A chemical material storage management system based on the Internet of Things

Abstract

The invention relates to the technical field of chemical material storage management, and in particular to a chemical material storage management system based on the Internet of Things, which comprises a data collection module, an environment analysis module and a material analysis module, wherein: the data collection module collects material inventory data, historical warehouse environment data, and real-time warehouse environment data; the environment analysis module uses a decision tree algorithm to establish a decision tree model according to the historical warehouse environment data; and uses the constructed decision tree model to analyze the environmental state of the real-time warehouse environment data; if the environmental analysis result is abnormal, the feature with the highest importance score is selected according to the decision tree model, and is sent to the material analysis module as an abnormal environment data feature; the material analysis module calculates the Pearson correlation coefficient between the abnormal environment data feature and different material inventory data features, and adjusts the inventory corresponding to the material inventory data feature according to the calculated Pearson correlation coefficient.

Description

Technical Field

[0001] The present invention relates to the technical field of chemical material storage management, and in particular to a chemical material storage management system based on the Internet of Things. Background Art

[0002] Traditional chemical material storage management systems usually rely on static and periodic monitoring methods, and the management of warehouse environment and material inventory is mainly achieved through regular inspections and manual recording. This method has obvious disadvantages. The environmental monitoring methods in traditional systems are relatively simple, usually relying on limited sensors, and cannot fully and meticulously reflect the multi-dimensional changes in the warehouse environment. For example, important indicators such as temperature, humidity, and gas concentration may be ignored or delayed. Due to the lack of advanced data analysis tools, traditional systems cannot use historical data for in-depth analysis and prediction, cannot identify the key features of environmental anomalies, and cannot accurately judge the environmental status and warn of potential environmental risks. This leads to the situation that when the environment is abnormal, it is difficult for managers to detect problems in a timely manner and take effective countermeasures, which increases the instability of the storage environment and potential safety hazards.

[0003] In addition, the disadvantages of traditional systems in environmental analysis and processing are particularly obvious. They are unable to identify and process key abnormal characteristics, resulting in difficulty in timely resolution of environmental abnormalities, which increases the difficulty and risk of warehouse management. At the same time, traditional systems lack a refined adjustment mechanism and are unable to make dynamic adjustments based on environmental changes, resulting in insufficient accuracy and effectiveness in material inventory management, making it difficult to ensure the safe storage of materials and the stability of the environment. These disadvantages make it difficult for traditional chemical material warehouse management systems to cope with complex and changeable storage environments, and it is difficult to achieve the intelligent and scientific level of modern management systems. Summary of the invention

[0004] The purpose of the present invention is to provide a chemical material storage management system based on the Internet of Things to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solution: a chemical material storage management system based on the Internet of Things, characterized in that it includes a data collection module, an environmental analysis module and a material analysis module, wherein:

[0006] The data collection module collects material inventory data and historical warehouse environment data, as well as real-time warehouse environment data;

[0007] The environmental analysis module uses a decision tree algorithm to establish a decision tree model based on the historical warehouse environmental data, selects the feature with the largest information gain for splitting, and constructs a decision tree model; and uses the constructed decision tree model to analyze the environmental status of the real-time warehouse environmental data; wherein:

[0008] If the environmental analysis result is abnormal, the sum of information gains is calculated based on the abnormal features in the decision tree model as the importance score, and the feature with the highest importance score is selected and sent to the material analysis module as the abnormal environmental data feature;

[0009] The material analysis module calculates the Pearson correlation coefficient between the abnormal environment data feature and different material inventory data features, sorts the correlations between the abnormal environment data feature and the different material inventory data features according to the calculated Pearson correlation coefficient, and selects the material inventory data features corresponding to the abnormal environment data features in descending order of the Pearson correlation coefficient, and adjusts the inventory corresponding to the material inventory data features, wherein:

[0010] After the adjustment, the data collection module re-collects the real-time warehouse environment data, and the environment analysis module performs environment analysis on the re-collected real-time warehouse environment data until the environment analysis result is normal or the material inventory data features corresponding to all abnormal environment data features are selected.

[0011] As a further improvement of the present technical solution, the data collection module establishes a connection with the inventory management system to collect material inventory data, indicating the inventory quantity of each material in the current warehouse; and extracts past warehouse environment data from the historical database, including but not limited to temperature, humidity and gas concentration; and deploys a variety of sensors at various key locations in the warehouse to collect real-time warehouse environment data.

