A radiation source intelligent monitoring method and system based on the Internet of Things

Through the IoT-based method, the radiation source detection data is monitored and processed in real time, and the abnormality and risk score is calculated using encoder and fuzzy logic, the problems of low data processing efficiency and inaccurate risk assessment in the existing technology are solved, achieving efficient Intelligent supervision of radiation sources.

CN118708977BActive Publication Date: 2024-12-03STATE GRID JIANGSU ELECTRIC POWER CO LIANYUNGANG POWER SUPPLY CO +2
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
CN202411186993.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-28
Publication Date
2024-12-03
Estimated Expiration
2044-08-28

AI Technical Summary

Technical Problem

When existing radiation source intelligent supervision technology processes large-scale, real-time updated data flows, the data structure is insufficient and the lack of dynamic threshold adjustment mechanisms, resulting in low access and processing efficiency, difficulty in identifying and responding to rapidly changing environmental conditions, and traditional numerical processing The method lacks processing accuracy and flexibility in the face of uncertain data, which affects the refinement of risk assessment and the quality of decision making.

Method used

Using an Internet of Things method, the radiation source detection data flow is monitored in real time, and the geographic, time and radiation intensity characteristic data are recorded synchronously, and transmitted to the timing database for indexing and aggregation. The encoder learns and optimizes parameters to calculate abnormal indicators and fuzzy weight scores. , combining anomaly identification threshold and fuzzy weight score to calculate risk scores, and dynamically adjust the threshold to improve response flexibility and risk assessment accuracy.

Benefits of technology

It realizes efficient access and processing speed of data, enhances the ability to predict and identify abnormal states, improves the meticulousness and response speed of risk assessment, and improves the accuracy and efficiency of radiation safety management.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118708977B_ABST
    Figure CN118708977B_ABST
Patent Text Reader

Abstract

The present invention discloses a radiation source intelligent supervision method and system based on the Internet of Things. The method comprises: synchronously recording the geographic, temporal and radiation intensity characteristic data set of each data point, and transmitting the data set to a time series database to perform index processing, aggregation and normalization on the data to generate a radiation data set; performing characteristic index calculation, applying an encoder for learning, optimizing encoder parameters, calculating the abnormal index of each data point, and calculating the abnormal recognition threshold; converting the radiation intensity into a fuzzy value, and performing aggregation operations to obtain the fuzzy matching degree, and then calculating the fuzzy weight score of each data point; combining the abnormal recognition threshold with the fuzzy weight score, calculating the risk score of each data point, and calculating the risk assessment result of each data point in combination with the risk classification threshold.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of data management and relates to an intelligent monitoring method and system for radiation sources based on the Internet of Things. Background Art

[0002] The field of data management technology encompasses all aspects of data collection, storage, maintenance, retrieval, and analysis. It is a core area of ​​computer science and information technology, focusing on effectively processing data to support business and decision-making. Technologies in this area include, but are not limited to, database management system design and optimization, data security and privacy protection, the construction of data warehouses and data lakes, and complex data query and information extraction techniques. Data management also involves data integration and cleansing to ensure data quality and consistency, as well as data governance, the process of ensuring that data is used appropriately and in accordance with regulations and standards throughout the organization.

[0003] The Intelligent Radiation Source Monitoring System utilizes advanced information technology and hardware and software to monitor and manage radiation sources in real time. It is primarily used to ensure radiation safety and prevent hazards caused by radiation leakage or improper use. By monitoring the status and environment of radiation sources, the system can promptly detect anomalies and implement appropriate preventive or response measures. Furthermore, the Intelligent Radiation Source Monitoring System can help relevant agencies optimize resource allocation and improve emergency response efficiency, thereby more effectively protecting public health and safety.

[0004] Existing intelligent radiation source monitoring technologies often struggle with processing large-scale, real-time data streams due to insufficient data structure and a lack of effective dynamic threshold adjustment mechanisms. This leads to inefficient access and processing, making it difficult to effectively identify and respond to rapidly changing environmental conditions. Furthermore, traditional numerical processing methods often employ simplified approaches when dealing with data containing uncertainty. This not only limits processing accuracy but also reduces operational flexibility, impacting the sophistication of risk assessments and the quality of decision-making. Summary of the Invention

[0005] In order to solve the deficiencies in the prior art, the present invention provides a method and system for intelligent monitoring of radiation sources based on the Internet of Things.

[0006] The present invention adopts the following technical solutions.

[0007] A first aspect of the present invention provides a method for intelligently monitoring radiation sources based on the Internet of Things, the method comprising:

[0008] S1: Based on the Internet of Things, real-time monitoring of radiation source detection data streams is performed, and the geographic, temporal, and radiation intensity characteristic data sets of each data point are synchronously recorded. The data is then transferred to a time series database for indexing, aggregation, and normalization to generate a radiation dataset.

