Community LPG intelligent turnover warehouse early warning system and method based on multi-source data fusion

Abstract

The present invention provides a community LPG smart turnover warehouse early warning system and method based on multi-source data fusion, including: a multi-dimensional data acquisition module, which is used to collect multi-source monitoring data in the turnover warehouse in real time based on multi-source sensing equipment, and pre-process the multi-source monitoring data, and perform multi-source data fusion on the pre-processed multi-source monitoring data based on data fusion logic to obtain fused data; a data analysis module, which is used to build a multi-dimensional early warning model for the turnover warehouse, and perform real-time analysis of the fused data based on the multi-dimensional early warning model for the turnover warehouse; an early warning module, which is used to determine the operating status of the turnover warehouse based on the real-time analysis results, and when there is an abnormality in the operating status, perform corresponding early warning operations based on the abnormal phenomenon. The accuracy of early warning monitoring is improved, emergency measures are taken in a timely manner, and the operating reliability and safety of the turnover warehouse are guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a community LPG smart turnover warehouse early warning system and method based on multi-source data fusion. Background Art

[0002] The community LPG smart turnover warehouse is a distribution point set up in the community, which can store bottled gas. The setting of the community LPG smart turnover warehouse is conducive to providing convenient services for community users while reducing logistics costs;

[0003] Due to the particularity of the LPG smart turnover warehouse, it is necessary to conduct real-time and effective early warning monitoring to facilitate timely safety response when abnormalities occur in the community LPG smart turnover warehouse;

[0004] However, at present, the early warning monitoring of LPG smart turnover warehouse can only monitor unilateral data, and the quality of the monitoring data cannot be guaranteed, which seriously affects the early warning monitoring effect of LPG smart turnover warehouse. Moreover, the current early warning monitoring rules are not flexible enough, and the early warning monitoring mechanism cannot be effectively updated and maintained. In addition, when analyzing the monitoring data, a combination of manual and machine analysis is often used, which greatly reduces the monitoring accuracy of LPG smart turnover warehouse, so that the operation safety of LPG smart turnover warehouse cannot be guaranteed.

[0005] Therefore, in order to overcome the above-mentioned defects, the present invention provides a community LPG intelligent turnover warehouse early warning system and method based on multi-source data fusion. Summary of the invention

[0006] The present invention provides a community LPG smart turnover warehouse early warning system and method based on multi-source data fusion, which is used to comprehensively and effectively monitor the turnover warehouse by adopting multi-source sensing equipment, realize accurate and effective acquisition of multi-source monitoring data in the turnover warehouse, and pre-process the acquired multi-source monitoring data to ensure the accuracy and reliability of the obtained multi-source monitoring data. At the same time, the pre-processed multi-source monitoring data is data-fused according to the data fusion logic to ensure that when the turnover warehouse is analyzed for early warning, monitoring data from different aspects can be comprehensively considered, thereby improving the accuracy of early warning monitoring. Secondly, a multi-dimensional early warning model for the turnover warehouse is constructed, and the fused data is analyzed in real time to realize accurate and effective acquisition of the real-time operation status of the turnover warehouse, thereby facilitating the timely discovery of abnormal situations and ensuring the timeliness of early warning. Finally, whether there is an abnormal situation in the turnover warehouse is judged according to the operating status of the turnover warehouse, and when an abnormality exists, a corresponding early warning operation is performed according to the abnormal phenomenon, which is convenient for the staff to quickly and effectively determine the abnormal type according to the early warning information, and also convenient for timely taking emergency measures, thereby ensuring the operational reliability and operational safety of the turnover warehouse.

[0007] The present invention provides a community LPG intelligent turnover warehouse early warning system based on multi-source data fusion, comprising:

[0008] The multi-dimensional data acquisition module is used to collect multi-source monitoring data in the turnover warehouse in real time based on multi-source sensing equipment, pre-process the multi-source monitoring data, and perform multi-source data fusion on the pre-processed multi-source monitoring data based on data fusion logic to obtain fused data;

[0009] The data analysis module is used to build a multi-dimensional early warning model for turnover warehouses and to perform real-time analysis on fused data based on the multi-dimensional early warning model for turnover warehouses;

[0010] The early warning module is used to determine the operating status of the turnover warehouse based on the real-time analysis results, and when there is an abnormality in the operating status, perform corresponding early warning operations based on the abnormal phenomenon.

[0011] Preferably, a community LPG intelligent turnover warehouse early warning system based on multi-source data fusion, a multi-dimensional data acquisition module, includes:

[0012] A position determination unit for:

[0013] Obtain early warning monitoring services for turnover warehouses, analyze the early warning monitoring services, and determine monitoring items for turnover warehouses;

[0014] Determine the sensor equipment and operating conditions of each monitoring project based on the business attributes of the monitoring project, and determine the deployment location of each sensor equipment in the turnover warehouse based on the spatial structure and operating conditions of the turnover warehouse;

[0015] A deployment guidance unit is used to generate a deployment map based on the deployment location and distribute the deployment map to each staff terminal for deployment guidance;

[0016] The data collection unit is used to adapt the monitoring parameters of each deployed sensor device based on the deployment guidance results, and control each sensor device to collect multi-source monitoring data in the turnover warehouse in real time based on the monitoring parameter adaptation results.

[0017] Preferably, a community LPG intelligent turnover warehouse early warning system based on multi-source data fusion, a multi-dimensional data acquisition module, includes:

[0018] A data acquisition unit, used to acquire the obtained multi-source monitoring data, and perform data cleaning on the multi-source monitoring data to obtain initial multi-source monitoring data;

[0019] Data pre-processing unit, used for:

[0020] Determine the data characteristics of each individual monitoring data in the multi-source monitoring data based on the representation object of the initial multi-source monitoring data, and determine the data dimension of each individual monitoring data based on the data characteristics;

[0021] The dimension distribution of multi-source monitoring data is obtained based on the data dimension of each single monitoring data, and a unified dimension is determined based on the dimension distribution;

[0022] Based on the unified dimension, the dimension reduction index of the data dimension of each single monitoring data is determined, and the data dimension of the multi-source monitoring data is unified based on the dimension reduction index to complete the preprocessing of the multi-source monitoring data.

[0023] Preferably, a community LPG intelligent turnover warehouse early warning system based on multi-source data fusion, a multi-dimensional data acquisition module, includes:

[0024] Fusion preparation unit for:

[0025] Extract the time parameters of the preprocessed multi-source monitoring data, and align the multi-source monitoring data within the same monitoring time period based on the time parameters;

[0026] Determine the data source of the multi-source monitoring data based on the data alignment result, and determine the monitoring business characteristics of the multi-source monitoring data based on the data source;

[0027] A fusion logic determination unit, used to determine the business association attributes between different monitoring data in the multi-source monitoring data based on the monitoring business characteristics, and determine the correlation and difference distribution characteristics between the multi-source monitoring data based on the business association attributes;

[0028] The data fusion logic of multi-source monitoring data is obtained based on the correlation and difference distribution characteristics;

[0029] Data fusion unit for:

[0030] Extracting principal component data of each source of monitoring data from multi-source monitoring data based on monitoring business characteristics, and obtaining characteristic representation data of each source of monitoring data based on the principal component data;

[0031] Based on the monitoring accuracy of each sensor device, the feature characterization data of the source monitoring data are assigned corresponding weights, and based on the weight assignment result, the feature characterization data and the corresponding weights are simultaneously input into the neural network for data fusion processing to obtain fused data.

