Humidity sensor location optimization method and system based on wireless sensor network

Abstract

The present invention discloses a humidity sensor position optimization method and system based on a wireless sensor network. By constructing a humidity monitoring model based on three-dimensional visualization, humidity sensor distribution information is imported. Within a preset time interval T, humidity monitoring area data is collected. After the monitoring data is serialized, it is imported into an LSTM prediction model to generate prediction data. In combination with the monitoring and prediction data, an isolated forest algorithm is used to identify abnormal data sets and mark the sensors to be tested. For abnormal sensors, adjacent sensors are marked by a monitoring model, integrated into sample data, and linear regression is used to judge the linear relationship. If there is no linear relationship, it is marked as abnormal. Finally, the abnormal equipment is located in the monitoring model and the data is analyzed, a maintenance plan is formulated and the sensor position is optimized. Through the present invention, it is possible to dynamically, quickly and accurately evaluate abnormal sensors, and quickly locate abnormal sensors, effectively improve the accuracy and efficiency of humidity monitoring, and reduce human participation.

Description

Technical Field

[0001] The present invention relates to the field of sensor data analysis, and more specifically, to a humidity sensor position optimization method and system based on a wireless sensor network. Background Art

[0002] In industrial automation, humidity control is crucial to ensuring product quality, optimizing production processes, and saving energy. However, traditional humidity monitoring methods often have problems such as slow response speed, low accuracy, and weak processing capabilities, which are difficult to meet the needs of modern industry. In addition, traditional humidity sensor monitoring is often based on simple numerical analysis and relies on manual inspection. It lacks predictive analysis, abnormal trend assessment, and location optimization of multiple humidity sensors. It lacks comprehensive analysis of multi-sensor data, making it difficult to accurately and quickly evaluate and diagnose abnormal sensors, and it is difficult to adapt to the high-precision humidity control application needs of industries such as industry and chemical industry. Summary of the invention

[0003] The present invention overcomes the defects of the prior art and proposes a humidity sensor position optimization method and system based on a wireless sensor network.

[0004] A first aspect of the present invention provides a humidity sensor location optimization method based on a wireless sensor network, comprising:

[0005] Based on the humidity monitoring area, a monitoring model based on 3D visualization is constructed to obtain the distribution information of humidity sensors and import the distribution information into the monitoring model;

[0006] In one analysis cycle, the humidity monitoring area is monitored for humidity. Based on the preset time interval T, data of each humidity sensor is collected at multiple time points to obtain monitoring data. The data collection is based on a wireless sensor network.

[0007] Based on the time dimension, the monitoring data is serialized to obtain the sequence data of each humidity sensor. The sequence data is imported into the LSTM-based prediction model for sequence prediction to generate the prediction data of each humidity sensor.

[0008] All monitoring data are associated with prediction data to form a monitoring data set. Based on the isolation forest algorithm, abnormal data points in the monitoring data set are judged and abnormal data sets are screened out.

[0009] Through the abnormal data set, the corresponding sensor to be tested is marked, and a sensor to be tested is selected. Based on the monitoring model, the sensor to be tested is used as the center to mark the adjacent sensors of the sensor to be tested, and the monitoring data of the adjacent sensors and the monitoring data of the sensor to be tested are integrated into sample data. The sample data is linearly fitted based on linear regression to determine whether there is a linear relationship. If there is no linear relationship, the sensor to be tested is marked as an abnormal sensor;

[0010] Through abnormal sensors and corresponding monitoring data, abnormal equipment location and abnormal data analysis are performed in the monitoring model to formulate sensor equipment maintenance plans;

[0011] Through abnormal equipment positioning and sensor equipment maintenance plans, the position of abnormal sensors is optimized, and the position of abnormal sensors is reset based on the monitoring feedback data after the implementation of the maintenance plan.

[0012] In this solution, based on the humidity monitoring area, a monitoring model based on three-dimensional visualization is constructed to obtain the distribution information of humidity sensors and import the distribution information into the monitoring model, specifically:

[0013] Acquire spatial information of the humidity monitoring area, wherein the spatial information includes area, contour and monitoring range information of the area;

[0014] According to the spatial information, a monitoring model based on three-dimensional visualization is constructed;

[0015] Import the distribution information into the monitoring model for humidity visualization monitoring.

[0016] In this solution, within one analysis cycle, the humidity monitoring area is monitored for humidity, and based on a preset time interval T, data is collected from each humidity sensor at multiple time points to obtain monitoring data. The data collection is based on a wireless sensor network, specifically:

[0017] Set an analysis week and a preset time interval T, and divide the analysis period into multiple time points based on the preset time interval T;

[0018] Data from each humidity sensor is collected at multiple time points to obtain monitoring data.

[0019] In this solution, the monitoring data is serialized based on the time dimension to obtain the sequence data of each humidity sensor, and the sequence data is imported into the LSTM-based prediction model for sequence prediction to generate the prediction data of each humidity sensor, specifically:

[0020] Using a humidity sensor as the analysis unit, the monitoring data is preprocessed by data cleaning and standardization;

[0021] Based on the time dimension, the monitoring data is serialized to form sequence data;

[0022] Build a prediction model based on LSTM, set the prediction period to the preset time interval T, import the sequence data into the prediction model for sequence prediction, and generate prediction sequence data;

[0023] The predicted sequence data is analyzed to obtain the predicted data.