[0012] As a further improvement of the technical solution, the environment analysis module includes a model training unit, and the model training unit establishes a decision tree model according to the historical warehouse environment data, specifically including:

[0013] Set the historical warehouse environment data to ,in For the The feature vector of the records, is the corresponding environmental state, where each feature vector Contains several features;

[0014] Select the best split feature at each node and calculate the information gain , where the data set contained in the current node is , then a feature Information gain The calculation formula is as follows:

[0015] ,in It is a dataset Entropy of It is a feature The set of all values ​​of ; By feature The value of Partitioned subsets; Is a subset The number of samples; It is a dataset The total number of samples; Is a subset The entropy of ; where: the entropy is calculated by counting the number of samples in each category in the statistical data set and dividing it by the total number of samples in the data set to obtain the probability of each category, multiplying the negative logarithm of the probability of each category by the probability, and then summing them to obtain the entropy;

[0016] The training of the decision tree model starts from the root node, selects the feature with the largest information gain for splitting, and recursively constructs child nodes until the stopping condition is met.

[0017] As a further improvement of the technical solution, the environmental analysis module includes a prediction unit, which inputs the real-time warehouse environmental data into the trained decision tree model, specifically including:

[0018] Input real-time warehouse environment data into the trained decision tree model;

[0019] Starting from the root node, according to the value of the data feature value, the corresponding decision path is selected according to the judgment condition, and the data features involved in the decision path are recorded;

[0020] Following the decision path, after entering the next node, continue to select the corresponding decision path to enter the next node according to the judgment conditions of the current node and the value of the data feature value;

[0021] From the root node to the leaf node, the decision path is selected according to the value of the data feature value until the leaf node is reached. The leaf node corresponds to the environmental state, which indicates the environmental state corresponding to the real-time warehouse environmental data.

[0022] As a further improvement of the technical solution, the environment analysis module includes an abnormal data unit, which receives data features recorded when the environment state is abnormal, that is, abnormal features, and calculates the sum of information gains of the abnormal features on all nodes as the importance score, and the formula is as follows:

[0023] feature Importance Rating Expressed as ,in It is a feature The set of all nodes that appear, It is a feature At the node The information gain at

[0024] After calculating the importance scores corresponding to all abnormal features, the abnormal data unit selects the feature with the largest importance score as the feature with the greatest impact on the environmental anomaly, as the abnormal environmental data feature.

[0025] As a further improvement of the present technical solution, the material analysis module includes a correlation analysis unit, which calculates the Pearson correlation coefficient between the abnormal environment data characteristics and different material inventory data characteristics. The calculation formula of the Pearson correlation coefficient is as follows:

[0026] ,in is the Pearson correlation coefficient, with a value range of [-1,1]; and They are the data values ​​of abnormal environment data characteristics and material inventory data characteristics; and They are the means corresponding to the abnormal environmental data characteristics and material inventory; is the sample size.

[0027] As a further improvement of the present technical solution, the material analysis module includes an adjustment amount calculation unit, which sorts the correlation between the abnormal environment data features and different material inventory data features according to the calculated Pearson correlation coefficient, and selects the material inventory data features corresponding to the abnormal environment data features in order from large to small Pearson correlation coefficients, adjusts the material inventory, and determines the adjustment amount according to the absolute value of the Pearson correlation coefficient. The adjustment amount The calculation formula is as follows:

[0028] ,in is the basic adjustment amount set, is the absolute value of the Pearson correlation coefficient.

[0029] Compared with the prior art, the present invention has the following beneficial effects:

[0030] 1. This environmental analysis module in the chemical material storage management system based on the Internet of Things not only uses the decision tree algorithm to establish a decision tree model based on historical warehouse environmental data, but also selects the features with the largest information gain for splitting, and gradually builds an accurate decision tree model; and after the model is built, the module can also effectively use the decision tree model to analyze the real-time warehouse environmental data and quickly determine whether the current environmental state is normal; when the environmental analysis results show an abnormality, the module will record all the features that affect the abnormal state, and calculate the total information gain of these features in the nodes of the decision tree to obtain their importance scores. Subsequently, the module will select the features with the highest importance score as the main abnormal environmental data features and provide them to the material analysis module for subsequent processing. This method of identifying key features through information gain scoring not only improves the accuracy of environmental analysis, but also provides clear guidance for subsequent steps, thereby enhancing the intelligence level of the overall system.