[0032] Preferably, a community LPG intelligent turnover warehouse early warning system based on multi-source data fusion, the data fusion unit includes:

[0033] A data processing subunit, for:

[0034] Classify the received feature representation data according to the difference representation standard based on the neural network, and determine the causal chain between different feature representation data according to the business association attributes based on the classification results;

[0035] The feature representation data is split into fields, and the split fields are converted into data vectors;

[0036] A fusion directed graph between different feature representation data is obtained based on the causal chain, and node similarity values ​​between different feature representation data are determined based on the weight assignment result and the fusion directed graph;

[0037] The data fusion unit is used to determine the data fusion node between different feature representation data based on the node similarity value, and fuse the data vectors corresponding to the different feature representation data based on the data fusion node to obtain fused data.

[0038] Preferably, a community LPG intelligent turnover warehouse early warning system based on multi-source data fusion, the data analysis module includes:

[0039] A data retrieval unit is used to retrieve historical monitoring data and determine the warning object and the warning trigger condition and warning level corresponding to the warning object based on the turnover warehouse warning monitoring target;

[0040] Model building unit for:

[0041] Determine the warning features based on the warning object and the corresponding warning trigger conditions and warning levels, and select the prediction model framework from the model library based on the warning features;

[0042] Iteratively train the prediction model framework based on historical monitoring data and early warning features, and perform performance verification on the iterative training results one by one to obtain the multi-dimensional performance evaluation index value of the model after each training;

[0043] The multi-dimensional performance evaluation index values ​​are weighted averaged, and the weighted average result is compared with the benchmark threshold value. When the weighted average result is greater than the benchmark threshold, the iterative training result of the prediction model framework is determined to be qualified, and a multi-dimensional early warning model for the turnover warehouse is obtained;

[0044] Model optimization unit for:

[0045] Based on the operation requirements of the turnover warehouse multi-dimensional early warning model, the production environment is adapted to parameters, and based on the parameter adaptation results, the multi-dimensional early warning model of the turnover warehouse is adaptively adjusted in terms of the available space of the production environment;

[0046] Deploy the turnover warehouse multi-dimensional early warning model after the adaptive parameter level adjustment in the production environment, and set the update cycle of the turnover warehouse multi-dimensional early warning model based on the deployment results;

[0047] Based on the set results, the multi-dimensional early warning model of the turnover warehouse is periodically updated and optimized.

[0048] Preferably, a community LPG intelligent turnover warehouse early warning system based on multi-source data fusion, the data analysis module includes:

[0049] A data transmission unit, used to obtain the fused data in each monitoring time period, and transmit the fused data in each monitoring time period to the turnover warehouse multi-dimensional early warning model in real time based on the data interface;

[0050] Data parsing unit, used for:

[0051] Based on the multi-dimensional early warning model of the turnover warehouse, the fused data is split and analyzed to obtain the independent operation status results of different monitoring items in the turnover warehouse;

[0052] At the same time, the fusion data is correlated and analyzed based on the multi-dimensional early warning model of the turnover warehouse, and the limiting factors between different monitoring items in the turnover warehouse are determined based on the correlation analysis results;

[0053] The independent operation status results of different monitoring items are corrected based on the limiting factors, and the real-time analysis results of the fused data are obtained based on the correction results.

[0054] Preferably, a community LPG intelligent turnover warehouse early warning system based on multi-source data fusion, the early warning module includes:

[0055] A result acquisition unit is used to acquire the obtained real-time analysis results, and classify the real-time analysis results into environmental objects and human objects to obtain a two-dimensional state assessment data group;

[0056] A state determination unit for:

[0057] Based on the purpose of turnover warehouse early warning monitoring, the benchmark safe operation status parameters under each dimension are determined respectively, and the difference between the status assessment data group of each dimension and the corresponding benchmark safe operation status parameters are analyzed respectively;

[0058] Based on the difference analysis results, the deviation items and deviation parameters of the deviation items under each dimension are obtained, and based on the influence coefficient of the activity behavior of the human object on the state parameters of the monitoring items in the environmental object, the deviation parameters of the deviation items existing in the environmental object are corrected;

[0059] Based on the correction results, the deviation items and deviation parameters of the deviation items in each dimension are summarized to obtain the operating status of the turnover warehouse.

[0060] Preferably, a community LPG intelligent turnover warehouse early warning system based on multi-source data fusion, the early warning module includes:

[0061] A status acquisition unit, used to acquire the operating status of the turnover bin, and when the operating status of the turnover bin does not meet the operating requirements, determine an abnormal monitoring item based on the operating status;

[0062] Early warning unit for:

[0063] Determine the abnormal monitoring subject object based on the abnormal monitoring item, and determine the abnormal phenomenon of the abnormal monitoring subject object based on the operating status of the turnover warehouse;

[0064] Based on abnormal phenomena, corresponding warning measures are matched from the warning mechanism, and object warning operations are performed based on the corresponding warning measures.

[0065] The present invention provides a community LPG intelligent turnover warehouse early warning method based on multi-source data fusion, comprising:

[0066] Based on the multi-source sensor equipment, multi-source monitoring data in the turnover warehouse is collected in real time, and the multi-source monitoring data is pre-processed, and multi-source data fusion is performed on the pre-processed multi-source monitoring data based on the data fusion logic to obtain fused data;

[0067] Construct a multi-dimensional early warning model for turnover warehouses, and perform real-time analysis of fused data based on the multi-dimensional early warning model for turnover warehouses;

[0068] The operating status of the turnover warehouse is determined based on the real-time analysis results, and when there are abnormalities in the operating status, corresponding early warning operations are performed based on the abnormal phenomena.

[0069] Compared with the prior art, the present invention has the following beneficial effects:

[0070] 1. The working principle and beneficial effects of the above technical solution are: by adopting multi-source sensing equipment to conduct comprehensive and effective monitoring of the turnover warehouse, the multi-source monitoring data in the turnover warehouse can be accurately and effectively obtained, and the obtained multi-source monitoring data can be preprocessed to ensure the accuracy and reliability of the obtained multi-source monitoring data. At the same time, the pre-processed multi-source monitoring data is fused according to the data fusion logic to ensure that when conducting early warning analysis on the turnover warehouse, the monitoring data from different aspects can be comprehensively considered, thereby improving the accuracy of early warning monitoring. Secondly, a multi-dimensional early warning model for the turnover warehouse is constructed, and the fused data is analyzed in real time to achieve accurate and effective acquisition of the real-time operation status of the turnover warehouse, thereby facilitating the timely discovery of abnormal situations and ensuring the timeliness of early warning. Finally, according to the operating status of the turnover warehouse, it is judged whether there is an abnormal situation in the turnover warehouse, and when an abnormality exists, the corresponding early warning operation is performed according to the abnormal phenomenon, which is convenient for the staff to quickly and effectively determine the abnormal type according to the early warning information, and also convenient for timely taking emergency measures, thereby ensuring the operational reliability and operational safety of the turnover warehouse.

