Professional warehouse control method and system based on artificial intelligence and storage medium

Abstract

The invention relates to the technical field of data processing, and discloses a professional warehouse control method and system based on artificial intelligence and a storage medium. The method comprises the steps of collecting professional warehouse data through a multi-source sensing device, and performing fusion to obtain a physical resource multi-source fusion data set; carrying out classification by adopting a transfer learning algorithm to obtain engineering material classification result data; mapping the reserve quota standard to obtain state evaluation data; carrying out demand prediction to obtain deployment scheme data; decomposing the operation process to obtain operation instruction data; and performing carbon emission analysis to obtain environment management decision data. According to the invention, the efficiency and accuracy of professional warehouse control based on artificial intelligence are improved.

Description

Technical Field

[0001] The present application relates to the field of data processing, and in particular to a professional warehouse control method, system and storage medium based on artificial intelligence. Background Art

[0002] With the continuous expansion of the scale of power grid enterprises and the continuous development of their business, the management of professional warehouses faces increasingly complex challenges. The existing professional warehouse control methods mainly adopt the traditional manual management mode, managing materials through regular inventory, manual records and experience judgment. Although some companies have introduced warehouse management systems, they are mainly limited to basic in and out warehouse management and inventory statistics, and lack intelligent analysis and decision-making capabilities. In terms of hazardous waste resource management, existing technologies mainly rely on manual judgment and fixed processes for processing, and cannot dynamically adjust the operation plan according to actual conditions. At the same time, under the background of dual carbon goals, the energy consumption and carbon emissions of professional warehouses are becoming increasingly prominent, but the existing management methods lack refined control measures for energy consumption and carbon emissions.

[0003] The existing technology has the following shortcomings: First, the material classification management lacks intelligent means and cannot effectively process multi-source heterogeneous data, resulting in low material classification accuracy and poor efficiency; second, in the management of hazardous waste resources, the existing methods cannot achieve intelligent decomposition of work tasks and optimal allocation of resources, which increases safety risks; third, the management of energy consumption and carbon emissions is relatively extensive, lacks scientific evaluation methods and optimization strategies, and it is difficult to meet the requirements of refined management; finally, there is a lack of effective data association and coordination mechanism between the various management links, which affects the overall management effect. Summary of the invention

[0004] The present application provides a professional warehouse control method, system and storage medium based on artificial intelligence, which are used to realize intelligent management and green operation of professional warehouses through technical means such as multi-source data fusion, intelligent classification optimization, and low-carbon index monitoring.

[0005] In the first aspect, the present application provides a professional warehouse control method based on artificial intelligence, and the professional warehouse control method based on artificial intelligence includes: collecting physical inventory information, physical data of outbound waiting, physical status of tools and tools, and location data of delivery vehicles within the professional warehouse through multi-source sensing equipment, performing quality inspection and preprocessing on the collected data, and obtaining a multi-source fusion data set of physical resources through data fusion processing; classifying and processing the material information through feature extraction based on the multi-source fusion data set of physical resources, and optimizing the extracted features using a transfer learning algorithm to obtain engineering material classification result data; based on the multi-source fusion data set of physical resources and the engineering material classification result data, The warehouse reserve quota standards are digitally mapped, and the mapping results are dynamically monitored based on low-carbon indicators to obtain green professional warehouse status assessment data; based on the green professional warehouse status assessment data, the physical resource demand of the power supply station is predicted and analyzed, and the predicted results are subjected to multi-objective optimization calculations to obtain hazardous waste resource allocation plan data; based on the hazardous waste resource allocation plan data, the warehouse operation process is decomposed and processed, and the decomposed temporary storage management unit is controlled and calculated to obtain hazardous waste resource operation instruction data; based on the hazardous waste resource operation instruction data, the professional warehouse carbon emission data is calculated and processed, and the calculation results are subjected to dual-carbon target optimization analysis to obtain green professional warehouse environmental management decision data.

[0006] In a second aspect, the present application provides a professional warehouse control system based on artificial intelligence, and the professional warehouse control system based on artificial intelligence includes:

[0007] The acquisition module is used to collect the physical inventory information of the professional warehouse, the physical data of the outbound waiting for use, the physical status of tools and tools, and the location data of the delivery vehicles through multi-source sensing equipment, and perform quality inspection and pre-processing on the collected data. After data fusion processing, a multi-source fusion data set of physical resources is obtained;

[0008] A classification module is used to classify material information by feature extraction based on the multi-source fusion data set of physical resources, and optimize the extracted features by using a transfer learning algorithm to obtain engineering material classification result data;

[0009] A mapping module is used to digitally map the reserve quota standard of the professional warehouse based on the multi-source fusion data set of physical resources and the classification result data of engineering materials, and dynamically monitor the mapping results based on the low-carbonization index to obtain the green professional warehouse status evaluation data;

[0010] A prediction module is used to predict and analyze the physical resource demand of the power supply station based on the green professional warehouse status evaluation data, perform multi-objective optimization calculations on the prediction results, and obtain hazardous waste resource allocation plan data;

[0011] A decomposition module, used to decompose the storage operation process based on the hazardous waste resource allocation plan data, and control and calculate the decomposed temporary storage management unit to obtain hazardous waste resource operation instruction data;

[0012] The calculation module is used to calculate and process the carbon emission data of the professional warehouse according to the hazardous waste material operation instruction data, perform dual-carbon target optimization analysis on the calculation results, and obtain green professional warehouse environmental management decision data.