[0009] S2: Calculate the characteristic index of each data point in the radiation dataset, and use the encoder to learn the characteristic index to optimize the encoder parameters. Use the optimized encoder parameters and characteristic index to calculate the anomaly index of each data point, and then calculate the anomaly recognition threshold;

[0010] S3: Convert the radiation intensity of each data point in the radiation dataset into a fuzzy value, and perform aggregation operation on the fuzzy values ​​of all data points to obtain the fuzzy matching degree of the entire radiation dataset, and then calculate the fuzzy weight score of each data point;

[0011] S4: The anomaly identification threshold is combined with the fuzzy weight score to calculate the risk score of each data point, and the risk assessment result of each data point is calculated in combination with the risk classification threshold.

[0012] Preferably, in S1, based on the Internet of Things device, the radiation source detection data stream from the radiation detector is monitored in real time, including the geographic location tag, timestamp and radiation intensity. The radiation detector is used as a data point, and the geographic, time and radiation intensity characteristic data set of each data point is synchronously recorded, specifically:

[0013] ,

[0014] in, Represents the geographical, temporal and radiation intensity characteristic data set of the i-th data point;

[0015] Represents geographical features, Represents time characteristics, Indicates the radiation intensity characteristics;

[0016] Represents a geographic location tag. Indicates a timestamp, Indicates radiation intensity;

[0017] 、 represents the scaling factor, Indicates the time offset, An exponent representing the adjustment of radiation intensity.

[0018] Preferably, in S1, the geographical, temporal and radiation intensity characteristic data sets of each data point are transferred to a time series database, and the data are indexed to obtain an index data stream. :

[0019] ,

[0020] in, represents the index function, represents the eigenvalue extraction function, Indicates that the index is The weight corresponding to the feature, represents the index of the feature, for The index is The eigenvalues ​​of the features.

[0021] Preferably, in S1, the index data stream Perform aggregation and normalization to generate a radiation dataset:

[0022] ,

[0023] in, represents the total number of data points, Represents an aggregation operation, represents the normalization parameter.

[0024] Preferably, in S2, the calculation formula of the characteristic index is:

[0025] ,

[0026] in, Representative radiation dataset Middle The characteristic index of the data point, Representative radiation dataset Middle The value of the data point.

[0027] Preferably, in S2, the optimization formula for the encoder parameters is:

[0028] ,

[0029] in, 、 Represent the encoder parameters before and after optimization, represents the learning rate, (,) represents the loss function, Represents a set of characteristic indicators, represents the regularization coefficient; represents the square of the modulus; Represents the encoder parameters Gradient operation.

[0030] Preferably, in S2, the calculation formula for the abnormality index of each data point is:

[0031] ,

[0032] in, Representative Anomaly indicators for data points, represents the reconstruction error, represents the weight factor, represents the average position of the feature index data set in the latent space, Representative potential locations of data points, represents the normalization factor, is a constant.

[0033] Preferably, in S2, the calculation formula for the abnormality recognition threshold is:

[0034] ,

[0035] ,

[0036] in, represents the anomaly recognition threshold, represents the dynamic adjustment coefficient, represents the sensitivity of the control adjustment, represents the base of natural logarithm, represents the initial anomaly identification threshold, Represents the parameter that controls the multiple of the standard deviation, 、 represents the adjustment factor; Represents the average value of the abnormal indicators of all data points, Represents the standard deviation of the anomaly indicator for all data points; represents the abnormal indicator dataset; is the data skewness coefficient.

[0037] Preferably, in S3, the formula for converting radiation intensity into fuzzy value is:

[0038] ,

[0039] in, Representative The fuzzy value of the data point, Is the radiation dataset Middle The radiation intensity of each data point, To control the slope of the curve, Is the radiation dataset The mean radiation intensity of all data points in , Represents the base of natural logarithms.

[0040] Preferably, in S3, the fuzzy values ​​of all data points are aggregated to obtain the fuzzy matching degree of the entire radiation dataset:

[0041] ,

[0042] in, Represents the degree of fuzzy matching, is the total number of data points, It is The weight of the data point.

[0043] Preferably, in S3, the calculation formula for the fuzzy weight score of each data point is:

[0044] ,

[0045] in, Representative The fuzzy weight score of each data point.

[0046] Preferably, in S4, the risk score calculation formula for each data point is:

[0047] ,

[0048] in, Indicates the Risk score for each data point, is the parameter for adjusting the fuzzy weight gain, represents the anomaly recognition threshold, Representative The fuzzy weight score of the data points, Represents the base of natural logarithms.

[0049] Preferably, in S4, the risk classification threshold calculation formula is:

[0050] ,

[0051] in, is the risk classification threshold, is the coefficient for adjusting the threshold; Represents the average value of the abnormal indicators of all data points, Represents the standard deviation of the anomaly indicator for all data points.

[0052] Preferably, in S4, the formula for calculating the risk assessment result of each data point is:

[0053] ,

[0054] in, Representative The safety level index of each data point, 、 is the adjustment factor to control the threshold sensitivity, Indicates the Risk score for each data point, is the risk classification threshold.

