysis on the window variance to generate sensor anomaly points.

[0074] A correlation coefficient analysis system is used to pairwise compare and analyze the working data collected by all sensors when the device is working properly to generate sensor correlation coefficients.

[0075] A fault category determination index module is used to set one sensor as sensor A and the sensor with the highest correlation coefficient with it as sensor B, and perform correlation analysis on the anomaly points of sensor A and the sampling points collected by sensor B to generate a fault category determination index.

[0076] A fault category output module is used to compare the fault category determination index with a threshold and output the fault determination category.

[0077] The above formulas are all dimensionless and take their numerical values for calculation. The formulas are obtained by software simulation of a large amount of collected data to get a formula closest to the real situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0078] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. Those skilled in the art can realize that the units and algorithm steps of the examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed by hardware or software methods depends on the specific application and design constraints of the technical solution.

[0079] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units. They can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0080] The above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in this application, and all should be covered within the protection scope of this application.

Claims

1. A collaborative analysis method for multi-sensor fusion data of industrial situation awareness, characterized in that The specific steps include: S1. Set multiple sensors on the device. When the device is operating stably, collect working data through the sensors, unify the dimension of the working data, and perform standardization processing on the working data to generate standard data; S2. Process the standard data, set a collection window, and perform correlation analysis on the standard data within the collection window to generate a window variance, where the window variance is used to reflect the variance of the standard data within the set collection window; S3. Perform correlation analysis on the window variance to generate sensor anomaly points; S4. When the device is operating normally, perform pairwise comparison analysis on the working data collected by all sensors to generate sensor correlation coefficients; S5. Set one of the sensors as sensor A, and set the sensor with the highest sensor correlation coefficient with it as sensor B. Perform correlation analysis on the anomaly points of sensor A and the sampling points collected by sensor B to generate a fault category determination index; S6. Compare the fault category determination index with the threshold value and output the fault determination category; The sensor includes a temperature sensor, a pressure sensor, and a humidity sensor, and the working data is used to reflect the data collected by the k-th sensor at the sampling point i, where i is used to index the sampling point, k is used to index the sensor, the value range of k is a positive integer from 1 to n, the value range of i is a positive integer from 1 to s, the acquisition frequencies of all sensors are set to be the same, and the sampling points are synchronized; Unify the dimension of the working data, and perform standardization processing on the working data collected by s sensors to generate standard data , and the formula is as follows: Standard data Is working data Data obtained after standardization processing, standard data Used to intuitively reflect the data fluctuation degree obtained by the sensor; Set the acquisition window size to t sampling points, and perform correlation analysis on the standard data within the acquisition window t to generate the window variance , and the formula is as follows: Among them, the sampling window is used to intercept segments of the standard data of the k-th sensor. The intercepted segments are continuous, and the size range of the intercepted segments is t sampling points. The acquisition window is numbered, and j is used to index the number. The window variance is used to reflect the variance of the standard data of the j-th acquisition window, is used to reflect the mean value of the standard data within the j-th acquisition window, where the subscript k is used to index the sensor; Perform correlation analysis on the window variance to generate sensor anomaly points , and the anomaly formula is as follows: When it indicates that the standard data collected in the j-th acquisition window of the k-th sensor has mutated, which is an outlier and is marked as a device failure or a sensor failure.

2. The collaborative analysis method for industrial situation awareness multi-sensor fusion data according to claim 1, wherein: When the device is operating normally, perform pairwise comparison analysis on the working data collected by all sensors to generate sensor correlation coefficients. The sensor correlation coefficients are used to reflect the correlation between sensors, and the formula is: wherein, is the data of the q-th sampling point of one of the sensors, is the average value of the data sampled by the sensor, is the data of the q-th sampling point of another sensor, is the average value of the data sampled by the sensor, is the correlation coefficient between the two selected sensors. The closer it is to 1, the higher the degree of correlation.

3. The industrial situation awareness multi-sensor fusion data collaborative analysis method according to claim 2, wherein: The sensor where the abnormal point appears is set as the A sensor, and the correlation coefficient between the sensors therewith The corresponding sensor with the maximum value is set as the B sensor, and the correlation analysis is performed on the standard data collected by the A sensor and the B sensor to generate a fault category determination index GP; among them, the abnormal points of the A sensor are: , and the sampling points collected by the corresponding B sensor are: , and the correlation analysis is performed on the abnormal points of the A sensor and the sampling points collected by the corresponding B sensor to generate a fault category determination index GP, and the formula therefor is: The fault category determination index GP is used to determine the fault determination of sensor A.

4. An industrial situation awareness multi-sensor fusion data collaborative analysis method according to claim 3, characterized in that: Threshold Set to 0.9r - 1.1r. When , the output fault determination category is that the device has a fault. At this time, the device needs to be repaired. When is not within the threshold range, the output fault determination category is that sensor A has a fault. At this time, sensor A needs to be repaired.

5. An industrial situation awareness multi-sensor fusion data collaborative analysis system for implementing the industrial situation awareness multi-sensor fusion data collaborative analysis method described in claim 1, characterized in that, It includes: A sensor cooperation system for setting multiple sensors on the device. When the device is operating stably, collect working data through the sensors, unify the dimension of the working data, and perform standardization processing on the working data to generate standard data; A data processing system for processing the standard data, setting a collection window, and performing correlation analysis on the standard data within the collection window to generate a window variance, where the window variance is used to reflect the variance of the standard data within the set collection window; A window analysis system for performing correlation analysis on the window variance to generate sensor anomaly points; A correlation coefficient analysis system for performing pairwise comparison analysis on the working data collected by all sensors when the device is operating normally to generate sensor correlation coefficients; A fault category determination index module for setting one of the sensors as sensor A, and setting the sensor with the highest sensor correlation coefficient with it as sensor B. Perform correlation analysis on the anomaly points of sensor A and the sampling points collected by sensor B to generate a fault category determination index; The fault category output module is used to compare the fault category determination index with the threshold value and output the fault determination category.

Images