[0024] In this solution, all monitoring data are associated with prediction data to form a monitoring data set. Based on the isolation forest algorithm, abnormal data points in the monitoring data set are judged and abnormal data sets are screened out. Specifically:

[0025] Based on each humidity sensor, the corresponding monitoring data is associated with the prediction data to form a monitoring data set;

[0026] Based on the isolation forest model, set the number of trees, the estimated proportion of abnormal data, and the random number seed parameters;

[0027] Import the monitoring data set into the isolation forest model for data prediction and analysis, make anomaly judgments for each data point, and generate a judgment list;

[0028] According to the judgment list, abnormal data is extracted from the monitoring data set to generate an abnormal data set.

[0029] In this solution, the abnormal data set is used to mark the corresponding sensor to be tested, and a sensor to be tested is selected. Based on the monitoring model, the sensor to be tested is centered, and the adjacent sensors of the sensor to be tested are marked. The monitoring data of the adjacent sensors and the monitoring data of the sensor to be tested are integrated into sample data. The linear fitting of the sample data is performed based on linear regression to determine whether there is a linear relationship. If there is no linear relationship, the sensor to be tested is marked as an abnormal sensor. Specifically:

[0030] Through the abnormal data set, the humidity sensor corresponding to the data collection is analyzed and marked as the sensor to be tested;

[0031] A sensor to be tested is selected, and in the monitoring model, humidity sensors within a preset radius of the sensor to be tested are marked with the sensor to be tested as the center, and multiple adjacent sensors are selected from the humidity sensors within the preset radius according to the spatial relationship, so that the multiple adjacent sensors are continuous with the sensor to be tested in spatial position;

[0032] Integrate the monitoring data of adjacent sensors and the monitoring data of the sensor to be tested into sample data, and sort the sample data according to the continuity of spatial position;

[0033] Based on the linear regression algorithm, the sample data is linearly fitted and the determination coefficient after fitting is calculated to determine whether the determination coefficient is less than a predetermined threshold. If so, it is determined that there is no linearity in the sample data and the sensor to be tested is marked as an abnormal sensor.

[0034] In this solution, the abnormal sensors and corresponding monitoring data are used to locate abnormal equipment and analyze abnormal data in the monitoring model, and a sensor equipment maintenance plan is formulated, specifically:

[0035] Based on the abnormal sensor and the corresponding monitoring data, the abnormal equipment is located in the monitoring model, the sensor abnormality is evaluated in combination with the prediction data associated with the abnormal sensor, an abnormality report is generated, and a sensor equipment maintenance plan is formulated.

[0036] The second aspect of the present invention further provides a humidity sensor location optimization system based on a wireless sensor network, the system comprising: a memory, a processor, the memory comprising a humidity sensor location optimization program based on a wireless sensor network, the humidity sensor location optimization program based on a wireless sensor network when executed by the processor implements the following steps:

[0037] Based on the humidity monitoring area, a monitoring model based on 3D visualization is constructed to obtain the distribution information of humidity sensors and import the distribution information into the monitoring model;

[0038] In one analysis cycle, the humidity monitoring area is monitored for humidity. Based on the preset time interval T, data of each humidity sensor is collected at multiple time points to obtain monitoring data. The data collection is based on a wireless sensor network.

[0039] Based on the time dimension, the monitoring data is serialized to obtain the sequence data of each humidity sensor. The sequence data is imported into the LSTM-based prediction model for sequence prediction to generate the prediction data of each humidity sensor.

[0040] All monitoring data are associated with prediction data to form a monitoring data set. Based on the isolation forest algorithm, abnormal data points in the monitoring data set are judged and abnormal data sets are screened out.

[0041] Through the abnormal data set, the corresponding sensor to be tested is marked, and a sensor to be tested is selected. Based on the monitoring model, the sensor to be tested is used as the center to mark the adjacent sensors of the sensor to be tested, and the monitoring data of the adjacent sensors and the monitoring data of the sensor to be tested are integrated into sample data. The sample data is linearly fitted based on linear regression to determine whether there is a linear relationship. If there is no linear relationship, the sensor to be tested is marked as an abnormal sensor;

[0042] Through abnormal sensors and corresponding monitoring data, abnormal equipment location and abnormal data analysis are performed in the monitoring model to formulate sensor equipment maintenance plans;

[0043] Through abnormal equipment positioning and sensor equipment maintenance plans, the position of abnormal sensors is optimized, and the position of abnormal sensors is reset based on the monitoring feedback data after the implementation of the maintenance plan.

[0044] The third aspect of the present invention also provides a computer-readable storage medium, which includes a humidity sensor location optimization program based on a wireless sensor network. When the humidity sensor location optimization program based on a wireless sensor network is executed by a processor, the steps of the humidity sensor location optimization method based on a wireless sensor network as described in any of the above items are implemented.