[0055] A second aspect of the present invention provides an intelligent monitoring system for radiation sources based on the Internet of Things, comprising:

[0056] The radiation data acquisition module is used to monitor the radiation detection data stream in real time based on the Internet of Things, synchronously record the geographical, temporal and radiation intensity characteristic data sets of each data point, and transmit them to the time series database for indexing, aggregation and normalization of the data to generate a radiation data set;

[0057] The anomaly recognition threshold calculation module is used to calculate the characteristic index of each data point in the radiation data set, and use the encoder to learn the characteristic index to optimize the encoder parameters. The optimized encoder parameters and characteristic index are used to calculate the anomaly index of each data point, and then the anomaly recognition threshold is calculated;

[0058] The fuzzy weight scoring module is used to convert the radiation intensity of each data point in the radiation dataset into a fuzzy value, and perform aggregation operations on the fuzzy values ​​of all data points to obtain the fuzzy matching degree of the overall data of the radiation dataset, and then calculate the fuzzy weight score of each data point;

[0059] The risk assessment module is used to combine the anomaly identification threshold with the fuzzy weight score to calculate the risk score of each data point, and to calculate the risk assessment result of each data point in combination with the risk classification threshold.

[0060] A terminal includes a processor and a storage medium; the storage medium is used to store instructions;

[0061] The processor is configured to operate according to the instructions to execute the steps of the method.

[0062] A computer-readable storage medium stores a computer program, which implements the steps of the method when executed by a processor.

[0063] Compared with the prior art, the beneficial effects of the present invention include at least:

[0064] The present invention is based on Internet of Things devices, monitors the data stream from radiation detectors in real time, synchronously records the geographical, temporal and radiation intensity characteristic data sets of each data point, and transmits them to a time series database to index, aggregate and normalize the data to generate a radiation data set, which can achieve efficient data access and improve processing speed.

[0065] The present invention utilizes the inherent structure of encoder learning data to enhance the ability to predict and identify abnormal states, optimizes encoder parameters, and uses the optimized encoder parameters and characteristic indicators to calculate the abnormality index of each data point, thereby calculating the abnormality recognition threshold. This can achieve dynamic adjustment of the abnormality recognition threshold and improve flexible response to environmental changes.

[0066] The present invention converts radiation intensity data into fuzzy values ​​and performs aggregation operations on the fuzzy values ​​of all data points to obtain the fuzzy matching degree of the overall radiation data set, and then calculates the fuzzy weight score of each data point, which improves the flexibility and accuracy of data processing and makes risk assessment more detailed and comprehensive.

[0067] The present invention comprehensively applies anomaly recognition thresholds and fuzzy weight scores to calculate the risk score of each data point, and combines risk classification thresholds to calculate the risk assessment results of each data point, thereby improving the accuracy of risk assessment results and enhancing the precision and response speed of radiation safety management. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 This is a principle block diagram of the present invention;

[0069] Figure 2 Flowchart of the steps for obtaining a radiation dataset of the present invention;

[0070] Figure 3 Flowchart of the steps for obtaining abnormal indicators of the present invention;

[0071] Figure 4 Flowchart of the steps for obtaining abnormality identification parameters of the present invention;

[0072] Figure 5 Flowchart of the steps for obtaining fuzzy weighted scores of the present invention;

[0073] Figure 6 This is a flow chart of the steps for obtaining the security level of the present invention. DETAILED DESCRIPTION

[0074] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. The embodiments described in this application are only part of the embodiments of the present invention, not all of them. Based on the spirit of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0075] Embodiment 1 of the present invention provides a method for intelligent monitoring of radiation sources based on the Internet of Things, such as Figure 1 As shown, the method includes:

[0076] S1: Based on the Internet of Things, real-time monitoring of radiation detection data streams is carried out, and the geographical, temporal, and radiation intensity characteristic data sets of each data point are synchronously recorded. The data are then transferred to a time series database for indexing, aggregation, and normalization to generate a radiation dataset.

[0077] Further preferably, based on the Internet of Things devices, the radiation source detection data stream from the radiation detector is monitored in real time, the radiation detector is used as the data point, the geographic tag, timestamp, and radiation intensity of each detection point are synchronously recorded, the information is aggregated into a time series database in real time, indexed and stored, and a radiation data set is generated, involving the time series database index and storage structure, such as Figure 2 As shown, specifically including:

[0078] (1) Use IoT devices to receive real-time data from radiation detectors, including geotags, timestamps, and radiation intensity, using the formula:

[0079] ,

[0080] Generate primary data streams;

[0081] in, represents a set of data points, Represents a geographic location tag. Indicates a time stamp. represents the radiation intensity, 、 represents the scaling factor, represents the time offset, An index indicating the adjustment of radiation intensity;

[0082] (2) Transfer the primary data stream to the time series database and index the data to assist in querying using the formula:

[0083] ,

[0084] Get index data stream ;

[0085] in, represents the index function, represents the characteristic function, represents the weight, The index representing the feature;

[0086] (3) Using index data stream, through data processing and aggregation algorithms, adopt the formula:

[0087] ,

[0088] Generate radiation dataset;

[0089] in, represents a radiation dataset, represents the total number of data points, Represents an aggregation operation, represents the normalization parameter.