[0071] 2. By determining the time parameters of the pre-processed multi-source monitoring data, the multi-source monitoring data within the same monitoring time period can be aligned according to the time parameters, so as to facilitate the effective fusion of the multi-source monitoring data within the same monitoring time period, which provides convenience for the fusion of multi-source data. Secondly, according to the data alignment results, the correlation and difference distribution characteristics between the multi-source monitoring data within the same monitoring time period are determined to achieve accurate and effective acquisition of the data fusion logic of the multi-source monitoring data. Finally, by determining the characteristic representation data of the monitoring data from each source in the multi-source monitoring data, and assigning corresponding weights to different characteristic representation data according to the monitoring accuracy of each sensor device, it is convenient to lock the importance of different monitoring data according to the weight when performing data fusion, and realize accurate and reliable data fusion processing of the multi-source monitoring data according to the weight, which ensures the effect of data fusion and improves the accuracy and reliability of early warning of turnover warehouses through data fusion results.

[0072] 3. By retrieving historical monitoring data and determining the early warning features according to the turnover warehouse early warning monitoring targets, it is possible to select a suitable prediction model framework from the model library based on the early warning features. Secondly, the prediction model framework is iteratively trained through historical monitoring data and early warning features, and the performance of the iterative training results is verified to ensure the accuracy and reliability of the final training results. Finally, the production environment of the turnover warehouse multidimensional early warning model is parameter adapted to ensure the security and reliability of the deployment environment of the turnover warehouse multidimensional early warning model. At the same time, an update cycle is set for the turnover warehouse multidimensional early warning model to achieve regular updates of the turnover warehouse multidimensional early warning model, ensuring the operation effect of the turnover warehouse multidimensional early warning model and the reliability of the turnover warehouse early warning.

[0073] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures specifically pointed out in this application document.

[0074] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0076] Figure 1 This is a structural diagram of a community LPG smart turnover warehouse early warning system based on multi-source data fusion in an embodiment of the present invention;

[0077] Figure 2It is a structural diagram of a multi-dimensional data acquisition module in a community LPG smart turnover warehouse early warning system based on multi-source data fusion in an embodiment of the present invention;

[0078] Figure 3 The present invention is a flowchart of a community LPG smart turnover warehouse early warning method based on multi-source data fusion in an embodiment of the present invention. DETAILED DESCRIPTION

[0079] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0080] Embodiment 1:

[0081] This embodiment provides a community LPG smart turnover warehouse early warning system based on multi-source data fusion, such as Figure 1 As shown, including:

[0082] The multi-dimensional data acquisition module is used to collect multi-source monitoring data in the turnover warehouse in real time based on multi-source sensing equipment, pre-process the multi-source monitoring data, and perform multi-source data fusion on the pre-processed multi-source monitoring data based on data fusion logic to obtain fused data;

[0083] The data analysis module is used to build a multi-dimensional early warning model for turnover warehouses and to perform real-time analysis on fused data based on the multi-dimensional early warning model for turnover warehouses;

[0084] The early warning module is used to determine the operating status of the turnover warehouse based on the real-time analysis results, and when there is an abnormality in the operating status, perform corresponding early warning operations based on the abnormal phenomenon.

[0085] In this embodiment, the LPG smart turnover warehouse is a distribution point set up near the residents. This distribution point is the "smart turnover warehouse", which contains a variety of sensing equipment, smoke sensors, temperature sensors, microwaves, videos, etc., which can distinguish between people and animals, monitor the position of items in the warehouse, and issue an early warning when the temperature or smoke reaches the warning value. Among them, LPG is bottled gas, natural gas, and liquefied petroleum gas.

[0086] In this embodiment, the multi-source sensing device refers to various types of sensors, including smoke sensors, temperature sensors, microwaves, and video surveillance devices.

[0087] In this embodiment, the multi-source monitoring data refers to a variety of data obtained by monitoring the environment in the turnover warehouse and the activities of people through multi-source sensing equipment.

[0088] In this embodiment, preprocessing the multi-source monitoring data refers to performing operations such as data cleaning and format unification on the monitoring data from multiple different sources, thereby facilitating the fusion of the monitoring data from different sources.

[0089] In this embodiment, the data fusion logic refers to the standard for fusing multi-source monitoring data, including fusing multi-source monitoring data according to data relevance and difference.

[0090] In this embodiment, multi-source data fusion refers to associating monitoring data from different sources to obtain more comprehensive and rich monitoring results in the turnover warehouse, and the fusion result is the fusion data.

[0091] In this embodiment, constructing a multi-dimensional early warning model for a turnover warehouse refers to constructing the model framework after training it with historical monitoring data and early warning standards, in order to effectively analyze the multi-source monitoring data in the turnover warehouse.

[0092] In this embodiment, the abnormal phenomenon refers to an abnormal situation occurring in the turnover warehouse, such as an abnormal situation such as excessive temperature, smoke, or animal intrusion, so as to facilitate corresponding early warning operations according to the type of abnormality.

[0093] The working principle and beneficial effects of the above technical solution are: by adopting multi-source sensing equipment to conduct comprehensive and effective monitoring of the turnover warehouse, the multi-source monitoring data in the turnover warehouse can be accurately and effectively acquired, and the acquired multi-source monitoring data can be preprocessed to ensure the accuracy and reliability of the obtained multi-source monitoring data. At the same time, the pre-processed multi-source monitoring data is fused according to the data fusion logic to ensure that when conducting early warning analysis on the turnover warehouse, the monitoring data from different aspects can be comprehensively considered, thereby improving the accuracy of early warning monitoring. Secondly, a multi-dimensional early warning model for the turnover warehouse is constructed, and the fused data is analyzed in real time to achieve accurate and effective acquisition of the real-time operation status of the turnover warehouse, thereby facilitating the timely discovery of abnormal situations and ensuring the timeliness of early warning. Finally, according to the operating status of the turnover warehouse, it is judged whether there is an abnormal situation in the turnover warehouse, and when an abnormality exists, the corresponding early warning operation is performed according to the abnormal phenomenon, which is convenient for the staff to quickly and effectively determine the type of abnormality based on the early warning information, and also convenient for timely taking emergency measures, thereby ensuring the operational reliability and operational safety of the turnover warehouse.