[0031] 2. The material analysis module in the chemical material warehouse management system based on the Internet of Things accurately quantifies the relationship between environmental anomalies and material inventory by calculating the Pearson correlation coefficient between abnormal environmental data characteristics and different material inventory data characteristics; by calculating the correlation coefficient, the module can effectively identify which material inventory characteristics have the greatest impact on environmental anomalies, and sort them in order from large to small according to the correlation coefficient, and process these characteristics one by one; when making inventory adjustments, the module will determine the adjustment amount based on the absolute value of the correlation coefficient to ensure the effectiveness and scientificity of the adjustment measures. Specifically, the calculation formula of the adjustment amount is to multiply the set basic adjustment amount by the absolute value of the correlation coefficient, so that the adjustment amount is closely related to the correlation between the abnormal environmental data characteristics and the inventory data. This refined processing method not only improves the accuracy of inventory adjustment, but also prevents further deterioration of the warehouse environment to a large extent, ensuring the safety and stability of material storage.

[0032] 3. This chemical material storage management system based on the Internet of Things solves the problem of abnormal warehouse environment caused by unscientific storage of chemical materials through the loop iteration formed; first, the data collection module continuously collects material inventory data, historical warehouse environment data and real-time warehouse environment data to form a comprehensive data foundation. Then, the environmental analysis module uses the decision tree model to analyze the real-time warehouse environment data, quickly identify environmental anomalies and record key abnormal features. Then, these abnormal features and their importance scores are passed to the material analysis module, which determines and adjusts the inventory with the greatest impact on the environment by calculating the Pearson correlation coefficient between the abnormal environment data features and the material inventory data features; after the adjustment is completed, the data collection module re-collects the latest real-time warehouse environment data and starts a new round of environmental analysis and adjustment process. This loop iteration mechanism can timely discover and correct the unscientific aspects of chemical material storage through real-time monitoring and dynamic adjustment, effectively solve the abnormal warehouse environment problems, ensure the stability and safety of the storage environment, and the close cooperation between the modules enables the system to efficiently respond to environmental changes and maintain the best storage conditions, thereby improving the scientificity and intelligence level of chemical material storage management. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a schematic diagram of the overall module of the present invention;

[0034] Figure 2 It is a schematic diagram of the environment analysis module unit of the present invention;

[0035] Figure 3 It is a schematic diagram of the material analysis module unit of the present invention.

[0036] In the figure: 100, data collection module; 200, environmental analysis module; 201, model training unit; 202, prediction unit; 203, abnormal data unit; 300, material analysis module; 301, correlation analysis unit; 302, adjustment amount calculation unit. DETAILED DESCRIPTION

[0037] The following will be combined with the accompanying drawings in the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0038] See also Figure 1-3 The present invention provides a technical solution: a chemical material storage management system based on the Internet of Things, including a data collection module 100, an environmental analysis module 200 and a material analysis module 300.

[0039] The data collection module 100 establishes a connection with the inventory management system to collect material inventory data, indicating the inventory quantity of each material in the current warehouse; extracts past warehouse environment data from the historical database, including but not limited to temperature, humidity, gas concentration, and environmental status, etc., as historical warehouse environment data; deploys a variety of sensors at various key locations in the warehouse, such as temperature sensors, humidity sensors, gas detection sensors, etc., to collect real-time warehouse environment data.

[0040] The model training unit 201 in the environment analysis module 200 uses a decision tree algorithm to establish a decision tree model based on historical warehouse environment data, specifically including:

[0041] Data preparation: Set the historical warehouse environment data to ,in For the The feature vector of the records, is the corresponding environmental state, i.e., label (normal or abnormal), where each feature vector Contains several features, such as temperature ,humidity and gas concentration Etc., the eigenvector is expressed as ;

[0042] Decision tree model training: select the best split feature at each node and calculate information gain , where the data set contained in the current node is , then a feature Information gain The calculation formula is as follows:

[0043] ,in It is a dataset Entropy of It is a feature The set of all values ​​of ; By feature The value of Partitioned subsets; Is a subset The number of samples; It is a dataset The total number of samples; Is a subset The entropy of ; where: the entropy is calculated by counting the number of samples in each category in the statistical data set and dividing it by the total number of samples in the data set to obtain the probability of each category, multiplying the negative logarithm of the probability of each category by the probability, and then summing them to obtain the entropy;

[0044] The training of the decision tree model starts from the root node, selects the feature with the largest information gain for splitting, and recursively constructs child nodes until the stopping condition is met.

[0045] The prediction unit 202 in the environment analysis module 200 inputs the real-time warehouse environment data into the trained decision tree model, specifically including:

[0046] Input real-time warehouse environment data into the trained decision tree model;

[0047] Starting from the root node, according to the value of the data feature value, the corresponding decision path is selected according to the judgment condition, and the data features involved in the decision path are recorded;

[0048] Following the decision path, after entering the next node, continue to select the corresponding decision path to enter the next node according to the judgment conditions of the current node and the value of the data feature value;

[0049] From the root node to the leaf node, the decision path is selected according to the value of the data feature value until the leaf node is reached. The leaf node corresponds to the environmental state, which indicates the environmental state corresponding to the real-time warehouse environmental data; wherein: when the environmental state is abnormal, the recorded data features are sent to the abnormal data unit 203.