[0045] The present invention discloses a humidity sensor position optimization method and system based on a wireless sensor network. By constructing a humidity monitoring model based on three-dimensional visualization, humidity sensor distribution information is imported. Within a preset time interval T, humidity monitoring area data is collected. After the monitoring data is serialized, it is imported into an LSTM prediction model to generate prediction data. In combination with the monitoring and prediction data, an isolated forest algorithm is used to identify abnormal data sets and mark the sensors to be tested. For abnormal sensors, adjacent sensors are marked by a monitoring model, integrated into sample data, and linear regression is used to judge the linear relationship. If there is no linear relationship, it is marked as abnormal. Finally, the abnormal equipment is located in the monitoring model and the data is analyzed, a maintenance plan is formulated and the sensor position is optimized. Through the present invention, it is possible to dynamically, quickly and accurately evaluate abnormal sensors, and quickly locate abnormal sensors, effectively improve the accuracy and efficiency of humidity monitoring, and reduce human participation. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 A flow chart of a humidity sensor location optimization method based on a wireless sensor network according to the present invention is shown;

[0047] Figure 2 A block diagram of a humidity sensor location optimization system based on a wireless sensor network according to the present invention is shown. DETAILED DESCRIPTION

[0048] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0049] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0050] Figure 1 The flowchart of a humidity sensor location optimization method based on a wireless sensor network of the present invention is shown.

[0051] like Figure 1 As shown, the first aspect of the present invention provides a humidity sensor location optimization method based on a wireless sensor network, comprising:

[0052] S102, based on the humidity monitoring area, construct a monitoring model based on three-dimensional visualization, obtain humidity sensor distribution information, and import the distribution information into the monitoring model;

[0053] S104, within an analysis cycle, humidity monitoring is performed on the humidity monitoring area, and data is collected from each humidity sensor at multiple time points based on a preset time interval T, and monitoring data is obtained, wherein the data collection is based on a wireless sensor network;

[0054] S106, based on the time dimension, serializing the monitoring data to obtain sequence data of each humidity sensor, importing the sequence data into the LSTM-based prediction model for sequence prediction, and generating prediction data of each humidity sensor;

[0055] S108, associating all monitoring data with prediction data to form a monitoring data set, judging abnormal data points in the monitoring data set based on an isolation forest algorithm, and filtering out abnormal data sets;

[0056] S110, using the abnormal data set, marking the corresponding sensor to be tested, selecting a sensor to be tested, marking the adjacent sensors of the sensor to be tested based on the monitoring model and with the sensor to be tested as the center, integrating the monitoring data of the adjacent sensors with the monitoring data of the sensor to be tested into sample data, performing linear fitting of the sample data based on linear regression to determine whether there is a linear relationship, and if there is no linear relationship, marking the sensor to be tested as an abnormal sensor;

[0057] S112, using abnormal sensors and corresponding monitoring data, abnormal equipment positioning and abnormal data analysis are performed in the monitoring model, and a sensor equipment maintenance plan is formulated;

[0058] S114 optimizes the position of the abnormal sensor through abnormal equipment positioning and sensor equipment maintenance plan, and resets the position of the abnormal sensor based on the monitoring feedback data after the maintenance plan is implemented.

[0059] It should be noted that there are generally multiple adjacent sensors, and the data collection is based on a wireless sensor network. The humidity sensor and the analysis platform communicate and transmit data through the network by building wireless sensors. The network nodes include analysis terminals, user terminals, and sensor terminals.

[0060] By locating the abnormal equipment and maintaining the sensor equipment, the position of the abnormal sensor is optimized, and the position of the abnormal sensor is reset based on the monitoring feedback data after the maintenance plan is implemented. By locating the abnormal equipment, the degree of data deviation is analyzed, and the position of the abnormal sensor is reset through the monitoring model and related distribution information. The reset position is within a feasible position range. The feedback data can be used to determine the improvement of the abnormality after resetting the equipment position or maintaining the equipment, which helps to improve the utilization rate of the sensor.

[0061] According to an embodiment of the present invention, based on the humidity monitoring area, a monitoring model based on three-dimensional visualization is constructed, the humidity sensor distribution information is obtained, and the distribution information is imported into the monitoring model, specifically:

[0062] Acquire spatial information of the humidity monitoring area, wherein the spatial information includes area, contour and monitoring range information of the area;

[0063] According to the spatial information, a monitoring model based on three-dimensional visualization is constructed;

[0064] Import the distribution information into the monitoring model for humidity visualization monitoring.

[0065] It should be noted that the monitoring model is used to visualize the distribution of sensors and humidity value monitoring distribution, which facilitates sensor management in the area.

[0066] According to an embodiment of the present invention, within one analysis cycle, the humidity monitoring area is monitored for humidity, and based on a preset time interval T, data is collected from each humidity sensor at multiple time points to obtain monitoring data, specifically:

[0067] Set an analysis week and a preset time interval T, and divide the analysis period into multiple time points based on the preset time interval T;

[0068] Data from each humidity sensor is collected at multiple time points to obtain monitoring data.

[0069] It should be noted that the monitoring data generally includes information such as humidity value, sensor temperature, sensor working status, etc. T is a default preset value. The larger T is, the fewer corresponding time points and the less data volume. Each humidity sensor corresponds to independent monitoring data.

[0070] According to an embodiment of the present invention, the monitoring data is serialized based on the time dimension to obtain the sequence data of each humidity sensor, and the sequence data is imported into the LSTM-based prediction model for sequence prediction to generate the prediction data of each humidity sensor, specifically:

[0071] Using a humidity sensor as the analysis unit, the monitoring data is preprocessed by data cleaning and standardization;

[0072] Based on the time dimension, the monitoring data is serialized to form sequence data;

[0073] Build a prediction model based on LSTM, set the prediction period to the preset time interval T, import the sequence data into the prediction model for sequence prediction, and generate prediction sequence data;

[0074] The predicted sequence data is analyzed to obtain the predicted data.