[0094] Embodiment 2:

[0095] Based on Example 1, this embodiment provides a community LPG smart turnover warehouse early warning system based on multi-source data fusion, and a multi-dimensional data acquisition module, including:

[0096] A position determination unit for:

[0097] Obtain early warning monitoring services for turnover warehouses, analyze the early warning monitoring services, and determine monitoring items for turnover warehouses;

[0098] Determine the sensor equipment and operating conditions of each monitoring project based on the business attributes of the monitoring project, and determine the deployment location of each sensor equipment in the turnover warehouse based on the spatial structure and operating conditions of the turnover warehouse;

[0099] A deployment guidance unit is used to generate a deployment map based on the deployment location and distribute the deployment map to each staff terminal for deployment guidance;

[0100] The data collection unit is used to adapt the monitoring parameters of each deployed sensor device based on the deployment guidance results, and control each sensor device to collect multi-source monitoring data in the turnover warehouse in real time based on the monitoring parameter adaptation results.

[0101] In this embodiment, the early warning monitoring service is known in advance and is used to characterize the types of items to be monitored on the turnover warehouse and the specific monitoring requirements corresponding to each monitoring item.

[0102] In this embodiment, the monitoring items refer to specific businesses that need to be monitored on the turnover warehouse, including temperature, humidity, smoke, and human activities.

[0103] In this embodiment, the business attributes refer to the monitoring requirements and required monitoring conditions corresponding to the monitoring items.

[0104] In this embodiment, the operating conditions refer to the working environment required by each sensor device, etc.

[0105] In this embodiment, adapting the monitoring parameters of each deployed sensor device refers to configuring the sensitivity and monitoring angle of each sensor device.

[0106] The working principle and beneficial effects of the above technical solution are: by analyzing the early warning monitoring business of the turnover warehouse, the monitoring items required for the turnover warehouse can be accurately and effectively acquired; secondly, according to the business attributes of the monitoring items, the required sensor equipment and the operating conditions of the sensor equipment are locked, so as to facilitate the location deployment and parameter configuration of each sensor equipment; finally, according to the determined deployment location, the deployment guidance and the determined parameter configuration are used to adapt the monitoring parameters of each sensor equipment, thereby ensuring the operating stability of each sensor equipment, thereby realizing comprehensive and effective monitoring of the turnover warehouse through multi-source sensor equipment, and realizing accurate and effective acquisition of multi-source monitoring data.

[0107] Embodiment 3:

[0108] Based on Example 1, this embodiment provides a community LPG smart turnover warehouse early warning system based on multi-source data fusion, and a multi-dimensional data acquisition module, including:

[0109] A data acquisition unit, used to acquire the obtained multi-source monitoring data, and perform data cleaning on the multi-source monitoring data to obtain initial multi-source monitoring data;

[0110] Data pre-processing unit, used for:

[0111] Determine the data characteristics of each individual monitoring data in the multi-source monitoring data based on the representation object of the initial multi-source monitoring data, and determine the data dimension of each individual monitoring data based on the data characteristics;

[0112] The dimension distribution of multi-source monitoring data is obtained based on the data dimension of each single monitoring data, and a unified dimension is determined based on the dimension distribution;

[0113] Based on the unified dimension, the dimension reduction index of the data dimension of each single monitoring data is determined, and the data dimension of the multi-source monitoring data is unified based on the dimension reduction index to complete the preprocessing of the multi-source monitoring data.

[0114] In this embodiment, the initial multi-source monitoring data refers to the multi-source monitoring data without other interference factors obtained after operations such as removing abnormal data and missing data segments in the multi-source monitoring data.

[0115] In this embodiment, the representation object refers to the record subject corresponding to the initial multi-source monitoring data, such as temperature and humidity.

[0116] In this embodiment, the data feature refers to parameters such as the value fluctuation range corresponding to the monitoring data determined according to the characterization object and the data type of the corresponding monitoring data.

[0117] In this embodiment, the data dimension is used to characterize the complexity of different monomer monitoring data, wherein the monomer monitoring data is each type of data in the multi-source monitoring data, such as one monomer monitoring data for each of temperature and humidity.

[0118] In this embodiment, unifying the dimension means unifying all the individual monitoring data into the same dimension, thereby facilitating unification of multi-source monitoring data, and further facilitating fusion processing of the multi-source monitoring data.

[0119] In this embodiment, the dimensionality reduction index refers to a data index for reducing high-dimensional monomer monitoring data, for example, it may be to retain the principal component, thereby achieving the goal of dimensionality reduction.

[0120] The working principle and beneficial effects of the above technical solution are: by cleaning the obtained multi-source monitoring data, the accuracy and reliability of the multi-source monitoring data are ensured; secondly, by determining the data characteristics of each individual monitoring data in the multi-source monitoring data, the data dimensions of different individual monitoring data are unified according to the data characteristics, thereby ensuring the consistency of the data dimensions between the individual monitoring data, thereby facilitating the fusion processing of the multi-source monitoring data, improving the accuracy and reliability of the preprocessing of the multi-source monitoring data, and providing reliable guarantee for turnover warehouse early warning.

[0121] Embodiment 4:

[0122] Based on Example 1, this example provides a community LPG smart turnover warehouse early warning system based on multi-source data fusion, such as Figure 2 As shown, the multi-dimensional data acquisition module includes:

[0123] Fusion preparation unit for:

[0124] Extract the time parameters of the preprocessed multi-source monitoring data, and align the multi-source monitoring data within the same monitoring time period based on the time parameters;

[0125] Determine the data source of the multi-source monitoring data based on the data alignment result, and determine the monitoring business characteristics of the multi-source monitoring data based on the data source;

[0126] A fusion logic determination unit, used to determine the business association attributes between different monitoring data in the multi-source monitoring data based on the monitoring business characteristics, and determine the correlation and difference distribution characteristics between the multi-source monitoring data based on the business association attributes;

[0127] The data fusion logic of multi-source monitoring data is obtained based on the correlation and difference distribution characteristics;

[0128] Data fusion unit for:

[0129] Extracting principal component data of each source of monitoring data from multi-source monitoring data based on monitoring business characteristics, and obtaining characteristic representation data of each source of monitoring data based on the principal component data;

[0130] Based on the monitoring accuracy of each sensor device, the feature characterization data of the source monitoring data are assigned corresponding weights, and based on the weight assignment result, the feature characterization data and the corresponding weights are simultaneously input into the neural network for data fusion processing to obtain fused data.

[0131] In this embodiment, the time parameter refers to the specific time information of the acquired multi-source monitoring data, so as to facilitate grouping of the multi-source monitoring data within the same time period.

[0132] In this embodiment, data alignment refers to aggregating multi-source monitoring data within the same monitoring time period, so as to facilitate effective fusion of data within the same monitoring time period when performing data fusion.

[0133] In this embodiment, the data source refers to the device terminal corresponding to the multi-source monitoring data, that is, the corresponding sensor device.

[0134] In this embodiment, the monitoring service characteristics refer to the service characteristics corresponding to each monitoring data in the multi-source monitoring data, including the type of monitoring data, the type of monitoring, the scope of monitoring, and the like.

[0135] In this embodiment, the business association attribute is used to characterize the correlation between different monitoring data, for example, temperature will affect the internal gas pressure of bottled gas in the turnover warehouse.

[0136] In this embodiment, the difference distribution feature refers to the unrelated parameters between the multi-source monitoring data, that is, the unique attributes that characterize each monitoring item.