[0090] The following are examples of the above formulas:

[0091] (1) Formula:

[0092] ,

[0093] Parameter details:

[0094] : Geographic location tag, assumed to be a longitude value.

[0095] : Geolocation scaling factor, assumed to be 1.1 (to simulate geolocation adjustments).

[0096] : Timestamp, assumed to be in hours.

[0097] : Time scaling factor, assumed to be 1.05 (indicating time adjustment).

[0098] : Time offset, assumed to be 2 hours (indicates the offset from the base time).

[0099] : Radiation intensity, assumed to be in microsieverts.

[0100] : Exponential adjustment parameter for radiation intensity, assumed to be 0.5 (square root adjustment of radiation intensity).

[0101] Calculation process and examples:

[0102] Assumptions (longitude), (Hour), (microsievert). The calculation process is as follows:

[0103] ,

[0104] ,

[0105] ,

[0106] The result is This represents a reading of 4 microsieverts at an adjusted geographic location of 33 degrees longitude and 12.5 hours later.

[0107] (2) Formula:

[0108] ,

[0109] Parameter details:

[0110] : Weight coefficient, set .

[0111] : Function processing function, assuming It means directly taking the value.

[0112] Calculation process and examples:

[0113] use The result,

[0114] The calculation process is as follows:

[0115] ,

[0116] ,

[0117] ,

[0118] The index value 27.6 represents the weighted index, which is used for retrieval and query.

[0119] (3) Formula:

[0120] ,

[0121] Parameter details:

[0122] : Normalization parameter, set to 2 (assuming simplified normalization).

[0123] : The total number of data points, assuming there is only 1 data point, that is .

[0124] Calculation process and examples:

[0125] use ,

[0126] The calculation process is as follows:

[0127] (because there is only one data point)

[0128] ,

[0129] The value 5.25 represents the total value of the dataset after normalization for analysis or storage.

[0130] S2: Calculate the characteristic index of each data point in the radiation dataset, and use the encoder to learn the characteristic index to optimize the encoder parameters. Use the optimized encoder parameters and characteristic index to calculate the anomaly index of each data point, and then calculate the anomaly recognition threshold;

[0131] Further preferably, the features of each data point are extracted from the radiation data set, the intrinsic structure of the data is learned by applying feature encoding, the parameters of the encoder are optimized by the gradient descent method, the anomaly index of each data point is calculated according to the reconstruction error and the latent space distribution, and the recognition threshold is dynamically updated according to the distribution of the anomaly index to obtain the anomaly recognition parameter.

[0132] like Figure 3 As shown in the figure, the steps for obtaining abnormal indicators are as follows:

[0133] (1) Extract the features of each data point from the radiation dataset and apply the feature extraction algorithm to identify the intrinsic structure of the data using the formula:

[0134] ,

[0135] Generate feature set;

[0136] in, Representative The feature set of data points, Representative dataset The data points, The sum square is to increase nonlinearity and handle large value variation;

[0137] (2) Apply the encoder to the feature set for learning and use the gradient descent method to optimize the encoder parameters using the formula:

[0138] ,

[0139] Generate optimized encoder parameters;

[0140] in, 、 Represent the new and old encoder parameters respectively, represents the learning rate, represents the loss function, Represents the feature set, represents the regularization coefficient; Represents the encoder parameters Gradient operation.

[0141] (3) Using the optimized encoder parameters and feature set, calculate the reconstruction error of each data point and its position in the latent space using the formula:

[0142] ,

[0143] Generate anomaly indicators;

[0144] in, Representative Anomaly indicators for data points, represents the reconstruction error, represents the weight factor, Represents the average position of the feature index data set in the latent space (take the feature index set All ), Representative The potential location of the data point (take ), represents the normalization factor, is a small constant.

[0145] The following are examples of the above formulas:

[0146] Dataset assumptions

[0147] ,

[0148] ,

[0149] formula:

[0150] ,

[0151] Calculation process:

[0152] for :

[0153] ,

[0154] for :

[0155] ,

[0156] for :

[0157] ,

[0158] formula:

[0159] ,

[0160] Assume parameters:

[0161] Initial parameters ,

[0162] Learning rate ,

[0163] Regularization coefficient ,

[0164] Assume that the gradient of the loss function (Calculated based on the actual loss function)

[0165] Calculation process:

[0166] Calculate the regularization term:

[0167] ,

[0168] Update parameters:

[0169] ,

[0170] formula:

[0171] ,

[0172] Assume parameters:

[0173] Assumed to be 0.05 (calculated by model reconstruction error)

[0174] Average position (Assuming that, according to calculated)

[0175] use Value

[0176] ,

[0177] ,

[0178] (Avoiding zero denominator)

[0179] Calculation process:

[0180] for :

[0181] ,

[0182] like Figure 4 As shown, the steps for obtaining abnormal identification parameters are as follows:

[0183] (1) Using anomaly indicators, extract statistical characteristics, including mean and standard deviation, and determine the baseline anomaly threshold using the formula:

[0184] ,

[0185] and

[0186] ,

[0187] Generate statistical results;

[0188] in, represents the average value of the abnormal index, represents the standard deviation of the abnormal indicator, represents the total number of data points;