[0050] Since the feature selected by each node is determined by maximizing the information gain, the feature with a larger information gain has a greater impact on the classification result. The abnormal data unit 203 in the environment analysis module 200 receives the data features recorded when the environment state is abnormal, that is, the abnormal features, and calculates the sum of the information gains of the abnormal features on all nodes as the importance score, and the formula is as follows:

[0051] feature Importance Rating Expressed as ,in It is a feature The set of all nodes that appear, It is a feature At the node The information gain at

[0052] After calculating the importance scores corresponding to all abnormal features, the abnormal data unit 203 selects the feature with the largest importance score as the feature with the greatest impact on the environmental anomaly, as the abnormal environmental data feature.

[0053] The environment analysis module 200 realizes accurate analysis and prediction of the warehouse environment state through the model training unit 201 and the prediction unit 202. The model training unit 201 uses the decision tree algorithm to establish a decision tree model based on the historical warehouse environment data, selects the best split feature by calculating the information gain, and recursively constructs the child node to ensure the accuracy and effectiveness of the model; after the real-time environment data is input, the module can quickly analyze according to the decision tree model to determine whether the current environment state is abnormal; when the environment state is abnormal, the abnormal data unit 203 calculates the importance score of the abnormal feature and selects the feature with the highest score as the main abnormal data. This process not only improves the accuracy of environmental monitoring, but also effectively identifies and processes the features that have the greatest impact on environmental abnormalities through the calculation of importance scores, ensuring the safety and stability of the warehouse environment.

[0054] The correlation analysis unit 301 in the material analysis module 300 calculates the Pearson correlation coefficient between the abnormal environment data characteristics and the different material inventory data characteristics. The calculation formula of the Pearson correlation coefficient is as follows:

[0055] ,in is the Pearson correlation coefficient, with a value range of [-1,1]; and They are the data values ​​of abnormal environment data characteristics and material inventory data characteristics; and They are the means corresponding to the abnormal environmental data characteristics and material inventory; is the sample size.

[0056] The adjustment amount calculation unit 302 in the material analysis module 300 sorts the correlations between the abnormal environment data features and the different material inventory data features according to the calculated Pearson correlation coefficient, and selects the material inventory data features corresponding to the abnormal environment data features in order from large to small Pearson correlation coefficients, adjusts the material inventory, and determines the adjustment amount according to the absolute value of the Pearson correlation coefficient. The adjustment amount The calculation formula is as follows:

[0057] ,in is the basic adjustment amount set, is the absolute value of the Pearson correlation coefficient.

[0058] After the adjustment amount calculation unit 302 adjusts the material inventory, the data collection module 100 re-collects the real-time warehouse environment data, and the environment analysis module 200 performs environment analysis on the re-collected real-time warehouse environment data until the environment analysis results are normal or the material inventory data features corresponding to all abnormal environment data features are selected.

[0059] The material analysis module 300 realizes accurate analysis and processing of the relationship between environmental anomalies and material inventory through the correlation analysis unit 301 and the adjustment amount calculation unit 302; the correlation analysis unit 301 calculates the Pearson correlation coefficient between the abnormal environmental data characteristics and the material inventory data characteristics, accurately quantifies the relationship between the two, and sorts them according to the size of the correlation coefficient to identify the material inventory characteristics that have the greatest impact on environmental anomalies; the adjustment amount calculation unit 302 determines the adjustment amount according to the absolute value of the correlation coefficient, and scientifically adjusts the material inventory in turn; after the adjustment, real-time feedback is given to the data collection module 100 for a new round of data collection and environmental analysis to ensure the effectiveness and timeliness of the adjustment measures. Through this accurate correlation analysis and scientific inventory adjustment method, the module effectively prevents further deterioration of the warehouse environment, ensures the safe storage of chemical materials, and significantly improves the level of intelligent warehouse management.