[0137] In this embodiment, the principal component data refers to the core data of each source monitoring data in the multi-source monitoring data, that is, the key data that can characterize the function of each source monitoring data.

[0138] In this embodiment, the characteristic characterization data refers to the characteristic data of the principal component data in the monitoring data from each source, that is, the key data in the principal component data.

[0139] The working principle and beneficial effects of the above technical solution are as follows: by determining the time parameters of the pre-processed multi-source monitoring data, the multi-source monitoring data within the same monitoring time period can be aligned according to the time parameters, so as to facilitate the effective fusion of the multi-source monitoring data within the same monitoring time period, and provide convenience for the fusion of multi-source data; secondly, according to the data alignment results, the correlation and difference distribution characteristics between the multi-source monitoring data within the same monitoring time period are determined, so as to achieve accurate and effective acquisition of the data fusion logic of the multi-source monitoring data; finally, by determining the characteristic representation data of the monitoring data from each source in the multi-source monitoring data, and assigning corresponding weights to different characteristic representation data according to the monitoring accuracy of each sensor device, it is convenient to lock the importance of different monitoring data according to the weight when performing data fusion, so as to achieve accurate and reliable data fusion processing of the multi-source monitoring data according to the weight, thereby ensuring the effect of data fusion and improving the accuracy and reliability of early warning of turnover warehouses through data fusion results.

[0140] Embodiment 5:

[0141] Based on Example 4, this embodiment provides a community LPG smart turnover warehouse early warning system based on multi-source data fusion, and the data fusion unit includes:

[0142] A data processing subunit, for:

[0143] Classify the received feature representation data according to the difference representation standard based on the neural network, and determine the causal chain between different feature representation data according to the business association attributes based on the classification results;

[0144] The feature representation data is split into fields, and the split fields are converted into data vectors;

[0145] A fusion directed graph between different feature representation data is obtained based on the causal chain, and node similarity values ​​between different feature representation data are determined based on the weight assignment result and the fusion directed graph;

[0146] The data fusion unit is used to determine the data fusion node between different feature representation data based on the node similarity value, and fuse the data vectors corresponding to the different feature representation data based on the data fusion node to obtain fused data.

[0147] In this embodiment, the difference characterization standard is known in advance and is used to characterize the characteristics of different monitoring data.

[0148] In this embodiment, the causal chain is used to characterize the interaction relationship between the feature characterization data corresponding to different monitoring data, so as to facilitate locking the fusion order between different monitoring data.

[0149] In this embodiment, field splitting refers to splitting feature representation data according to semantic structures.

[0150] In this embodiment, the data vector refers to the conversion of the split fields into corresponding vector forms through a neural network, thereby facilitating the fusion processing of different monitoring data.

[0151] In this embodiment, the fusion directed graph determines the fusion direction between different feature representation data according to the causal chain, thereby facilitating the fusion of different feature representation data.

[0152] In this embodiment, the node similarity value is used to characterize the relevant data segments between different feature representation data, so as to facilitate the determination of the data fusion node between different feature representation data according to the node similarity value.

[0153] In this embodiment, the data fusion node refers to a fusion segment between different feature representation data, that is, a parameter connecting different monitoring data.

[0154] The working principle and beneficial effects of the above technical solution are: by classifying different types of feature representation data, and according to the classification results and business-related attributes, the causal chain between different feature representation data is accurately and effectively determined; secondly, by splitting the feature representation data into fields and converting the split fields into corresponding data vectors, convenience is provided for data fusion; finally, the fusion directed graph between different feature representation data is determined through the causal chain, and the node similarity values ​​between different feature representation data are locked according to the weight assignment results and the fusion directed graph, and finally, different feature representation data are effectively fused according to the node similarity values, so as to facilitate effective early warning operations on turnover warehouses through fused data.

[0155] Embodiment 6:

[0156] Based on Example 1, this example provides a community LPG smart turnover warehouse early warning system based on multi-source data fusion, and the data analysis module includes:

[0157] A data retrieval unit is used to retrieve historical monitoring data and determine the warning object and the warning trigger condition and warning level corresponding to the warning object based on the turnover warehouse warning monitoring target;

[0158] Model building unit for:

[0159] Determine the warning features based on the warning object and the corresponding warning trigger conditions and warning levels, and select the prediction model framework from the model library based on the warning features;

[0160] Iteratively train the prediction model framework based on historical monitoring data and early warning features, and perform performance verification on the iterative training results one by one to obtain the multi-dimensional performance evaluation index value of the model after each training;

[0161] The multi-dimensional performance evaluation index values ​​are weighted averaged, and the weighted average result is compared with the benchmark threshold value. When the weighted average result is greater than the benchmark threshold, the iterative training result of the prediction model framework is determined to be qualified, and a multi-dimensional early warning model for the turnover warehouse is obtained;

[0162] Model optimization unit for:

[0163] Based on the operation requirements of the turnover warehouse multi-dimensional early warning model, the production environment is adapted to parameters, and based on the parameter adaptation results, the multi-dimensional early warning model of the turnover warehouse is adaptively adjusted in terms of the available space of the production environment;

[0164] Deploy the turnover warehouse multi-dimensional early warning model after the adaptive parameter level adjustment in the production environment, and set the update cycle of the turnover warehouse multi-dimensional early warning model based on the deployment results;

[0165] Based on the set results, the multi-dimensional early warning model of the turnover warehouse is periodically updated and optimized.

[0166] In this embodiment, the historical monitoring data is set in advance and is used as training parameters for training the multi-dimensional early warning model of the turnover warehouse.

[0167] In this embodiment, the turnover warehouse early warning monitoring target is known in advance, for example, it can be a high temperature warning, a high pressure warning, and an abnormal human activity warning, including the sensitivity of the warning.

[0168] In this embodiment, the warning object refers to warnings for temperature, humidity, pressure, etc.

[0169] In this embodiment, the warning triggering condition is set according to the warning requirement. For example, the warning condition may be triggered when the temperature in the turnover warehouse is higher than 30 degrees Celsius.

[0170] In this embodiment, the warning characteristics refer to the method of issuing warnings to the turnover warehouse, the object of the warning, the sensitivity of the warning, etc.

[0171] In this embodiment, the prediction model framework is a model framework that is selected from the model library and is suitable for the current turnover warehouse early warning, and is one of the models in the model library.

[0172] In this embodiment, successive performance verification refers to performing performance verification on each iterative training result to ensure that the final training result meets the performance requirements, wherein the multi-dimensional performance evaluation index value is the result obtained after each performance verification.

[0173] In this embodiment, the reference threshold is set in advance.

[0174] In this embodiment, the purpose of performing parameter adaptation on the production environment based on the operating requirements of the turnover warehouse multidimensional early warning model is to adjust the environment in which the turnover warehouse multidimensional early warning model needs to be deployed, thereby ensuring that the turnover warehouse multidimensional early warning model can operate stably in the corresponding production environment.