[0189] (2) According to the statistical results, set the initial anomaly identification threshold, refer to the skewness and kurtosis of the data distribution, adjust the threshold to identify potential anomalies, and use the formula:

[0190] ,

[0191] Generate preliminary anomaly identification thresholds;

[0192] in, represents the initial anomaly identification threshold, is a parameter that controls the multiple of the standard deviation. 、 is the adjustment factor; is the data skewness coefficient;

[0193] (3) Dynamically adjust the initial anomaly identification threshold to match the actual distribution of data and recent abnormal activities, using the formula:

[0194] ,

[0195] Generate anomaly identification parameters;

[0196] in, represents the anomaly recognition threshold, is the dynamic adjustment coefficient, Controls the sensitivity of adjustments, Represents the base of natural logarithms.

[0197] The following are examples of the above formulas:

[0198] formula:

[0199] ,

[0200] ,

[0201] in:

[0202] : Average value of abnormal indicators

[0203] : Standard deviation of abnormal indicators

[0204] : The total number of data points

[0205] : No. Anomaly indicators for data points

[0206] Calculation process:

[0207] Suppose there is an abnormal indicator dataset .

[0208] First calculate the average :

[0209] ,

[0210] Next, calculate the standard deviation :

[0211] ,

[0212] formula:

[0213] ,

[0214] in:

[0215] : Preliminary anomaly identification threshold

[0216] : Standard deviation multiple, control threshold strictness

[0217] 、 : Adjustment coefficient, reflecting the effects of skewness and kurtosis

[0218] : Data skewness coefficient

[0219] Calculation process:

[0220] Assumptions ,and (The data are less skewed).

[0221] Calculating the threshold :

[0222] ,

[0223] formula:

[0224] ,

[0225] in:

[0226] : Adjusted anomaly recognition threshold

[0227] : Dynamic adjustment coefficient

[0228] : Control adjustment sensitivity

[0229] : natural logarithm base

[0230] Calculation process:

[0231] Assumptions .

[0232] Based on the previous ,and .

[0233] calculate :

[0234] ,

[0235] S3: Convert the radiation intensity of each data point in the radiation dataset into a fuzzy value, and perform aggregation operation on the fuzzy values ​​of all data points to obtain the fuzzy matching degree of the entire radiation dataset, and then calculate the fuzzy weight score of each data point;

[0236] Further preferably, the fuzzy weight score involves a fuzzy value and a degree of aggregation;

[0237] For each data point in the radiation dataset, fuzzification is performed according to the radiation intensity, the real value is converted into a fuzzy value, the fuzzy values ​​of multiple data points are aggregated through fuzzy operation, the fuzzy matching degree of the overall data is calculated, the uncertainty level of each data point is evaluated, and the fuzzy weight score is obtained, such as Figure 5 As shown, specifically including:

[0238] (1) Each data point in the radiation data set is fuzzified according to the radiation intensity and converted into a fuzzy value using the formula:

[0239] ,

[0240] Generate fuzzy value sets;

[0241] in, Representative The fuzzy value of the data point, It is a dataset Middle The radiation intensity of each data point, Controls the slope of the curve, is the mean radiation intensity;

[0242] (2) Aggregate the fuzzy values ​​through fuzzy operation and calculate the fuzzy matching degree of the overall data using the formula:

[0243] ,

[0244] Generate fuzzy matching degree;

[0245] in, Represents the degree of fuzzy matching, N is the total number of data points, It is The weight of each data point;

[0246] (3) Evaluate the degree of fuzzy matching and the fuzzy value of each data point, calculate the uncertainty level of each data point, and obtain the fuzzy weight score using the formula:

[0247] ,

[0248] Generate fuzzy weighted score set;

[0249] in, Representative The fuzzy weight score of each data point.

[0250] The following are examples of the above formulas:

[0251] formula:

[0252] ,

[0253] Parameter explanation and examples:

[0254] : No. The radiation intensity of each data point.

[0255] : Controls the slope of the curve, a common value is in the range of 0.1 to 1.

[0256] : The average value of the radiation intensity of the data points.

[0257] Assume data:

[0258] Unit radiation intensity.

[0259] Average unit radiation intensity.

[0260] .

[0261] Calculation process:

[0262] calculate .

[0263] Calculate the exponential part: .

[0264] Compute the exponential function: Very close to 0.

[0265] Calculate fuzzy value :

[0266] .

[0267] Result description:

[0268] Given Well above average , making the fuzzy value Close to 1, indicating the influence of high radiation intensity.

[0269] formula:

[0270] ,

[0271] Parameter explanation and examples:

[0272] : fuzzy value.

[0273] : Weight, assuming that each .

[0274] Assume data:

[0275] .

[0276] Data points.

[0277] Calculation process:

[0278] Calculate the weighted sum: .

[0279] Calculate the weighted fuzzy value sum:

[0280] .

[0281] Calculate the degree of fuzzy matching .

[0282] Result description:

[0283] Fuzzy matching degree A value of 0.5 indicates an average blur level.

[0284] formula:

[0285] ,

[0286] Parameter explanation and examples:

[0287] : fuzzy value.