[0013] The third aspect of the present application provides a computer-readable storage medium, which stores instructions. When the computer-readable storage medium is run on a computer, it enables the computer to execute the above-mentioned professional warehouse control method based on artificial intelligence.

[0014] In the technical solution provided in this application, the physical information of the internal inventory of professional warehouses, the physical data of outbound waiting, the physical status of tools and the location data of delivery vehicles are collected through multi-source sensing equipment, and the collected data are quality inspected and pre-processed, so as to realize the accurate acquisition and effective integration of multi-dimensional data, and provide a reliable data basis for subsequent analysis; the material characteristics are optimized by using the transfer learning algorithm, which significantly improves the accuracy and efficiency of material classification, making the management of different types of materials more accurate; the reserve quota standards of professional warehouses are digitally mapped, and dynamic monitoring is carried out based on low-carbon indicators, so as to realize the refinement and scientificization of reserve management, and provide a quantitative basis for green operation; by predicting and analyzing the physical resource needs of power supply stations and performing multi-objective optimization operations, the accuracy and efficiency of resource allocation are improved, and resource waste is reduced; in terms of hazardous waste resource management, the standardization and safety of hazardous waste resource treatment are realized through the decomposition and processing of warehousing operation processes and the control and calculation of temporary storage management units; finally, through the calculation and processing of carbon emission data of professional warehouses and the optimization analysis of dual carbon targets, a complete green management system is established, which effectively reduces energy consumption and environmental impact. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0016] Figure 1 This is a schematic diagram of an embodiment of a professional warehouse control method based on artificial intelligence in an embodiment of the present application;

[0017] Figure 2 This is a schematic diagram of an embodiment of a professional warehouse control system based on artificial intelligence in an embodiment of the present application. DETAILED DESCRIPTION

[0018] Embodiments of the present application provide a method, system and storage medium for professional warehouse control based on artificial intelligence. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0019] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 In the embodiment of the present application, an embodiment of the professional warehouse control method based on artificial intelligence includes:

[0020] Step S101: collect physical information of the internal inventory of professional warehouses, physical data of physical objects to be used in warehouses, physical status of tools and tools, and location data of delivery vehicles through multi-source sensing equipment, perform quality inspection and pre-processing on the collected data, and obtain a multi-source fusion data set of physical resources through data fusion processing;

[0021] Step S102: Classify the material information by extracting features based on the multi-source fusion data set of physical resources, and optimize the extracted features by using a transfer learning algorithm to obtain engineering material classification result data;

[0022] Step S103: Digitally map the reserve quota standard of the professional warehouse based on the multi-source fusion data set of physical resources and the classification result data of engineering materials, dynamically monitor the mapping results based on the low-carbonization index, and obtain the green professional warehouse status evaluation data;

[0023] Step S104: Based on the green professional warehouse status assessment data, forecast and analyze the physical resource demand of the power supply station, perform multi-objective optimization calculations on the forecast results, and obtain hazardous waste resource allocation plan data;

[0024] Step S105: Based on the hazardous waste resource allocation plan data, the storage operation process is decomposed and processed, and the decomposed temporary storage management unit is controlled and calculated to obtain hazardous waste resource operation instruction data;

[0025] Step S106: Calculate and process the carbon emission data of the professional warehouse based on the hazardous waste resource operation instruction data, perform dual carbon target optimization analysis on the calculation results, and obtain green professional warehouse environmental management decision data.

[0026] It is understandable that the execution subject of the present application can be a professional warehouse control system based on artificial intelligence, or a terminal or a server, which is not limited here. The present application embodiment is described by taking the server as the execution subject as an example.

[0027] Specifically, the physical data in the professional warehouse is collected through the deployed multi-source sensing equipment. The multi-source sensing equipment includes material identification readers, environmental parameter detectors, video collectors and vehicle positioning devices, which are responsible for collecting the physical information of the inventory in the professional warehouse, the physical data of the outbound physical data, the physical status of tools and the location data of the distribution vehicles. Among them, the physical information of inventory refers to the basic information of the materials currently stored in the professional warehouse, including material codes, specifications, models, quantities, and storage time; the physical data of outbound physical data refers to the information of materials that have been processed but not actually out of the warehouse; the physical status of tools refers to the operating status data of various tools and equipment; the location data of the distribution vehicle refers to the real-time location coordinates of the vehicle responsible for material transportation. The collected raw data is firstly tested for quality, mainly including data integrity check, outlier identification and redundant data cleaning. The integrity check is verified by the set data format template to identify and mark the missing fields; outlier identification uses statistical analysis methods to calculate the mean and standard deviation of the data, and marks the data that deviates from the mean by more than 3 standard deviations as outliers; redundant data cleaning deletes duplicate records by comparing the timestamp and content of the data. In the preprocessing stage, the marked problematic data are processed: for missing data, short-term missing data are interpolated with adjacent values, and long-term missing data are replaced with historical averages; for outliers, validity judgments are made based on business rules, and invalid data are removed if confirmed.

[0028] In the data fusion processing stage, the preprocessed multi-source data is integrated. First, a unified data structure template is established, including time dimension, space dimension, object dimension and attribute dimension. Then, data from different sources are mapped according to the dimensional correspondence, and data of the same material at different times, locations and states are associated to form a complete data chain. Finally, through data aggregation operation, the statistical characteristics of each dimension are calculated to generate a multi-source fusion data set of physical resources. In the feature extraction stage, the material information in the multi-source fusion data set of physical resources is deeply analyzed. First, the basic characteristics of the materials are extracted, including explicit characteristics such as specifications, materials, and uses, as well as implicit characteristics such as frequency of use, storage conditions, and hazard level. Then, the features are optimized using the transfer learning algorithm, which can learn the feature expression method from the existing material classification experience and transfer the learned feature knowledge to the new material classification task. Through feature optimization, a more discriminative feature representation is obtained, and then the engineering material classification result data is generated.