[0175] In this embodiment, adaptive parameter level adjustment of the multidimensional warning model of the turnover warehouse according to the available space of the production environment based on the parameter adaptation result means adaptively adjusting the size of the multidimensional warning model of the turnover warehouse according to the available space of the production environment, thereby ensuring that the multidimensional warning model of the turnover warehouse can be deployed in the production environment. For example, the multidimensional warning model of the turnover warehouse can be compressed or streamlined.

[0176] The working principle and beneficial effects of the above technical solution are: by retrieving historical monitoring data and determining the early warning features according to the turnover warehouse early warning monitoring targets, a suitable prediction model framework is selected from the model library according to the early warning features; secondly, the prediction model framework is iteratively trained through historical monitoring data and early warning features, and the performance of the iterative training results is verified to ensure the accuracy and reliability of the training results obtained in the end; finally, the production environment of the multidimensional early warning model of the turnover warehouse is parameter adapted to ensure the security and reliability of the deployment environment of the multidimensional early warning model of the turnover warehouse; at the same time, an update cycle is set for the multidimensional early warning model of the turnover warehouse to achieve regular updates of the multidimensional early warning model of the turnover warehouse, thereby ensuring the operation effect of the multidimensional early warning model of the turnover warehouse and the reliability of the early warning for the turnover warehouse.

[0177] Embodiment 7:

[0178] Based on Example 1, this example provides a community LPG smart turnover warehouse early warning system based on multi-source data fusion, and the data analysis module includes:

[0179] A data transmission unit, used to obtain the fused data in each monitoring time period, and transmit the fused data in each monitoring time period to the turnover warehouse multi-dimensional early warning model in real time based on the data interface;

[0180] Data parsing unit, used for:

[0181] Based on the multi-dimensional early warning model of the turnover warehouse, the fused data is split and analyzed to obtain the independent operation status results of different monitoring items in the turnover warehouse;

[0182] At the same time, the fusion data is correlated and analyzed based on the multi-dimensional early warning model of the turnover warehouse, and the limiting factors between different monitoring items in the turnover warehouse are determined based on the correlation analysis results;

[0183] The independent operation status results of different monitoring items are corrected based on the limiting factors, and the real-time analysis results of the fused data are obtained based on the correction results.

[0184] In this embodiment, the independent operation status result refers to the analysis result corresponding to each type of monitoring data obtained after the fusion data is split and analyzed through the turnover warehouse multi-dimensional early warning model.

[0185] In this embodiment, the limiting factor is used to characterize the mutual limiting degree between different monitoring items, and the purpose is to revise the independent operation status results.

[0186] The working principle and beneficial effect of the above technical solution are: through the obtained multi-dimensional early warning model of the turnover warehouse, the obtained fusion data is received and analyzed in real time, so as to realize real-time and effective determination of the operating status of the turnover warehouse, thereby facilitating timely and effective early warning operations on the turnover warehouse according to the analysis results.

[0187] Embodiment 8:

[0188] Based on Example 1, this embodiment provides a community LPG smart turnover warehouse early warning system based on multi-source data fusion, and the early warning module includes:

[0189] A result acquisition unit is used to acquire the obtained real-time analysis results, and classify the real-time analysis results into environmental objects and human objects to obtain a two-dimensional state assessment data group;

[0190] A state determination unit for:

[0191] Based on the purpose of turnover warehouse early warning monitoring, the benchmark safe operation status parameters under each dimension are determined respectively, and the difference between the status assessment data group of each dimension and the corresponding benchmark safe operation status parameters are analyzed respectively;

[0192] Based on the difference analysis results, the deviation items and deviation parameters of the deviation items under each dimension are obtained, and based on the influence coefficient of the activity behavior of the human object on the state parameters of the monitoring items in the environmental object, the deviation parameters of the deviation items existing in the environmental object are corrected;

[0193] Based on the correction results, the deviation items and deviation parameters of the deviation items in each dimension are summarized to obtain the operating status of the turnover warehouse.

[0194] In this embodiment, the environmental object refers to parameters such as the temperature, humidity, and gas pressure of the bottled gas in the turnover warehouse.

[0195] In this embodiment, the human object refers to a person or animal active in the turnover warehouse.

[0196] In this embodiment, the two-dimensional state assessment data group refers to the data obtained by classifying the real-time results, wherein the classification criteria are environmental objects and human objects.

[0197] In this embodiment, the benchmark safe operation state parameter refers to the state parameter achieved by the normal operation of the turnover warehouse in each dimension.

[0198] In this embodiment, the deviation item refers to the monitoring item that does not meet the benchmark safe operation status parameter in each dimension, for example, it may be the temperature in the environmental object.

[0199] In this embodiment, the deviation parameter refers to the difference between the operating state of the deviation item and the corresponding reference safe operating state parameter.

[0200] In this embodiment, the influence coefficient refers to the degree of influence of the activities of human objects on the state parameters of the monitored items in the environmental objects. For example, human activities will affect the type and temperature of the gas in the turnover warehouse. The influence coefficient is obtained by analyzing multiple monitoring data.

[0201] The working principle and beneficial effects of the above technical solution are: by classifying the real-time analysis results, and locking the benchmark safe operation status parameters under each dimension according to the classification results, it is possible to perform difference analysis on the status assessment data group under each dimension and the corresponding benchmark safe operation status parameters, thereby achieving effective determination of deviation items and deviation parameters, and achieving rapid and accurate determination of abnormal conditions in the turnover warehouse, thereby achieving acquisition of the operating status in the turnover warehouse, which provides convenience for turnover warehouse early warning.

[0202] Embodiment 9:

[0203] Based on Example 1, this embodiment provides a community LPG smart turnover warehouse early warning system based on multi-source data fusion, and the early warning module includes:

[0204] A status acquisition unit, used to acquire the operating status of the turnover bin, and when the operating status of the turnover bin does not meet the operating requirements, determine an abnormal monitoring item based on the operating status;

[0205] Early warning unit for:

[0206] Determine the abnormal monitoring subject object based on the abnormal monitoring item, and determine the abnormal phenomenon of the abnormal monitoring subject object based on the operating status of the turnover warehouse;

[0207] Based on abnormal phenomena, corresponding warning measures are matched from the warning mechanism, and object warning operations are performed based on the corresponding warning measures.

[0208] In this embodiment, the abnormal monitoring item refers to the abnormal situation existing in the turnover warehouse.

[0209] In this embodiment, the abnormal monitoring subject object refers to the specific correspondence of the abnormality occurring in the turnover warehouse, for example, it may be one or more of temperature, humidity, and gas type and content.

[0210] In this embodiment, the abnormal phenomenon refers to the specific degree of abnormality of the abnormal monitoring subject object.

[0211] The working principle and beneficial effects of the above technical solution are: by analyzing the operating status of the turnover warehouse, the abnormal monitoring items in the turnover warehouse can be quickly and effectively locked, and the abnormal phenomenon can be determined according to the locking result, and the corresponding early warning measures can be matched from the early warning mechanism according to the abnormal phenomenon to perform corresponding early warning operations, thereby ensuring the timeliness of early warning of abnormalities in the turnover warehouse and the reliability of solving abnormal phenomena, and ensuring the safe operation of the turnover warehouse.