[0075] It should be noted that the monitoring data is data collected based on multiple time points, has a certain sequence, and is suitable for analysis by the LSTM prediction model, which can quickly and accurately predict and analyze the data. Each humidity sensor corresponds to corresponding prediction data.

[0076] It can be understood here that T is an adjustable time interval. With respect to prediction accuracy and prediction requirements, adjusting the size of T can simultaneously regulate the collection density of the original data, thereby adjusting the prediction accuracy.

[0077] According to an embodiment of the present invention, all monitoring data are associated with prediction data to form a monitoring data set, and based on the isolation forest algorithm, abnormal data points of the monitoring data set are judged, and abnormal data sets are screened out, specifically:

[0078] Based on each humidity sensor, the corresponding monitoring data is associated with the prediction data to form a monitoring data set;

[0079] Based on the isolation forest model, set the number of trees, the estimated proportion of abnormal data, and the random number seed parameters;

[0080] Import the monitoring data set into the isolation forest model for data prediction and analysis, make anomaly judgments for each data point, and generate a judgment list;

[0081] According to the judgment list, abnormal data is extracted from the monitoring data set to generate an abnormal data set.

[0082] It should be noted that the judgment list stores the judgment result of each data point.

[0083] According to an embodiment of the present invention, the abnormal data set is used to mark the corresponding sensor to be tested, a sensor to be tested is selected, and based on the monitoring model, the sensor to be tested is centered on the adjacent sensors of the sensor to be tested, and the monitoring data of the adjacent sensors and the monitoring data of the sensor to be tested are integrated into sample data, and the linear fitting of the sample data is performed based on linear regression to determine whether there is a linear relationship. If there is no linear relationship, the sensor to be tested is marked as an abnormal sensor, specifically:

[0084] Through the abnormal data set, the humidity sensor corresponding to the data collection is analyzed and marked as the sensor to be tested;

[0085] A sensor to be tested is selected, and in the monitoring model, humidity sensors within a preset radius of the sensor to be tested are marked with the sensor to be tested as the center, and multiple adjacent sensors are selected from the humidity sensors within the preset radius according to the spatial relationship, so that the multiple adjacent sensors are continuous with the sensor to be tested in spatial position;

[0086] Integrate the monitoring data of adjacent sensors and the monitoring data of the sensor to be tested into sample data, and sort the sample data according to the continuity of spatial position;

[0087] Based on the linear regression algorithm, the sample data is linearly fitted and the determination coefficient after fitting is calculated to determine whether the determination coefficient is less than a predetermined threshold. If so, it is determined that there is no linearity in the sample data and the sensor to be tested is marked as an abnormal sensor.

[0088] It should be noted that the sensor to be tested is a sensor that may be abnormal, and the corresponding monitoring data of the sensor has a certain abnormal trend. The multiple adjacent sensors and the sensor to be tested are continuous in spatial position, that is, the multiple adjacent sensors and the sensor to be tested are connected to form a straight line (or curve, whose deviation must be within the expected range), and the sensor to be tested is in the middle position. For sensors that are continuous in space, their corresponding humidity values ​​have a certain continuity.

[0089] According to an embodiment of the present invention, the abnormal equipment positioning and abnormal data analysis are performed in the monitoring model through abnormal sensors and corresponding monitoring data, and a sensor equipment maintenance plan is formulated, specifically:

[0090] Based on the abnormal sensor and the corresponding monitoring data, the abnormal equipment is located in the monitoring model, the sensor abnormality is evaluated in combination with the prediction data associated with the abnormal sensor, an abnormality report is generated, and a sensor equipment maintenance plan is formulated.

[0091] It is worth mentioning here that traditional humidity sensor monitoring is often based on simple numerical analysis and relies on manual inspection. It lacks predictive analysis and abnormal trend assessment of multiple humidity sensors, lacks integration of multi-sensor data, and is difficult to accurately and quickly evaluate and diagnose abnormal sensors. It is difficult to adapt to the high-precision humidity control application needs of industries such as industry and chemical industry.

[0092] In the present invention, the first step is to collect monitoring data from multiple sensors, set corresponding intervals to obtain monitoring data, use the continuity of the time dimension, perform data prediction through LSTM, associate the monitoring data with the predicted data and perform through the isolation forest, effectively judge whether there is an abnormality in the data change trend, and make a predictive analysis on the trend of abnormal equipment, and screen out the sensors to be tested with abnormal trends; the second step is to use the continuity of data in space to screen out the adjacent sensors of the sensor to be tested, and combine the adjacent sensors for linear regression fitting and linear relationship evaluation. If there is no linear relationship, the abnormal sensor is screened out. Through the present invention, it is possible to dynamically, quickly and accurately evaluate abnormal sensors, and quickly locate abnormal sensors, effectively improve the accuracy and efficiency of humidity monitoring, and reduce human participation. At the same time, the present invention is highly practical and can be effectively applied to the numerical analysis of sensors in complex environments, especially for the multi-sensor state evaluation and analysis of large-scale industrial production.