[0029] In the digital mapping stage, the multi-source fusion data set of physical resources and the classification result data of engineering materials are compared with the reserve quota standards of professional warehouses. The reserve quota standards refer to the minimum and maximum inventory limits of various types of materials, which need to consider factors such as the importance of the materials, frequency of use, and procurement cycle. During the mapping process, the actual inventory level is compared with the quota standard, the inventory difference is calculated, and the evaluation is combined with low-carbon indicators. Low-carbon indicators include environmental protection indicators such as energy consumption, carbon emissions, and resource utilization. The dynamic monitoring stage continuously tracks inventory changes and trends in environmental indicators. By establishing a monitoring parameter matrix, the historical data of various indicators are recorded, the change rules are analyzed, and future trends are predicted. Based on the monitoring results, green professional warehouse status evaluation data is generated to reflect the operating status and environmental protection level of the professional warehouse.

[0030] In the forecasting and analysis phase, the demand for materials in the future is predicted based on the green professional warehouse status assessment data and the historical demand data of the physical resources of the power supply station. The forecasting process takes into account seasonal factors, the impact of special events, etc., and adopts the time series analysis method to obtain the demand forecast results. Then, multi-objective optimization calculations are performed, and multiple goals such as cost minimization, service level maximization, and environmental impact minimization are considered at the same time to generate hazardous waste material allocation plan data. In the operation process decomposition phase, the hazardous waste material allocation plan data is refined and the overall allocation plan is decomposed into specific operation tasks. First, identify the operation type, including warehousing, out of the warehouse, handling, storage, etc., and then establish a task chain according to the operation sequence. For the temporary storage management unit of hazardous waste materials, special consideration should be given to safety protection requirements, and special storage areas and operating procedures should be set up. Through operation decomposition and calculation, hazardous waste material operation instruction data is generated.

[0031] In the final stage, the carbon emissions during the operation are calculated based on the hazardous waste material operation instruction data and the energy consumption data of the professional warehouse. The calculation takes into account the energy consumption of each link such as equipment operation, lighting and heating, and transportation, and converts it into carbon emission equivalent. Then, the dual carbon target optimization analysis is carried out, namely the carbon peak and carbon neutrality targets, and energy conservation and emission reduction suggestions are put forward to generate green professional warehouse environmental management decision data.

[0032] For example, a professional warehouse is responsible for storing and managing various types of power equipment and materials. Data collected through multi-source sensor equipment shows that the inventory of transformer equipment fluctuates greatly, and a more reasonable reserve quota standard is established. The transfer learning algorithm analyzes historical data and finds that the use of transformers has obvious regional and seasonal characteristics. Combined with these characteristics, the reserve quota is optimized to ensure power supply reliability and avoid inventory backlogs. In terms of hazardous waste treatment, a partitioned storage and batch processing strategy is adopted to reduce the risk of handling waste transformer oil through precise control of operation instructions.

[0033] In the embodiment of the present application, multi-source sensing equipment is used to collect physical information of the internal inventory of professional warehouses, physical data of outbound waiting, physical status of tools and tools, and location data of delivery vehicles, and the collected data is quality inspected and preprocessed, so as to achieve accurate acquisition and effective integration of multi-dimensional data, and provide a reliable data basis for subsequent analysis; the characteristics of materials are optimized by using transfer learning algorithms, which significantly improves the accuracy and efficiency of material classification, making the management of different types of materials more accurate; the reserve quota standards of professional warehouses are digitally mapped, and dynamic monitoring is carried out based on low-carbon indicators, so as to achieve refinement and scientific management of reserves, and provide a quantitative basis for green operations; by predicting and analyzing the physical resource needs of power supply stations and performing multi-objective optimization operations, the accuracy and efficiency of resource allocation are improved, and resource waste is reduced; in terms of hazardous waste resource management, the standardization and safety of hazardous waste resource treatment are achieved through the decomposition and processing of warehousing operation processes and the control and calculation of temporary storage management units; finally, through the calculation and processing of carbon emission data of professional warehouses and the optimization analysis of dual carbon targets, a complete green management system is established, which effectively reduces energy consumption and environmental impact.

[0034] In a specific embodiment, the process of executing step S101 may specifically include the following steps:

[0035] (1) Receive the physical inventory information of the professional warehouse, the physical data of the outbound goods, the physical status of tools and tools, and the location data of the delivery vehicles through the data acquisition interface, and generate four types of original data streams;

[0036] (2) Classify and screen the four types of original data streams, determine the abnormal data judgment boundary through threshold analysis, eliminate the data that exceeds the threshold range, and fill the missing data with adjacent data to obtain preprocessed data;

[0037] (3) The preprocessed data is serially reorganized according to the time dimension, a data correspondence table is established according to the timestamp information, and a data alignment operation is performed based on the correspondence table to obtain time-series aligned data;

[0038] (4) Segment processing is performed on the time series alignment data, and a standardized interval is set for each type of data according to the numerical distribution characteristics. The maximum and minimum value mapping method is used to convert the intervals to obtain standardized data;

[0039] (5) Importing standardized data into the relationship matrix, establishing a three-dimensional association index based on material codes, storage location numbers, and time series numbers, and performing data cross-matching through the index to generate multi-dimensional association data;

[0040] (6) Based on the multidimensional correlation data, time series statistics are calculated, including mean, variance, and periodic characteristics, and layered aggregation is performed in combination with material attribute information to obtain a multi-source fusion data set of physical resources.