[0212] Embodiment 10:

[0213] This embodiment provides a community LPG smart turnover warehouse early warning method based on multi-source data fusion, such as Figure 3 As shown, including:

[0214] Step 1: Based on the multi-source sensing equipment, multi-source monitoring data in the turnover warehouse is collected in real time, and the multi-source monitoring data is pre-processed, and multi-source data fusion is performed on the pre-processed multi-source monitoring data based on the data fusion logic to obtain fused data;

[0215] Step 2: Build a multi-dimensional early warning model for turnover warehouses, and perform real-time analysis on the fused data based on the multi-dimensional early warning model for turnover warehouses;

[0216] Step 3: Determine the operating status of the turnover warehouse based on the real-time analysis results, and when there is an abnormality in the operating status, perform corresponding warning operations based on the abnormal phenomenon.

[0217] The working principle and beneficial effects of the above technical solution are: by adopting multi-source sensing equipment to conduct comprehensive and effective monitoring of the turnover warehouse, the multi-source monitoring data in the turnover warehouse can be accurately and effectively acquired, and the acquired multi-source monitoring data can be preprocessed to ensure the accuracy and reliability of the obtained multi-source monitoring data. At the same time, the pre-processed multi-source monitoring data is fused according to the data fusion logic to ensure that when conducting early warning analysis on the turnover warehouse, the monitoring data from different aspects can be comprehensively considered, thereby improving the accuracy of early warning monitoring. Secondly, a multi-dimensional early warning model for the turnover warehouse is constructed, and the fused data is analyzed in real time to achieve accurate and effective acquisition of the real-time operation status of the turnover warehouse, thereby facilitating the timely discovery of abnormal situations and ensuring the timeliness of early warning. Finally, according to the operating status of the turnover warehouse, it is judged whether there is an abnormal situation in the turnover warehouse, and when an abnormality exists, the corresponding early warning operation is performed according to the abnormal phenomenon, which is convenient for the staff to quickly and effectively determine the type of abnormality based on the early warning information, and also convenient for timely taking emergency measures, thereby ensuring the operational reliability and operational safety of the turnover warehouse.

[0218] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A community LPG intelligent turnover warehouse early warning system based on multi-source data fusion, characterized in that: include: The multi-dimensional data acquisition module is used to collect multi-source monitoring data in the turnover warehouse in real time based on multi-source sensing equipment, and pre-process the multi-source monitoring data, and perform multi-source data fusion on the pre-processed multi-source monitoring data based on data fusion logic to obtain fused data; wherein, LPG refers to bottled gas, natural gas, and liquefied petroleum gas; the smart turnover warehouse is a distribution point set up near residents; multi-source sensing equipment refers to various types of sensors; multi-source monitoring data refers to various data obtained after monitoring the environment and human activities in the turnover warehouse through multi-source sensing equipment; data fusion logic refers to the standard for data fusion of multi-source monitoring data; The data analysis module is used to build a multi-dimensional early warning model for turnover warehouses and to perform real-time analysis on fused data based on the multi-dimensional early warning model for turnover warehouses; The early warning module is used to determine the operating status of the turnover warehouse based on the real-time analysis results, and to perform corresponding early warning operations based on the abnormal phenomena when there are abnormalities in the operating status; Among them, the multi-dimensional data acquisition module includes: Fusion preparation unit for: Extract the time parameters of the preprocessed multi-source monitoring data, and align the multi-source monitoring data within the same monitoring time period based on the time parameters; where the time parameters refer to the specific time information of the multi-source monitoring data; Determine the data source of the multi-source monitoring data based on the data alignment result, and determine the monitoring business characteristics of the multi-source monitoring data based on the data source; wherein the data source refers to the sensor equipment corresponding to the multi-source monitoring data; the monitoring business characteristics refer to the business characteristics corresponding to each monitoring data in the multi-source monitoring data, including the type of monitoring data, the type of monitoring and the scope of monitoring; A fusion logic determination unit is used to: Determine the business correlation attributes between different monitoring data in the multi-source monitoring data based on the monitoring business characteristics, and determine the correlation and difference distribution characteristics between the multi-source monitoring data based on the business correlation attributes; wherein the business correlation attributes are used to characterize the correlation between different monitoring data; the difference distribution characteristics refer to the unrelated parameters between the multi-source monitoring data, that is, the unique attributes that characterize each monitoring item; The data fusion logic of multi-source monitoring data is obtained based on the correlation and difference distribution characteristics; Data fusion unit for: Extract the principal component data of each source monitoring data in the multi-source monitoring data based on the monitoring business characteristics, and obtain the characteristic representation data of each source monitoring data based on the principal component data; wherein the principal component data refers to the core data of each source monitoring data in the multi-source monitoring data; the characteristic representation data refers to the characteristic data of the principal component data in the monitoring data of each source, that is, the key data in the principal component data; Based on the monitoring accuracy of each sensor device, the characteristic representation data of the monitoring data from each source are assigned corresponding weights, and based on the weight assignment result, the characteristic representation data and the corresponding weights are simultaneously input into the neural network for data fusion processing to obtain fused data; The data fusion unit includes: A data processing subunit, for: Based on the neural network, the received feature representation data is classified according to the difference representation standard, and the causal chain between different feature representation data is determined according to the business association attributes based on the classification result; wherein the difference representation standard is used to characterize the characteristics of different monitoring data; the causal chain is used to characterize the interaction relationship between the feature representation data corresponding to different monitoring data; The feature representation data is split into fields, and the split fields are converted into data vectors; wherein field splitting refers to splitting the feature representation data according to the semantic structure; A fusion directed graph between different feature representation data is obtained based on the causal chain, and node similarity values ​​between different feature representation data are determined based on the weight assignment result and the fusion directed graph; wherein the fusion directed graph is the fusion direction between different feature representation data determined according to the causal chain; and the node similarity value is used to represent related data fragments between different feature representation data; The data fusion subunit is used to determine the data fusion node between different feature representation data based on the node similarity value, and fuse the data vectors corresponding to the different feature representation data based on the data fusion node to obtain fused data; wherein the data fusion node refers to the fusion segment between different feature representation data, that is, the parameter connecting different monitoring data; The data analysis module includes: A data retrieval unit is used to retrieve historical monitoring data and determine the warning object and the warning trigger condition and warning level corresponding to the warning object based on the turnover warehouse warning monitoring target; Model building unit for: Determine the warning features based on the warning object and the corresponding warning trigger conditions and warning levels, and select the prediction model framework from the model library based on the warning features; the warning features refer to the way to warn the turnover warehouse, the object of the warning, and the sensitivity of the warning; Iteratively train the prediction model framework based on historical monitoring data and early warning features, and perform performance verification on the iterative training results one by one to obtain the multi-dimensional performance evaluation index value of the model after each training; The multi-dimensional performance evaluation index values ​​are weighted averaged, and the weighted average result is compared with the benchmark threshold value. When the weighted average result is greater than the benchmark threshold, the iterative training result of the prediction model framework is determined to be qualified, and a multi-dimensional early warning model for the turnover warehouse is obtained; Model optimization unit for: Based on the operation requirements of the turnover warehouse multidimensional early warning model, the production environment is parameter-adapted, and based on the parameter adaptation results, the multidimensional early warning model of the turnover warehouse is adaptively adjusted in terms of the parameter level according to the available space of the production environment; wherein, the purpose of parameter adaptation of the production environment is to adjust the environment where the turnover warehouse multidimensional early warning model needs to be deployed; adaptively adjusting the parameter level of the turnover warehouse multidimensional early warning model refers to adaptively adjusting the size of the turnover warehouse multidimensional early warning model; Deploy the turnover warehouse multi-dimensional early warning model after the adaptive parameter level adjustment in the production environment, and set the update cycle of the turnover warehouse multi-dimensional early warning model based on the deployment results; Periodically update and optimize the turnover warehouse multi-dimensional early warning model based on the set results; The data analysis module also includes: A data transmission unit, used to obtain the fused data in each monitoring time period, and transmit the fused data in each monitoring time period to the turnover warehouse multi-dimensional early warning model in real time based on the data interface; Data parsing unit, used for: Based on the multi-dimensional early warning model of the turnover warehouse, the fused data is split and analyzed to obtain the independent operation status results of different monitoring items in the turnover warehouse; where the independent operation status results refer to the analysis results corresponding to each type of monitoring data; At the same time, the fusion data is correlated and analyzed based on the multi-dimensional early warning model of the turnover warehouse, and the limiting factors between different monitoring items in the turnover warehouse are determined based on the correlation analysis results; among them, the limiting factors are used to characterize the degree of mutual limitation between different monitoring items; The independent operation status results of different monitoring items are corrected based on the limiting factors, and the real-time analysis results of the fused data are obtained based on the correction results.