[0093] According to an embodiment of the present invention, it also includes:

[0094] In one analysis cycle, obtain abnormal data sets;

[0095] Through the abnormal data set, a humidity sensor is used as the analysis unit to calculate the linear change difference value between the monitoring data and the predicted data. Specifically, the linear change difference value is to perform linear fitting on the monitoring data and obtain a first change rate after fitting, and to perform linear fitting on the predicted data and obtain a second change rate after fitting. The difference between the first change rate and the second change rate is equal to the linear change difference value;

[0096] Determine whether the linear change difference value is greater than or less than a preset difference range. If so, dynamically adjust the preset time interval T based on the linear change difference value to obtain an adjusted T1, and re-set multiple time points based on T1 to collect data to obtain second monitoring data;

[0097] Predictive analysis and abnormality assessment of the humidity sensor are performed based on the second monitoring data.

[0098] It should be noted that abnormal assessment includes abnormal device positioning and abnormal data analysis. Whether the linear change difference value is greater than the preset difference value, it means that the current sensor is not in the expected normal working state. The larger the linear change difference value, the greater the deviation from the normal range. Further, the time interval is dynamically corrected through this value, so as to perform dynamic data collection and analysis, and realize dynamic adjustment of data collection amount and analysis and calculation amount.

[0099] Through the present invention, when the sensor is working well, the amount of data processing can be greatly reduced. When the sensor has an abnormal condition, the data collection process can be refined in real time, and the data can be modeled and analyzed in real time to improve the abnormality assessment capability. When there is a possibility of abnormality in the sensor operation, the amount of collected data can be dynamically increased for predictive analysis, the assessment accuracy can be dynamically improved, and the rationality of data collection can be effectively guaranteed.

[0100] The adjusted T1 is calculated as follows:

[0101] T1=T±[K×P×T+ln(T)];

[0102] Among them, T is the preset time interval, P is the linear change difference value, K is the correction coefficient, and the ± sign depends on whether the linear change difference value is greater than or less than the preset difference range. If it is greater, it is negative, and if it is less, it is positive.

[0103] Figure 2 A block diagram of a humidity sensor location optimization system based on a wireless sensor network according to the present invention is shown.

[0104] The second aspect of the present invention further provides a humidity sensor location optimization system 2 based on a wireless sensor network, the system comprising: a memory 21, a processor 22, the memory 21 comprising a humidity sensor location optimization program based on a wireless sensor network, the humidity sensor location optimization program based on a wireless sensor network when executed by the processor 22 implements the following steps:

[0105] Based on the humidity monitoring area, a monitoring model based on 3D visualization is constructed to obtain the distribution information of humidity sensors and import the distribution information into the monitoring model;

[0106] In one analysis cycle, the humidity monitoring area is monitored for humidity, and data of each humidity sensor is collected at multiple time points based on a preset time interval T, and monitoring data is obtained;

[0107] Based on the time dimension, the monitoring data is serialized to obtain the sequence data of each humidity sensor. The sequence data is imported into the LSTM-based prediction model for sequence prediction to generate the prediction data of each humidity sensor.

[0108] All monitoring data are associated with prediction data to form a monitoring data set. Based on the isolation forest algorithm, abnormal data points in the monitoring data set are judged and abnormal data sets are screened out.

[0109] Through the abnormal data set, the corresponding sensor to be tested is marked, and a sensor to be tested is selected. Based on the monitoring model, the sensor to be tested is used as the center to mark the adjacent sensors of the sensor to be tested, and the monitoring data of the adjacent sensors and the monitoring data of the sensor to be tested are integrated into sample data. The sample data is linearly fitted based on linear regression to determine whether there is a linear relationship. If there is no linear relationship, the sensor to be tested is marked as an abnormal sensor;

[0110] Through abnormal sensors and corresponding monitoring data, abnormal equipment location and abnormal data analysis are performed in the monitoring model to formulate sensor equipment maintenance plans;

[0111] Through abnormal equipment positioning and sensor equipment maintenance plans, the position of abnormal sensors is optimized, and the position of abnormal sensors is reset based on the monitoring feedback data after the implementation of the maintenance plan.

[0112] It should be noted that there are generally multiple adjacent sensors.

[0113] According to an embodiment of the present invention, based on the humidity monitoring area, a monitoring model based on three-dimensional visualization is constructed, the humidity sensor distribution information is obtained, and the distribution information is imported into the monitoring model, specifically:

[0114] Acquire spatial information of the humidity monitoring area, wherein the spatial information includes area, contour and monitoring range information of the area;

[0115] According to the spatial information, a monitoring model based on three-dimensional visualization is constructed;

[0116] Import the distribution information into the monitoring model for humidity visualization monitoring.

[0117] It should be noted that the monitoring model is used to visualize the distribution of sensors and humidity value monitoring distribution, which facilitates sensor management in the area.

[0118] According to an embodiment of the present invention, within one analysis cycle, the humidity monitoring area is monitored for humidity, and based on a preset time interval T, data is collected from each humidity sensor at multiple time points to obtain monitoring data, specifically:

[0119] Set an analysis week and a preset time interval T, and divide the analysis period into multiple time points based on the preset time interval T;

[0120] Data from each humidity sensor is collected at multiple time points to obtain monitoring data.

[0121] It should be noted that the monitoring data generally includes information such as humidity value, sensor temperature, sensor working status, etc. T is a default preset value. The larger T is, the fewer corresponding time points and the less data volume. Each humidity sensor corresponds to independent monitoring data.