[0288] : Fuzzy matching degree.

[0289] Calculation process:

[0290] For each calculate .

[0291] right .

[0292] Computes the product and accumulation of the numerator and denominator.

[0293] calculate (The rest of the calculation is similar).

[0294] Result description:

[0295] The value indicates the relative importance of the data point in the degree of fuzzy matching. A value of indicates that the data point deviates significantly from the overall fuzzy match.

[0296] S4: The anomaly identification threshold is combined with the fuzzy weight score to calculate the risk score of each data point, and the risk assessment result of each data point is calculated in combination with the risk classification threshold, that is, the safety level index of the data point, so as to achieve intelligent supervision. For example, when the safety level index of a data point is monitored to be lower than the set threshold or its corresponding preset level is higher than the set risk level, it will be inspected or other control measures will be taken.

[0297] Further preferably, the risk assessment of the radiation source data is performed by combining the abnormal identification parameters and the fuzzy weight score, and the safety level of each data point is determined by threshold comparison analysis, such as Figure 6 As shown, specifically including:

[0298] (1) Combine the anomaly identification parameters and the fuzzy weight score of each data point to calculate the preliminary risk score of each data point using the formula:

[0299] ,

[0300] Generate a preliminary risk score set;

[0301] in, Indicates the Risk score for each data point, is the parameter for adjusting the fuzzy weight gain;

[0302] (2) Calculate the weighted mean and weighted standard deviation of the preliminary risk score set for all data points using the formula:

[0303] ,

[0304] Determine thresholds for risk classification;

[0305] in, is the risk classification threshold, is the coefficient for adjusting the threshold;

[0306] (3) Calculate the safety level index of each data point based on the threshold using the formula:

[0307] ,

[0308] Generate a security rating for each data point;

[0309] in, Representative Safety level index of data points, adjustment factor Controls the threshold sensitivity.

[0310] The following are examples of the above formulas:

[0311] formula:

[0312] ,

[0313] in:

[0314] It is an anomaly recognition threshold set based on the system's historical data, such as 0.5.

[0315] It is The fuzzy weight score of each data point is assumed to be 0.8.

[0316] is a coefficient used to adjust the influence of fuzzy weight score, which is set as:

[0317] - Adjust the impact amplitude,

[0318] - adjust the steepness of the curve,

[0319] -Sets the center point of the function.

[0320] Calculation process:

[0321] calculate Adjustment part: ,

[0322] set up ,

[0323] but ,

[0324] Calculate the overall risk score :

[0325] ,

[0326] result Indicates the A risk score is calculated for each data point, with lower values ​​indicating lower risk.

[0327] Formula analysis:

[0328] ,

[0329] Assumptions:

[0330] is the average value of the risk score, set to 0.4.

[0331] is the standard deviation, set to 0.05.

[0332] is the risk level classification coefficient, which is set to 1.5.

[0333] Calculation process:

[0334] ,

[0335] result Indicates the threshold for risk classification, data points exceeding the value will be classified as high risk.

[0336] Formula analysis:

[0337] ,

[0338] Given:

[0339] - weighting factor affecting the portion exceeding the threshold,

[0340] -Affects the steepness of the curve.

[0341] Calculation process:

[0342] According to the threshold and risk scores :

[0343] ,

[0344] result Represents the safety level index, with lower values ​​indicating lower risk levels.

[0345] Embodiment 2 of the present invention provides an intelligent monitoring system for radiation sources based on the Internet of Things, including:

[0346] The radiation data acquisition module is used to monitor radiation detection data streams in real time based on the Internet of Things, synchronously record the geographic, temporal, and radiation intensity characteristic data sets of each data point, and transmit them to a time series database for indexing, aggregation, and normalization to generate a radiation dataset. This module can receive data streams from radiation detectors through IoT devices, record geographic tags, timestamps, and radiation intensity point by point, and enter the data into the time series database for indexing in real time to construct a radiation dataset.

[0347] The anomaly recognition threshold calculation module is used to calculate the characteristic index of each data point in the radiation data set, and use the encoder to learn the characteristic index to optimize the encoder parameters. The optimized encoder parameters and characteristic index are used to calculate the anomaly index of each data point, and then the anomaly recognition threshold is calculated. This module can separate the characteristics of each data point from the radiation data set, encode the characteristics, calculate the anomaly index and latent space distribution of each data point, adjust the recognition threshold according to the dynamic distribution of the anomaly index, and complete the formulation of anomaly recognition parameters.

[0348] The fuzzy weight scoring module is used to convert the radiation intensity of each data point in the radiation dataset into a fuzzy value, and perform aggregation operations on the fuzzy values ​​of all data points to obtain the fuzzy matching degree of the entire radiation dataset, and then calculate the fuzzy weight score of each data point. This module can fuzzify the data points in the radiation dataset, convert the radiation intensity into a fuzzy value, and perform aggregation operations on the fuzzy values ​​to evaluate and determine the uncertainty level of each data point and generate a fuzzy weight score.