[0041] Specifically, the data acquisition interface is a bridge connecting multi-source sensor equipment and data processing units. The physical inventory information of the professional warehouse is collected through RFID readers and barcode scanners, including basic information such as material codes, specifications, quantities, and storage time; the physical data of outbound materials comes from the outbound application records of the business system, which records the information of materials that have been approved but have not yet completed the actual outbound operation; the physical status data of tools is collected by various sensors, including equipment operation status, usage time, maintenance records and other information; the location data of the delivery vehicle is collected by the GPS positioning device in real time, and the latitude and longitude coordinates of the vehicle. These four types of data are transmitted according to a unified data format standard to form four independent original data streams.

[0042] When classifying and filtering the original data stream, a threshold judgment standard is set for each type of data. Taking the physical state of tools as an example, parameters such as equipment operating temperature and vibration frequency have their normal operating ranges, and data beyond this range will be marked as abnormal. The determination of the threshold is based on the statistical analysis results of the equipment technical specifications and historical operating data. For data marked as abnormal, the business rules are used to verify whether it is a real abnormality, and it is eliminated if it is confirmed to be abnormal. In the case of missing data, short-term data missing (such as data within a few minutes) is filled with valid data at nearby time points, and longer-term data missing is filled with the average value of historical data in this period. Serial reorganization of the time dimension is a key step in data preprocessing. Due to the different data collection frequencies of different sensor devices, the timestamps of the original data are not completely aligned. By establishing a data correspondence table, the timestamps of different data sources are mapped to a unified time scale. The correspondence table records the data status of each data source at each standard time point, including information such as actual collection time, data value, and data status mark. Based on this correspondence table, the data is interpolated or downsampled to make the sampling time points of all data sources completely aligned. Standardization is to eliminate the dimensional differences between different types of data. By analyzing the numerical distribution characteristics of each type of data, its range of variation and distribution law are determined, and then an appropriate standardization interval is set. The maximum and minimum value mapping method is a commonly used standardization method, which linearly maps the original data to the [0,1] interval. For data types with many outliers, it is also necessary to combine quantile analysis to determine the upper and lower limits of the mapping interval to avoid the influence of extreme values ​​on the standardization results.

[0043] The establishment of the relationship matrix adopts a three-dimensional data structure, with material code, storage location number and time series number as indexes of three dimensions. The material code uniquely identifies each material, the storage location number reflects the spatial distribution of the material, and the time series number records the time attribute of the data. Through the index of these three dimensions, all relevant data of any material at any time point and any location can be quickly located and associated. In the process of data cross-matching, the association relationship of materials is considered at the same time, such as the inventory change trend of the same type of materials, the substitution relationship between different materials, etc. The calculation of time series statistics is the last link of data fusion. The mean of each indicator of each type of material is calculated to reflect its overall level; the variance is calculated to reflect the degree of fluctuation of the data; and the periodic characteristics of the data are identified through autocorrelation analysis. Material attribute information includes material category, importance, frequency of use, etc. Based on these attributes, stratification is performed, and data aggregation is performed on materials at the same level to obtain statistical characteristics at multiple levels.

[0044] For example, for the management of power transformer equipment, the data acquisition interface simultaneously receives basic information of the transformer (such as specification model, production date), operating status data (such as load rate, oil temperature, insulation condition), inventory change records and maintenance records. When the original data was classified and screened, it was found that there were obvious abnormal values ​​in the oil temperature data. By checking the technical specifications of the equipment, it was determined that the normal oil temperature range was 0-90°C. The data outside this range was marked as abnormal and eliminated. During the data alignment process, data from different sources were unified to an hourly sampling frequency for subsequent analysis. During standardization, different mapping methods were used for different types of data, such as directly using the load rate as the standardized value and mapping the oil temperature data to the [0,1] interval. In the relationship matrix, indexes were established by transformer number, installation location and timestamp to facilitate tracking of the complete status history of each device. Finally, various statistical features were calculated to identify the laws of equipment operation and maintenance.

[0045] In a specific embodiment, the process of executing step S102 may specifically include the following steps:

[0046] (1) Establish a material feature vector matrix based on the multi-source fusion data set of physical resources, calculate statistical feature values ​​based on different material codes, sort the feature values ​​according to their numerical values, and obtain the initial sequence of material features;

[0047] (2) Component normalization calculation is performed on the initial sequence of material characteristics, an association mapping is established through material coding and storage location, and feature combination reconstruction is performed based on the association relationship to obtain material characteristic representation data;

[0048] (3) Group and screen the material characteristic data according to the material attributes, assign weights to each group of characteristics, calculate the importance of the characteristics based on the weight coefficients, and obtain the key characteristic data of the materials;

[0049] (4) Perform dimensionality reduction processing on key feature data of materials, construct feature mapping relationships through transfer learning algorithms, perform spatial transformation on the mapping results, and obtain feature optimized data;

[0050] (5) Construct a material classification decision tree based on feature optimization data, prune the decision tree nodes, select branches according to the information gain criterion, and obtain material classification rules;

[0051] (6) Establish a correspondence between material classification rules and material characteristics, classify material information, generate category identifiers through label correspondence, and obtain engineering material classification result data.