[0122] According to an embodiment of the present invention, the monitoring data is serialized based on the time dimension to obtain the sequence data of each humidity sensor, and the sequence data is imported into the LSTM-based prediction model for sequence prediction to generate the prediction data of each humidity sensor, specifically:

[0123] Using a humidity sensor as the analysis unit, the monitoring data is preprocessed by data cleaning and standardization;

[0124] Based on the time dimension, the monitoring data is serialized to form sequence data;

[0125] Build a prediction model based on LSTM, set the prediction period to the preset time interval T, import the sequence data into the prediction model for sequence prediction, and generate prediction sequence data;

[0126] The predicted sequence data is analyzed to obtain the predicted data.

[0127] It should be noted that the monitoring data is data collected based on multiple time points, has a certain sequence, and is suitable for analysis by the LSTM prediction model, which can quickly and accurately predict and analyze the data. Each humidity sensor corresponds to corresponding prediction data.

[0128] It can be understood here that T is an adjustable time interval. With respect to prediction accuracy and prediction requirements, adjusting the size of T can simultaneously regulate the collection density of the original data, thereby adjusting the prediction accuracy.

[0129] According to an embodiment of the present invention, all monitoring data are associated with prediction data to form a monitoring data set, and based on the isolation forest algorithm, abnormal data points of the monitoring data set are judged, and abnormal data sets are screened out, specifically:

[0130] Based on each humidity sensor, the corresponding monitoring data is associated with the prediction data to form a monitoring data set;

[0131] Based on the isolation forest model, set the number of trees, the estimated proportion of abnormal data, and the random number seed parameters;

[0132] Import the monitoring data set into the isolation forest model for data prediction and analysis, make anomaly judgments for each data point, and generate a judgment list;

[0133] According to the judgment list, abnormal data is extracted from the monitoring data set to generate an abnormal data set.

[0134] It should be noted that the judgment list stores the judgment result of each data point.

[0135] According to an embodiment of the present invention, the abnormal data set is used to mark the corresponding sensor to be tested, a sensor to be tested is selected, and based on the monitoring model, the sensor to be tested is centered on the adjacent sensors of the sensor to be tested, and the monitoring data of the adjacent sensors and the monitoring data of the sensor to be tested are integrated into sample data, and the linear fitting of the sample data is performed based on linear regression to determine whether there is a linear relationship. If there is no linear relationship, the sensor to be tested is marked as an abnormal sensor, specifically:

[0136] Through the abnormal data set, the humidity sensor corresponding to the data collection is analyzed and marked as the sensor to be tested;

[0137] A sensor to be tested is selected, and in the monitoring model, humidity sensors within a preset radius of the sensor to be tested are marked with the sensor to be tested as the center, and multiple adjacent sensors are selected from the humidity sensors within the preset radius according to the spatial relationship, so that the multiple adjacent sensors are continuous with the sensor to be tested in spatial position;

[0138] Integrate the monitoring data of adjacent sensors and the monitoring data of the sensor to be tested into sample data, and sort the sample data according to the continuity of spatial position;

[0139] Based on the linear regression algorithm, the sample data is linearly fitted and the determination coefficient after fitting is calculated to determine whether the determination coefficient is less than a predetermined threshold. If so, it is determined that there is no linearity in the sample data and the sensor to be tested is marked as an abnormal sensor.

[0140] It should be noted that the sensor to be tested is a sensor that may be abnormal, and the corresponding monitoring data of the sensor has a certain abnormal trend. The multiple adjacent sensors and the sensor to be tested are continuous in spatial position, that is, the multiple adjacent sensors and the sensor to be tested are connected to form a straight line (or curve, whose deviation must be within the expected range), and the sensor to be tested is in the middle position. For sensors that are continuous in space, their corresponding humidity values ​​have a certain continuity.

[0141] According to an embodiment of the present invention, the abnormal equipment positioning and abnormal data analysis are performed in the monitoring model through abnormal sensors and corresponding monitoring data, and a sensor equipment maintenance plan is formulated, specifically:

[0142] Based on the abnormal sensor and the corresponding monitoring data, the abnormal equipment is located in the monitoring model, the sensor abnormality is evaluated in combination with the prediction data associated with the abnormal sensor, an abnormality report is generated, and a sensor equipment maintenance plan is formulated.

[0143] It is worth mentioning here that traditional humidity sensor monitoring is often based on simple numerical analysis and relies on manual inspection. It lacks predictive analysis and abnormal trend assessment of multiple humidity sensors, lacks integration of multi-sensor data, and is difficult to accurately and quickly evaluate and diagnose abnormal sensors. It is difficult to adapt to the high-precision humidity control application needs of industries such as industry and chemical industry.

[0144] In the present invention, the first step is to collect monitoring data from multiple sensors, set corresponding intervals to obtain monitoring data, use the continuity of the time dimension, perform data prediction through LSTM, associate the monitoring data with the predicted data and perform through the isolation forest, effectively judge whether there is an abnormality in the data change trend, and make a predictive analysis on the trend of abnormal equipment, and screen out the sensors to be tested with abnormal trends; the second step is to use the continuity of data in space to screen out the adjacent sensors of the sensor to be tested, and combine the adjacent sensors for linear regression fitting and linear relationship evaluation. If there is no linear relationship, the abnormal sensor is screened out. Through the present invention, it is possible to dynamically, quickly and accurately evaluate abnormal sensors, and quickly locate abnormal sensors, effectively improve the accuracy and efficiency of humidity monitoring, and reduce human participation. At the same time, the present invention is highly practical and can be effectively applied to the numerical analysis of sensors in complex environments, especially for the multi-sensor state evaluation and analysis of large-scale industrial production.