[0349] The risk assessment module is used to combine the anomaly identification threshold with the fuzzy weight score to calculate the risk score for each data point, and then calculate the risk assessment result for each data point by combining the risk classification threshold. This module can combine the anomaly identification parameters with the fuzzy weight score to perform risk assessment on the data points. Through comparative analysis of the risk level threshold, risky data points are identified and marked, and recorded as the risk assessment result.

[0350] Embodiment 3 of the present invention provides a terminal, including a processor and a storage medium; the storage medium is used to store instructions; the processor is used to operate according to the instructions to execute the steps of the method.

[0351] Embodiment 4 of the present invention provides a computer-readable storage medium having a computer program stored thereon, which implements the steps of the method when executed by a processor.

[0352] Compared with the prior art, the beneficial effects of the present invention include at least:

[0353] The present invention is based on Internet of Things devices, monitors data streams from radiation detectors in real time, synchronously records geographic, temporal and radiation intensity characteristic data of each detection point, aggregates them into a time series database in real time, indexes, aggregates and normalizes the data, and generates a radiation dataset, which can achieve efficient data access and improve processing speed.

[0354] The present invention utilizes the inherent structure of encoder learning data to enhance the ability to predict and identify abnormal states, optimizes encoder parameters, and uses the optimized encoder parameters and characteristic indicators to calculate the abnormality index of each data point, thereby calculating the abnormality recognition threshold. This can achieve dynamic adjustment of the abnormality recognition threshold and improve flexible response to environmental changes.

[0355] The present invention uses fuzzy logic to convert radiation intensity data into fuzzy values, and performs aggregation operations on the fuzzy values ​​of all data points to obtain the fuzzy matching degree of the overall radiation data set, and then calculates the fuzzy weight score of each data point, which improves the flexibility and accuracy of data processing and makes risk assessment more detailed and comprehensive.

[0356] The present invention comprehensively applies anomaly recognition thresholds and fuzzy weight scores to calculate the risk score of each data point, and combines risk classification thresholds to calculate the risk assessment results of each data point, thereby improving the accuracy of risk assessment results and enhancing the precision and response speed of radiation safety management.

[0357] The present disclosure may be a system, method and / or computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for causing a processor to implement various aspects of the present disclosure.

[0358] A computer-readable storage medium can be a tangible device that can hold and store instructions for use by an instruction execution device. A computer-readable storage medium can be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanical encoding device, such as a punched card or raised structure in a groove on which instructions are stored, and any suitable combination thereof. As used herein, a computer-readable storage medium is not to be construed as a transient signal per se, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (e.g., a light pulse passing through a fiber optic cable), or an electrical signal transmitted through an electrical wire.

[0359] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions to be stored in the computer-readable storage medium in each computing / processing device.

[0360] The computer program instructions for performing the operations of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The computer-readable program instructions may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, the state information of the computer-readable program instructions is used to personalize an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), so that the electronic circuit can execute the computer-readable program instructions, thereby implementing various aspects of the present disclosure.

[0361] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. A radiation source intelligent supervision method based on the Internet of Things, characterized in that: The method comprises: S1: Based on the Internet of Things, real-time monitoring of radiation source detection data streams is carried out, and the geographic, time and radiation intensity characteristic data sets of each data point are synchronously recorded, and the data are transmitted to the time series database for indexing, aggregation and normalization to generate radiation data sets; S2: Calculate the characteristic index for each data point in the radiation data set, and use the encoder to learn the characteristic index to optimize the encoder parameters. Use the optimized encoder parameters and characteristic index to calculate the abnormal index of each data point, and then calculate the abnormal recognition threshold; The calculation formula of the characteristic index is: in, Representative radiation dataset Middle The characteristic index of the data point, Representative radiation dataset Middle The value of the data point; The optimization formula for encoder parameters is: in, , Respectively represent the encoder parameters before and after optimization, represents the learning rate, (,) represents the loss function, Represents a set of feature indicators, represents the regularization coefficient; represents the square of the modulus; Represents encoder parameters Gradient operation of ; The calculation formula for the anomaly index of each data point is: in, Representative Anomaly indicators for data points, represents the reconstruction error, represents the weight factor, represents the average position of the feature index data set in the latent space, Representative potential locations of data points, represents the normalization factor, is a constant; The calculation formula for the anomaly recognition threshold is: in, represents the anomaly recognition threshold, represents the dynamic adjustment coefficient, represents the sensitivity of the control adjustment, represents the base of natural logarithm, represents the initial anomaly identification threshold, represents the parameter that controls the standard deviation multiple, , represents the adjustment factor; Represents the average value of the abnormal indicators of all data points, Represents the standard deviation of the anomaly indicator for all data points; represents the abnormal indicator dataset; is the data skewness coefficient; S3: Convert the radiation intensity of each data point in the radiation data set into a fuzzy value, and perform aggregation operation on the fuzzy values ​​of all data points to obtain the fuzzy matching degree of the overall data of the radiation data set, and then calculate the fuzzy weight score of each data point; S4: The anomaly identification threshold is combined with the fuzzy weight score to calculate the risk score of each data point, and the risk assessment result of each data point is calculated in combination with the risk classification threshold.