[0052] Specifically, the material feature vector matrix is ​​a mathematical expression of the material characteristics in the multi-source fusion data set of physical resources. For each material, its basic attributes (such as specifications, models, materials), dynamic features (such as inventory change frequency, usage cycle) and associated features (such as matching relationship with other materials) are extracted to form a feature vector. The material code is the unique identifier of each material. The statistical value of each feature is calculated based on the material code, including the average level, degree of variation, change trend, etc. All feature values ​​are sorted, important features are highlighted, and an initial feature sequence is formed. The component normalization calculation standardizes the data of different dimensions in the initial feature sequence. For continuous features, the maximum and minimum value normalization method is used; for discrete features, one-hot encoding is used to convert them into numerical representations. The association mapping between material codes and storage locations reflects the spatial distribution characteristics of materials. By establishing a location coding system, the warehouse space is divided into different areas, and each area is given a unique location identifier. Based on this association relationship, the material features are recombined to generate feature representation data containing spatial information.

[0053] Feature characterization data is grouped according to material attributes, including functional categories (such as equipment, materials, tools), usage scenarios (such as indoors, outdoors), and hazard levels (such as flammable, explosive, corrosive), etc. A weight coefficient is set for each group of features, and the weight configuration takes into account factors such as the importance of the material, frequency of use, and difficulty of substitution. The importance score of each feature is obtained through weight calculation, and the key features with the most discriminative power for material classification are screened out. The transfer learning algorithm is the core of material feature optimization. The algorithm can learn feature expression methods from existing material classification experience. First, the feature extractor is trained in the source field (material categories with sufficient labeled data), and then the learned feature knowledge is transferred to the target field (new material categories) to achieve mapping transformation of the feature space. Dimensionality reduction processing is to reduce the dimension of the feature space and retain the most informative feature combination through methods such as principal component analysis or autoencoders.

[0054] The material classification decision tree is constructed based on the optimized feature data. Each node of the tree represents a feature judgment condition, and the branches represent different judgment results. In the process of building the decision tree, the information gain criterion is used to select the optimal split feature. The greater the information gain, the more helpful the feature is for classification. To prevent overfitting, the decision tree is pruned to delete the branch nodes that do not contribute much to the classification, and a concise and effective classification rule is obtained. When the material classification rules are applied to the actual classification process, it is necessary to establish a mapping relationship between the rules and the specific material characteristics. For each material, the classification rules are applied in turn to determine its category. Finally, a unique identifier is generated for each category through the label encoding system to form standardized engineering material classification result data.

[0055] For example, in the classification of power supply equipment, the feature vector is first constructed, including features such as the rated voltage, rated power, insulation level, and use environment requirements of the equipment. Taking the distribution transformer as an example, the feature vector contains multiple dimensions such as rated capacity (kVA), voltage ratio (kV), connection group, cooling method, and protection level. By analyzing historical data, the statistical values ​​of these features are calculated, such as the average service life and failure rate of transformers of different models. When normalizing the features, the rated capacity is mapped to the [0,1] interval, and the connection group is converted to a numerical representation using one-hot encoding. Combined with the storage location information of the transformer, a location encoding system is established, such as dividing the storage area according to the voltage level and establishing the association between the location and the voltage level. Then, based on the attributes of the transformer (such as indoor / outdoor, oil-immersed / dry type), the transformer is grouped and weights are assigned to different features, such as the weight of the insulation level is higher than that of the shell color. Through transfer learning, the classification experience of large transformers is transferred to the classification of small and medium-sized transformers to optimize the feature expression.

[0056] When building a decision tree, we first classify based on voltage level, and then further subdivide based on features such as capacity and cooling method. After pruning, classification rules such as "If the voltage ratio is 35 / 0.4kV and the capacity is greater than 1000kVA, it is classified as a large distribution transformer" are formed. Finally, these rules are applied to each transformer to generate classification result data containing category code, model specifications, applicable scenarios and other information. These classification results directly guide the storage management and deployment of transformers, and improve the management efficiency of professional warehouses.

[0057] In a specific embodiment, the process of executing step S103 may specifically include the following steps:

[0058] (1) Merge the physical resource multi-source fusion data set and the engineering material classification result data, establish a material value correspondence table, match and calculate the material code and reserve quota, and obtain the material reserve mapping data;

[0059] (2) Divide the material reserve mapping data into numerical intervals according to the professional warehouse reserve quota standard, quantify the interval values, establish a standard corresponding relationship through numerical conversion, and obtain the reserve quota mapping data;

[0060] (3) Construct a low-carbon index system based on the reserve quota mapping data, conduct quantitative analysis on the index values, classify the data according to the energy consumption level and carbon emissions, and obtain low-carbon index analysis data;

[0061] (4) Establish a monitoring parameter matrix based on low-carbon indicator analysis data, calculate the indicator change trend according to the time series, generate monitoring results through data comparison, and obtain dynamic monitoring data;

[0062] (5) Perform correlation analysis on the dynamic monitoring data and the reserve quota mapping data, perform data mining on the correlation results, generate status description information based on the mining results, and obtain status assessment results;

[0063] (6) Conduct green rating on the operation status of the professional warehouse based on the status assessment results, establish status evaluation indicators based on the rating results, form an evaluation report through indicator association, and obtain green professional warehouse status assessment data.