[0145] The third aspect of the present invention also provides a computer-readable storage medium, which includes a humidity sensor location optimization program based on a wireless sensor network. When the humidity sensor location optimization program based on a wireless sensor network is executed by a processor, the steps of the humidity sensor location optimization method based on a wireless sensor network as described in any of the above items are implemented.

[0146] The present invention discloses a humidity sensor position optimization method and system based on a wireless sensor network. By constructing a humidity monitoring model based on three-dimensional visualization, humidity sensor distribution information is imported. Within a preset time interval T, humidity monitoring area data is collected. After the monitoring data is serialized, it is imported into an LSTM prediction model to generate prediction data. In combination with the monitoring and prediction data, an isolated forest algorithm is used to identify abnormal data sets and mark the sensors to be tested. For abnormal sensors, adjacent sensors are marked by a monitoring model, integrated into sample data, and linear regression is used to judge the linear relationship. If there is no linear relationship, it is marked as abnormal. Finally, the abnormal equipment is located in the monitoring model and the data is analyzed, a maintenance plan is formulated and the sensor position is optimized. Through the present invention, it is possible to dynamically, quickly and accurately evaluate abnormal sensors, and quickly locate abnormal sensors, effectively improve the accuracy and efficiency of humidity monitoring, and reduce human participation.

[0147] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0148] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0149] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0150] Those skilled in the art can understand that: all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiments; and the aforementioned storage medium includes: mobile storage devices, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), disks or optical disks, and other media that can store program codes.

[0151] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.

[0152] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A humidity sensor location optimization method based on wireless sensor network, characterized in that: include: Based on the humidity monitoring area, a monitoring model based on three-dimensional visualization is constructed, and the distribution information of humidity sensors is obtained and imported into the monitoring model; In one analysis cycle, the humidity monitoring area is monitored for humidity. Based on the preset time interval T, data of each humidity sensor is collected at multiple time points to obtain monitoring data. The data collection is based on a wireless sensor network. Based on the time dimension, the monitoring data is serialized to obtain the sequence data of each humidity sensor. The sequence data is imported into the LSTM-based prediction model for sequence prediction to generate the prediction data of each humidity sensor. All monitoring data are associated with prediction data to form a monitoring data set. Based on the isolation forest algorithm, abnormal data points in the monitoring data set are judged and abnormal data sets are screened out. Through the abnormal data set, the corresponding sensor to be tested is marked, and a sensor to be tested is selected. Based on the monitoring model, the sensor to be tested is used as the center to mark the adjacent sensors of the sensor to be tested, and the monitoring data of the adjacent sensors and the monitoring data of the sensor to be tested are integrated into sample data. The sample data is linearly fitted based on linear regression to determine whether there is a linear relationship. If there is no linear relationship, the sensor to be tested is marked as an abnormal sensor; Through abnormal sensors and corresponding monitoring data, abnormal equipment location and abnormal data analysis are performed in the monitoring model to formulate sensor equipment maintenance plans; Through abnormal equipment positioning and sensor equipment maintenance plans, the position of abnormal sensors is optimized, and the position of abnormal sensors is reset based on the monitoring feedback data after the implementation of the maintenance plan.

2. A humidity sensor location optimization method based on a wireless sensor network according to claim 1, characterized in that: Based on the humidity monitoring area, a monitoring model based on three-dimensional visualization is constructed, the humidity sensor distribution information is obtained, and the distribution information is imported into the monitoring model, specifically: Acquire spatial information of the humidity monitoring area, wherein the spatial information includes area, contour and monitoring range information of the area; According to the spatial information, a monitoring model based on three-dimensional visualization is constructed; Import the distribution information into the monitoring model for humidity visualization monitoring.

3. A humidity sensor location optimization method based on a wireless sensor network according to claim 2, characterized in that: In one analysis cycle, the humidity monitoring area is monitored for humidity, and data from each humidity sensor is collected at multiple time points based on a preset time interval T, and monitoring data is obtained. The data collection is based on a wireless sensor network, specifically: Set an analysis week and a preset time interval T, and divide the analysis period into multiple time points based on the preset time interval T; Data from each humidity sensor is collected at multiple time points to obtain monitoring data.

4. The humidity sensor location optimization method based on wireless sensor network according to claim 3 is characterized in that: Based on the time dimension, the monitoring data is serialized to obtain the sequence data of each humidity sensor, and the sequence data is imported into the LSTM-based prediction model for sequence prediction to generate the prediction data of each humidity sensor, specifically: Using a humidity sensor as the analysis unit, the monitoring data is preprocessed by data cleaning and standardization; Based on the time dimension, the monitoring data is serialized to form sequence data; Build a prediction model based on LSTM, set the prediction period to the preset time interval T, import the sequence data into the prediction model for sequence prediction, and generate prediction sequence data; The predicted sequence data is analyzed to obtain the predicted data.