2. According to claim 1, a radiation source intelligent supervision method based on the Internet of Things is characterized by: In S1, based on IoT devices, the radiation source detection data stream from the radiation detector is monitored in real time, including geographic location tags, timestamps, and radiation intensity. The radiation detector is used as a data point, and the geographic, time, and radiation intensity feature data set of each data point is synchronously recorded, specifically: in, Represents the geographical, temporal and radiation intensity characteristic data set of the i-th data point; Represents geographical features, Represents time characteristics, Indicates the radiation intensity characteristics; Represents a geographic location tag. Indicates the timestamp, Indicates the radiation intensity; , represents the scaling factor, Indicates the time offset, An exponent representing the adjustment of radiation intensity.

3. The method for intelligently monitoring radiation sources based on the Internet of Things according to claim 2 is characterized in that: In S1, the geographic, time, and radiation intensity characteristic data set of each data point is transferred to the time series database, and the data is indexed to obtain the index data stream. : in, represents the index function, represents the eigenvalue extraction function, Indicates that the index is The weights corresponding to the features of represents the index of the feature, for The index is The eigenvalues ​​of the features.

4. The method for intelligently monitoring radiation sources based on the Internet of Things according to claim 3 is characterized in that: In S1, the index data stream Perform aggregation and normalization to generate a radiation dataset: in, represents the total number of data points, Represents an aggregation operation, Represents the normalization parameter.

5. The method for intelligently monitoring radiation sources based on the Internet of Things according to claim 1, characterized in that: In S3, the formula for converting radiation intensity into fuzzy value is: in, Representative The fuzzy value of the data point, is the radiation dataset Middle The radiation intensity of each data point, To control the slope of the curve, is the radiation dataset The mean radiation intensity of all data points in , Represents the base of natural logarithms.

6. The method for intelligently monitoring radiation sources based on the Internet of Things according to claim 5, characterized in that: In S3, the fuzzy values ​​of all data points are aggregated to obtain the fuzzy matching degree of the overall data of the radiation dataset: in, Represents the degree of fuzzy matching. is the total number of data points, It is The weight of the data point.

7. The method for intelligently monitoring radiation sources based on the Internet of Things according to claim 6, characterized in that: In S3, the calculation formula for the fuzzy weight score of each data point is: in, Representative The fuzzy weight score of each data point.

8. The method for intelligently monitoring radiation sources based on the Internet of Things according to claim 1, characterized in that: In S4, the risk score calculation formula for each data point is: in, Indicates Risk score for each data point, , is the parameter for adjusting the fuzzy weight gain, represents the anomaly recognition threshold, Representative The fuzzy weight score of the data points is Represents the base of natural logarithms.

9. The method for intelligently monitoring radiation sources based on the Internet of Things according to claim 1, characterized in that: In S4, the risk classification threshold calculation formula is: in, is the risk classification threshold, is the coefficient for adjusting the threshold; Represents the average value of the abnormal indicators of all data points, Represents the standard deviation of the anomaly indicator for all data points.

10. The method for intelligently monitoring radiation sources based on the Internet of Things according to claim 1, characterized in that: In S4, the risk assessment result calculation formula for each data point is: in, Representative The safety level index of the data points, , is the adjustment factor to control the threshold sensitivity, Indicates Risk score for each data point, is the risk classification threshold.

11. An intelligent monitoring system for radiation sources based on the Internet of Things, using the method according to any one of claims 1 to 10, characterized in that: The system comprises: The radiation data acquisition module is used to monitor the radiation detection data stream in real time based on the Internet of Things, synchronously record the geographical, time and radiation intensity characteristic data set of each data point, and transmit it to the time series database to index, aggregate and normalize the data to generate a radiation data set; The abnormality identification threshold calculation module is used to calculate the characteristic index of each data point in the radiation data set, and use the encoder to learn the characteristic index to optimize the encoder parameters, and use the optimized encoder parameters and characteristic indexes to calculate the abnormality index of each data point, and then calculate the abnormality identification threshold; The fuzzy weight scoring module is used to convert the radiation intensity of each data point in the radiation data set into a fuzzy value, and perform aggregation operations on the fuzzy values ​​of all data points to obtain the fuzzy matching degree of the overall data of the radiation data set, and then calculate the fuzzy weight score of each data point; The risk assessment module is used to combine the anomaly identification threshold with the fuzzy weight score to calculate the risk score of each data point, and calculate the risk assessment result of each data point in combination with the risk classification threshold.

12. A terminal comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is configured to operate according to the instructions to execute the steps of the method according to any one of claims 1-10.

13. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 10 are implemented.

Citation Information

Patent Citations

  • Power consumption information acquisition system operation state evaluation method based on fuzzy algorithm

    CN116976682A

  • Intelligent monitoring method for building safety big data and cloud platform system

    CN117236378A

  • Cable anomaly identification method and system based on time-frequency analysis

    CN118068132A

  • Building electrical safety automatic detection system

    CN118409155A

  • Power plant unattended equipment control method and device based on fuzzy logic

    CN118466413A