[0064] Specifically, the data merging process first integrates the basic information of materials in the multi-source fusion data set of physical resources with the classification result data of engineering materials. The material value correspondence table records the basic information such as the code, category, quantity, location, etc. of each material, as well as the material feature labels obtained from the classification results. The reserve quota is the minimum and maximum inventory limit set by the professional warehouse for different materials. The matching calculation of material codes and reserve quotas is based on historical demand data and business rules to generate material reserve mapping data containing the comparison relationship between actual inventory levels and quota standards.

[0065] When dividing the material reserve mapping data into numerical intervals, the following quantitative calculation formula is used:

[0066]

[0067] Among them, Q i represents the interval quantified value of the i-th type of materials, α is the weight coefficient of historical data, V k is the inventory in the kth historical period, W k is the period weight, R k is the demand response coefficient, β is the prediction factor, P i To predict demand, n is the number of historical data cycles.

[0068] When constructing a low-carbon index system, a multi-dimensional quantitative analysis method is used to target the energy consumption and carbon emission characteristics of different types of materials:

[0069]

[0070] Among them, E total Represents the comprehensive low-carbon index value, H j is the j-th energy consumption, Y j is the energy conversion coefficient, Z j is the carbon emission factor, γ j is the environmental impact weight, and m is the number of energy types.

[0071] The material reserve mapping data is classified according to the reserve quota standards of professional warehouses, and different inventory level thresholds are set for each category. For conventional materials, maintain a moderate inventory level; for key equipment and emergency materials, ensure sufficient reserves; for consumables and low-value consumables, adopt a dynamic replenishment strategy. The reserve quota mapping data is established through numerical conversion to achieve accurate management of material reserves. The low-carbon index system includes three dimensions: energy consumption index, carbon emission index and resource utilization index. The energy consumption index records the electricity consumption of storage equipment, the energy consumption of lighting and heating, etc.; the carbon emission index calculates the carbon emissions corresponding to various types of energy consumption; the resource utilization index reflects the material turnover efficiency and space utilization. Through quantitative analysis, the baseline value and target value of each indicator are determined, and the classification standard is established to generate low-carbon indicator analysis data. The monitoring parameter matrix is ​​a data structure for dynamically tracking low-carbon indicators, and the changes in various indicators are recorded in time series. By calculating statistical quantities such as moving average, coefficient of variation, and trend coefficient, the change trend and fluctuation law of the indicator are analyzed. Data comparison includes horizontal comparison (comparison with indicators of other similar warehouses) and vertical comparison (comparison with historical data) to generate specific monitoring result data. The correlation analysis between dynamic monitoring data and reserve quota mapping data focuses on the synergistic relationship in two dimensions: one is the impact of inventory levels on energy consumption, and the other is the impact of energy-saving measures on inventory management. Through data mining, key influencing factors are identified, such as the impact of different storage methods on energy consumption, the carbon emission intensity of different operating modes, etc. Based on the mining results, status description information is generated, including the compliance status of various indicators and improvement suggestions.

[0072] Green rating is a comprehensive evaluation of the operation status of professional warehouses. Rating indicators include energy efficiency, carbon emission intensity, resource recycling rate, etc. Improvement goals are set according to the rating results, and a status evaluation indicator system including short-term, medium-term and long-term goals is established. Finally, a complete green professional warehouse status evaluation report is formed through indicator correlation analysis.

[0073] For example: in the intelligent transformation of the power equipment warehouse, the main equipment such as transformers and switch cabinets are classified and coded, and the reserve quota standards are set in combination with historical demand data. Through data analysis, it is found that there are seasonal differences between the reserve level of transformers and actual demand, and the reserve quota range is adjusted accordingly. In terms of low-carbon transformation, it is found through monitoring that the energy consumption of the storage link of large transformers is high, mainly because the constant temperature and humidity requirements cause the air-conditioning system to run continuously. To address this problem, a partition storage strategy is adopted to store equipment with different environmental requirements separately, which significantly reduces energy consumption. Through continuous monitoring and data analysis, a correlation model between equipment storage and energy consumption is established to provide data support for inventory optimization. At the same time, an evaluation system including energy consumption indicators, carbon emission indicators and resource utilization indicators is constructed to achieve the green transformation of warehouse management.

[0074] In a specific embodiment, the process of executing step S104 may specifically include the following steps:

[0075] (1) The green professional warehouse status assessment data is decomposed according to the time dimension, missing values ​​are eliminated through data completion processing, and abnormal points are smoothed and corrected to obtain time series processing data;

[0076] (2) The time series processing data is correlated and matched with the historical records of the physical resource demand of the power supply station, and a time correlation matrix is ​​constructed based on the matching results. The demand feature sequence is obtained through data mapping;

[0077] (3) Perform periodic analysis on the demand feature sequence, extract trend items and seasonal items through numerical decomposition, calculate the demand change pattern based on the decomposition results, and obtain demand forecast data;

[0078] (4) Compare the demand forecast data with the physical resource inventory information of the power supply station, calculate the supply demand based on the difference, generate supply instructions through threshold judgment, and obtain resource allocation data;

[0079] (5) Screen the resource allocation data for hazardous waste resources, classify and sort them based on the hazardous waste classification, establish an allocation sequence table by priority sorting, and obtain the hazardous waste allocation sequence;

[0080] (6) Construct a hazardous waste allocation sequence according to time and space constraints, conduct a multi-objective evaluation of the plan, select the optimal plan based on the evaluation results, and obtain the hazardous waste resource allocation plan data.