5. A humidity sensor location optimization method based on a wireless sensor network according to claim 4, characterized in that: The method associates all monitoring data with prediction data to form a monitoring data set, and based on the isolation forest algorithm, determines abnormal data points in the monitoring data set and screens out abnormal data sets, specifically: Based on each humidity sensor, the corresponding monitoring data is associated with the prediction data to form a monitoring data set; Based on the isolation forest model, set the number of trees, the estimated proportion of abnormal data, and the random number seed parameters; Import the monitoring data set into the isolation forest model for data prediction and analysis, make anomaly judgments for each data point, and generate a judgment list; According to the judgment list, abnormal data is extracted from the monitoring data set to generate an abnormal data set.

6. A humidity sensor location optimization method based on a wireless sensor network according to claim 5, characterized in that: The abnormal data set is used to mark the corresponding sensor to be tested, and a sensor to be tested is selected. Based on the monitoring model, the sensor to be tested is used as the center to mark the adjacent sensors of the sensor to be tested, and the monitoring data of the adjacent sensors and the monitoring data of the sensor to be tested are integrated into sample data. The sample data is linearly fitted based on linear regression to determine whether there is a linear relationship. If there is no linear relationship, the sensor to be tested is marked as an abnormal sensor. Specifically: Through the abnormal data set, the humidity sensor corresponding to the data collection is analyzed and marked as the sensor to be tested; A sensor to be tested is selected, and in the monitoring model, humidity sensors within a preset radius of the sensor to be tested are marked with the sensor to be tested as the center, and multiple adjacent sensors are selected from the humidity sensors within the preset radius according to the spatial relationship, so that the multiple adjacent sensors are continuous with the sensor to be tested in spatial position; Integrate the monitoring data of adjacent sensors and the monitoring data of the sensor to be tested into sample data, and sort the sample data according to the continuity of spatial position; Based on the linear regression algorithm, the sample data is linearly fitted and the determination coefficient after fitting is calculated to determine whether the determination coefficient is less than a predetermined threshold. If so, it is determined that there is no linearity in the sample data and the sensor to be tested is marked as an abnormal sensor.

7. A humidity sensor location optimization method based on a wireless sensor network according to claim 6, characterized in that: By using abnormal sensors and corresponding monitoring data, abnormal equipment positioning and abnormal data analysis are performed in the monitoring model, and a sensor equipment maintenance plan is formulated, specifically: Based on the abnormal sensor and the corresponding monitoring data, the abnormal equipment is located in the monitoring model, the sensor abnormality is evaluated in combination with the prediction data associated with the abnormal sensor, an abnormality report is generated, and a sensor equipment maintenance plan is formulated.

8. A humidity sensor location optimization system based on wireless sensor network, characterized in that: The system includes: a memory and a processor, wherein the memory includes a humidity sensor location optimization program based on a wireless sensor network, and when the humidity sensor location optimization program based on a wireless sensor network is executed by the processor, the following steps are implemented: Based on the humidity monitoring area, a monitoring model based on three-dimensional visualization is constructed, and the distribution information of humidity sensors is obtained and imported into the monitoring model; In one analysis cycle, the humidity monitoring area is monitored for humidity. Based on the preset time interval T, data of each humidity sensor is collected at multiple time points to obtain monitoring data. The data collection is based on a wireless sensor network. Based on the time dimension, the monitoring data is serialized to obtain the sequence data of each humidity sensor. The sequence data is imported into the LSTM-based prediction model for sequence prediction to generate the prediction data of each humidity sensor. All monitoring data are associated with prediction data to form a monitoring data set. Based on the isolation forest algorithm, abnormal data points in the monitoring data set are judged and abnormal data sets are screened out. Through the abnormal data set, the corresponding sensor to be tested is marked, and a sensor to be tested is selected. Based on the monitoring model, the sensor to be tested is used as the center to mark the adjacent sensors of the sensor to be tested, and the monitoring data of the adjacent sensors and the monitoring data of the sensor to be tested are integrated into sample data. The sample data is linearly fitted based on linear regression to determine whether there is a linear relationship. If there is no linear relationship, the sensor to be tested is marked as an abnormal sensor; Through abnormal sensors and corresponding monitoring data, abnormal equipment location and abnormal data analysis are performed in the monitoring model to formulate sensor equipment maintenance plans; Through abnormal equipment positioning and sensor equipment maintenance plans, the position of abnormal sensors is optimized, and the position of abnormal sensors is reset based on the monitoring feedback data after the implementation of the maintenance plan.

9. A humidity sensor location optimization system based on a wireless sensor network according to claim 8, characterized in that: Based on the humidity monitoring area, a monitoring model based on three-dimensional visualization is constructed, the humidity sensor distribution information is obtained, and the distribution information is imported into the monitoring model, specifically: Acquire spatial information of the humidity monitoring area, wherein the spatial information includes area, contour and monitoring range information of the area; According to the spatial information, a monitoring model based on three-dimensional visualization is constructed; Import the distribution information into the monitoring model for humidity visualization monitoring.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a humidity sensor location optimization program based on a wireless sensor network. When the humidity sensor location optimization program based on a wireless sensor network is executed by a processor, the steps of the humidity sensor location optimization method based on a wireless sensor network as described in any one of claims 1 to 7 are implemented.

Images