2. The community LPG intelligent turnover warehouse early warning system based on multi-source data fusion according to claim 1 is characterized in that: Multi-dimensional data acquisition module, including: A position determination unit for: Obtain the early warning monitoring business for the turnover warehouse, analyze the early warning monitoring business, and determine the monitoring items for the turnover warehouse; the early warning monitoring business is used to characterize the types of items monitored for the turnover warehouse and the specific monitoring requirements corresponding to each monitoring item; the monitoring item refers to the specific business that needs to be monitored for the turnover warehouse, including but not limited to temperature, humidity, smoke, and human activities; Based on the business attributes of the monitoring items, the sensor equipment and operating conditions of each sensor equipment required for each monitoring item are determined, and based on the spatial structure and operating conditions of the turnover warehouse, the deployment location of each sensor equipment in the turnover warehouse is determined; wherein the business attributes refer to the monitoring requirements and required monitoring conditions corresponding to the monitoring items; A deployment guidance unit is used to generate a deployment map based on the deployment location and distribute the deployment map to each staff terminal for deployment guidance; The data collection unit is used to adapt the monitoring parameters of each deployed sensor device based on the deployment guidance results, and control each sensor device to collect multi-source monitoring data in the turnover warehouse in real time based on the monitoring parameter adaptation results.

3. The community LPG intelligent turnover warehouse early warning system based on multi-source data fusion according to claim 1 is characterized in that: Multi-dimensional data acquisition module, including: A data acquisition unit, used to acquire the obtained multi-source monitoring data, and perform data cleaning on the multi-source monitoring data to obtain initial multi-source monitoring data; Data pre-processing unit, used for: Based on the characterization object of the initial multi-source monitoring data, the data characteristics of each monomer monitoring data in the multi-source monitoring data are determined, and based on the data characteristics, the data dimensions of each monomer monitoring data are determined; wherein, the characterization object refers to the record subject corresponding to the initial multi-source monitoring data; the data characteristics refer to the value fluctuation range corresponding to the monitoring data determined according to the characterization object and the data type of the corresponding monitoring data; the monomer monitoring data refers to each type of data in the multi-source monitoring data; the data dimension is used to characterize the complexity of different monomer monitoring data; The dimension distribution of multi-source monitoring data is obtained based on the data dimension of each single monitoring data, and a unified dimension is determined based on the dimension distribution; Based on the unified dimension, the dimension reduction index of the data dimension of each individual monitoring data is determined, and the data dimension of the multi-source monitoring data is unified based on the dimension reduction index to complete the preprocessing of the multi-source monitoring data; wherein, the dimension reduction index refers to the data indicator for reducing the high-dimensional individual monitoring data.

4. The community LPG intelligent turnover warehouse early warning system based on multi-source data fusion according to claim 1 is characterized in that: Early warning module, including: A result acquisition unit is used to acquire the obtained real-time analysis results, and classify the real-time analysis results into environmental objects and human objects to obtain a two-dimensional state assessment data group; A state determination unit for: Based on the turnover warehouse early warning monitoring target, the benchmark safe operation state parameters under each dimension are determined respectively, and the difference between the state evaluation data group of each dimension and the corresponding benchmark safe operation state parameters are analyzed respectively; wherein, the benchmark safe operation state parameters refer to the state parameters achieved by the normal operation of the turnover warehouse; Based on the difference analysis results, the deviation items and deviation parameters of the deviation items in each dimension are obtained, and based on the influence coefficient of the activity behavior of the human object on the state parameters of the monitoring items in the environmental object, the deviation parameters of the deviation items existing in the environmental object are corrected; wherein, the deviation item refers to the monitoring item that does not meet the benchmark safe operation state parameters; the deviation parameter refers to the difference between the operation state of the deviation item and the corresponding benchmark safe operation state parameters; Based on the correction results, the deviation items and deviation parameters of the deviation items in each dimension are summarized to obtain the operating status of the turnover warehouse.

5. The community LPG intelligent turnover warehouse early warning system based on multi-source data fusion according to claim 1 is characterized in that: Early warning module, including: A state acquisition unit, used to acquire the operating state of the turnover warehouse, and when the operating state of the turnover warehouse does not meet the operating requirements, determine the abnormal monitoring items based on the operating state; Early warning unit for: Determine the abnormal monitoring subject object based on the abnormal monitoring item, and determine the abnormal phenomenon of the abnormal monitoring subject object based on the operating status of the turnover warehouse; Based on abnormal phenomena, corresponding warning measures are matched from the warning mechanism, and corresponding warning operations are performed based on the corresponding warning measures.

6. A community LPG smart turnover warehouse early warning method based on multi-source data fusion, the method is implemented by the system described in claim 1, characterized in that: The method comprises: Based on the multi-source sensor equipment, multi-source monitoring data in the turnover warehouse is collected in real time, and the multi-source monitoring data is pre-processed, and multi-source data fusion is performed on the pre-processed multi-source monitoring data based on the data fusion logic to obtain fused data; Construct a multi-dimensional early warning model for turnover warehouses, and perform real-time analysis of fused data based on the multi-dimensional early warning model for turnover warehouses; The operating status of the turnover warehouse is determined based on the real-time analysis results, and when there are abnormalities in the operating status, corresponding early warning operations are performed based on the abnormal phenomena.

Images