[0081] Specifically, the sequence decomposition of the time dimension is the basis for in-depth analysis of the green professional warehouse status assessment data. The green professional warehouse status assessment data contains time series information of multiple dimensions such as energy consumption data, carbon emission data, and material turnover data. The purpose of time series decomposition is to identify long-term trends, periodic changes, and random fluctuations in the data. The data completion process uses interpolation to fill missing data. For short-term missing data, linear interpolation is used, and for long-term missing data, interpolation is based on historical data of the same period. The moving average method is used for smooth correction of abnormal points. The data mean of the adjacent time period is calculated to replace the abnormal value to ensure the continuity and smoothness of the data. The association matching process between the time series processing data and the historical records of the physical resource demand of the power supply station first establishes a time correspondence. The historical records of the physical resource demand of the power supply station include material application records, actual collection records, and usage records. The time association matrix is ​​a two-dimensional data structure, with the horizontal axis representing the time series and the vertical axis representing different types of demand indicators. Through data mapping, data at different time points and different dimensions are integrated into a unified framework to form a complete demand feature sequence.

[0082] The periodic analysis of demand feature sequences adopts the time series decomposition method. The trend term reflects the long-term change direction of demand and is extracted by the moving average method; the seasonal term reflects the law of periodic changes, and the main cycle is identified by Fourier analysis. The decomposition results are used to construct a demand change law model, which comprehensively considers trend factors and seasonal factors to predict the demand in the future period. The demand forecast data is a quantitative estimate of the demand for materials in a certain period in the future. The comparison operation between the demand forecast data and the physical resource inventory information of the power supply station adopts the difference analysis method. The physical resource inventory information of the power supply station includes the number of stocks in stock, the number of in transit, and the number to be inspected. The replenishment demand is determined by calculating the difference between the predicted demand and the current inventory. The generation of replenishment instructions is based on threshold judgment. When the difference exceeds the preset threshold, the replenishment process is triggered to generate resource allocation data including material variety, quantity, and time requirements.

[0083] Hazardous waste resource screening is to identify and classify hazardous waste in resource allocation data. The classification of hazardous waste is based on national standards, including flammability, corrosiveness, toxicity and other characteristics. The classification and sorting of hazardous waste resources takes into account the hazardous characteristics, storage conditions and treatment requirements of the materials, and determines the allocation order through priority sorting. The determination of priority comprehensively considers the degree of hazard, storage time limit and treatment difficulty to form a hazardous waste allocation sequence including the allocation order. The construction of the plan under time and space constraints takes into account the storage space limitations, transportation time requirements and safety spacing requirements. The multi-objective evaluation includes three dimensions: cost minimization, risk minimization and efficiency maximization. The optimal allocation plan is selected through comprehensive evaluation to form a complete hazardous waste resource allocation plan data.

[0084] For example, in the management of waste parts of power transformers, historical data is analyzed. Through time series decomposition, it is found that the generation of waste parts has obvious seasonal characteristics. The amount generated during equipment maintenance in summer is large, while it is relatively small in spring and autumn. At the same time, the missing data in the daily maintenance records of the equipment are corrected through data completion processing, and a complete time series record is established. The processed data is associated with the maintenance plan of the power supply station, and it is found that there is a fixed time difference between the equipment overhaul time and the peak period of waste parts generation. This feature is used to predict the future generation of waste parts. Based on periodic analysis, the annual data is decomposed into long-term trends (increased maintenance caused by equipment aging) and seasonal fluctuations (affected by the maintenance cycle), and the generation of waste parts in the next maintenance cycle is accurately predicted.

[0085] The comparison between the forecast data and the current temporary storage capacity shows that the amount of waste parts in a specific period will exceed the temporary storage capacity, and the replenishment instructions for batch transfer are generated accordingly. In the hazardous waste screening stage, the waste parts are classified according to the hazard level, and the oil-containing parts and insulating materials are stored and processed separately. The final allocation plan takes into account the space limitations of the temporary storage point and meets the time limit requirements for hazardous waste treatment, achieving safe and efficient hazardous waste management.

[0086] In a specific embodiment, the process of executing step S105 may specifically include the following steps:

[0087] (1) Task decomposition of hazardous waste resource allocation plan data is performed, classification and aggregation are performed through job type identification, and priority is sorted according to job requirements to obtain a job decomposition sequence;

[0088] (2) Divide the operation decomposition sequence into links according to the warehouse operation process, arrange the time sequence of the operation content of each link, establish the operation chain through process association, and obtain the process decomposition data;

[0089] (3) Extract the temporary storage management unit from the process decomposition data, perform spatial layout analysis based on location information, perform zoning planning based on hazardous waste levels, and obtain temporary storage area division data;

[0090] (4) Matching the temporary storage area division data with the operation resources, allocating and calculating the operation personnel and equipment, adjusting them through resource constraints, and obtaining resource allocation data;

[0091] (5) Calculate the operation time of resource configuration data, determine the time window through the operation standard, make scheduling arrangements based on time constraints, and obtain operation scheduling data;

[0092] (6) Convert the operation scheduling data into specific operation steps, break down the key operation points, generate control signals through instruction coding, and obtain hazardous waste resource operation instruction data.