[0060] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A chemical material storage management system based on the Internet of Things, characterized in that: It comprises a data collection module (100), an environment analysis module (200) and a material analysis module (300), wherein: The data collection module (100) collects material inventory data, historical warehouse environment data, and real-time warehouse environment data; The environment analysis module (200) uses a decision tree algorithm to establish a decision tree model based on the historical warehouse environment data, selects the feature with the largest information gain for splitting, and constructs a decision tree model, specifically including: Set the historical warehouse environment data to ,in For the The feature vector of the records, is the corresponding environmental state, where each feature vector Contains several features; Select the best split feature at each node and calculate the information gain , where the data set contained in the current node is , then a feature Information gain The calculation formula is as follows: ,in It is a dataset Entropy of It is a feature The set of all values ​​of ; By feature The value of Partitioned subsets; Is a subset The number of samples; It is a dataset The total number of samples; Is a subset The entropy of ; where: the entropy is calculated by counting the number of samples in each category in the statistical data set and dividing it by the total number of samples in the data set to obtain the probability of each category, multiplying the negative logarithm of the probability of each category by the probability, and then summing them to obtain the entropy; The training of the decision tree model starts from the root node, selects the feature with the largest information gain for splitting, and recursively constructs child nodes until the stopping condition is met; The environmental analysis module (200) uses the constructed decision tree model to analyze the environmental status of the real-time warehouse environmental data; wherein: If the environmental analysis result is abnormal, the sum of information gains is calculated based on the abnormal features in the decision tree model as the importance score, and the feature with the highest importance score is selected and sent to the material analysis module (300) as the abnormal environmental data feature; The material analysis module (300) calculates the Pearson correlation coefficient between the abnormal environment data feature and different material inventory data features, sorts the correlations between the abnormal environment data feature and the different material inventory data features according to the calculated Pearson correlation coefficient, selects the material inventory data features corresponding to the abnormal environment data features in descending order of the Pearson correlation coefficient, and adjusts the inventory corresponding to the material inventory data features, specifically: The material analysis module (300) comprises a correlation analysis unit (301), wherein the correlation analysis unit (301) calculates the Pearson correlation coefficient between the abnormal environment data characteristics and different material inventory data characteristics. The calculation formula of the Pearson correlation coefficient is as follows: ,in is the Pearson correlation coefficient, with a value range of [-1,1]; and They are the data values ​​of abnormal environment data characteristics and material inventory data characteristics; and They are the means corresponding to the abnormal environmental data characteristics and material inventory; is the sample size; The material analysis module (300) comprises an adjustment amount calculation unit (302), wherein the adjustment amount calculation unit (302) sorts the correlations between the abnormal environment data features and different material inventory data features according to the calculated Pearson correlation coefficient, and selects the material inventory data features corresponding to the abnormal environment data features in order from large to small Pearson correlation coefficients, adjusts the material inventory, and determines the adjustment amount according to the absolute value of the Pearson correlation coefficient. The calculation formula is as follows: ,in is the basic adjustment amount set, is the absolute value of the Pearson correlation coefficient; After the adjustment, the data collection module (100) recollects the real-time warehouse environment data, and the environment analysis module (200) performs environment analysis on the recollected real-time warehouse environment data until the environment analysis result is normal or the material inventory data features corresponding to all abnormal environment data features are selected; The environment analysis module (200) comprises a prediction unit (202), wherein the prediction unit (202) inputs real-time warehouse environment data into a trained decision tree model, specifically comprising: Input real-time warehouse environment data into the trained decision tree model; Starting from the root node, according to the value of the data feature value, the corresponding decision path is selected according to the judgment condition, and the data features involved in the decision path are recorded; Following the decision path, after entering the next node, continue to select the corresponding decision path to enter the next node according to the judgment conditions of the current node and the value of the data feature value; From the root node to the leaf node, the decision path is selected according to the value of the data feature value until the leaf node is reached. The leaf node corresponds to the environmental state, which indicates the environmental state corresponding to the real-time warehouse environmental data; The environment analysis module (200) comprises an abnormal data unit (203), wherein the abnormal data unit (203) receives data features recorded when the environment state is abnormal, i.e., abnormal features, and calculates the sum of information gains of the abnormal features on all nodes as an importance score, and the formula is as follows: feature Importance Rating Expressed as ,in It is a feature The set of all nodes that appear, It is a feature At the node The information gain at After calculating the importance scores corresponding to all abnormal features, the abnormal data unit (203) selects the feature with the largest importance score as the feature with the greatest impact on the environmental anomaly, as the abnormal environmental data feature.

2. The chemical material storage management system based on the Internet of Things according to claim 1 is characterized in that: The data collection module (100) establishes a connection with the inventory management system and is used to collect material inventory data, indicating the inventory quantity of each material in the current warehouse; and extracting past warehouse environment data from a historical database, including but not limited to temperature, humidity, and gas concentration; and deploying a variety of sensors at various key locations in the warehouse to collect real-time warehouse environment data.

Images