[0093] Specifically, task decomposition is the process of converting hazardous waste allocation plan data into specific operation tasks. Operation type identification first divides the allocation tasks into basic types such as collection, classification, temporary storage, and transportation. Each type has its own specific operation requirements and operating specifications. Classification and summary integrate the same type of operation tasks for unified arrangement and management. Priority sorting considers factors such as the hazard level, processing time limit, and storage conditions of hazardous waste resources, determines the processing order according to the urgency and importance, and forms an operation decomposition sequence containing task types, priorities, and execution requirements. The link division of the warehousing operation process is based on the standard operation process specification, and the entire operation process is divided into on-site inspection, classified storage, daily maintenance, and outbound handover. The timing arrangement of the operation content of each link needs to consider the sequential dependency of the operation. For example, hazardous waste materials must complete the on-site inspection before they can be classified and stored. Process association forms a complete operation chain by establishing the connection relationship between the links, ensuring the smooth connection between the links, and finally generating process decomposition data including link division, timing relationship, and operation requirements.

[0094] The temporary storage management unit is the basic unit for temporary storage of hazardous waste materials, including storage areas, storage facilities and management requirements. The spatial layout analysis of location information takes into account the layout of the warehouse, channel settings and safety spacing requirements. When zoning is planned based on the hazardous waste level, materials with different hazardous characteristics are stored separately to ensure storage safety. High-hazardous waste materials require independent dedicated areas and are equipped with corresponding protective facilities. The result of zoning planning forms temporary storage area division data, which includes the location, area, storage capacity and safety requirements of each area. The matching operation of operation resources is to match the temporary storage area division data with the actual available operation resources. The allocation of operators takes into account workload balance and professional requirements to ensure that each operation area has operators with corresponding qualifications. Equipment allocation configures mechanical equipment such as forklifts and transporters according to operation requirements. Resource constraints include personnel quantity restrictions, equipment usage time restrictions and operation space restrictions. These constraints are met by adjusting the operation arrangement to finally form resource configuration data.

[0095] The operation time calculation is based on standard operation time and actual working conditions. The operation standard specifies the standard execution time for each operation step, and the time window takes into account the start time, end time and duration limit of the operation. Time constraints include statutory working hours, hazardous waste treatment time limits and equipment maintenance time. Scheduling arrangements need to reasonably allocate operation time under these constraints to avoid resource conflicts and time conflicts and form operation scheduling data. The detailed decomposition of operation steps is to convert operation scheduling data into specific execution instructions. Each operation task is decomposed into a series of standard operation steps, including operation methods, safety requirements and quality standards. The key points of the operation are refined, and the execution standards and precautions of each step are clearly specified. The instruction coding converts these operation requirements into standardized control signals to form complete hazardous waste material operation instruction data.

[0096] For example, in the task of handling oil leakage in power transformers, the deployment plan is broken down into tasks. The types of operations such as the collection of leaked oil, the preparation of temporary storage containers, and the dispatch of transfer vehicles are identified, and the priority order is determined according to the degree of leakage and environmental risks. Then, the links are divided according to the operation process, from sealing the leakage point to collecting waste oil, from temporary storage to final disposal, forming a complete processing chain.

[0097] In terms of temporary storage management, according to the hazardous characteristics of waste oil, a special anti-leakage storage area is set up, equipped with fire prevention, leakage prevention and ventilation facilities. When allocating operating resources, operators with hazardous chemicals handling qualifications are arranged, equipped with special explosion-proof pumps and sealed containers. In terms of time arrangement, considering the timeliness requirements of waste oil treatment, priority is given to transportation and disposal, and sufficient time for emergency treatment is reserved. The final generated operation instructions include specific operating requirements such as collection methods, protection requirements, and emergency measures to ensure that the entire treatment process is safe and efficient.

[0098] In a specific embodiment, the process of executing step S106 may specifically include the following steps:

[0099] (1) Conduct energy consumption analysis on hazardous waste resource operation instruction data, count electricity consumption through equipment operation records, calculate energy consumption based on operation duration, and obtain raw carbon emission data of professional warehouses;

[0100] (2) The original carbon emission data of professional warehouses are classified and counted according to the operation type, the carbon emission intensity is calculated by the unit operation volume, and the numerical conversion is performed based on the emission coefficient to obtain the carbon emission calculation data;

[0101] (3) Perform time series decomposition on carbon emission calculation data, identify carbon emission patterns through peak and valley distribution, classify data according to emission trends, and obtain carbon emission characteristic data;

[0102] (4) Analyze the differences between carbon emission characteristic data and dual carbon targets, set the deviation interval according to the target value, calculate the adjustment amount according to the interval size, and obtain carbon emission optimization data;

[0103] (5) Segment the carbon emission optimization data by scenario, identify high-carbon links through operation procedures, conduct quantitative evaluation based on energy-saving space, and obtain carbon emission control data;

[0104] (6) Match and analyze the carbon emission control data with environmental management requirements, generate management suggestions through decision-making rules, form control strategies based on the suggestions, and obtain green professional warehouse environmental management decision data.

[0105] Specifically, energy consumption analysis is a comprehensive assessment of energy consumption during hazardous waste material operations. The operating status of various types of electrical equipment is obtained through equipment operation records, including operating hours, power changes, and load conditions. Operation duration refers to the actual working time from the start to the end of the operation, and energy consumption is calculated based on the rated power of the equipment and actual operating conditions. By summarizing the energy consumption data of each operation link, a complete raw data of carbon emissions from professional warehouses is formed.

[0106] When classifying and counting the raw data of carbon emissions from professional warehouses, the data is grouped according to different operation types such as collection, classification, temporary storage, and transportation. The calculation formula for carbon emission intensity per unit of operation is:

[0107]

[0108] Among them, I carbon Indicates carbon emission intensity, F i is the consumption of the i-th energy source, D i is the energy density coefficient, U i is the carbon emission coefficient per unit of energy, Mi is the carbon equivalent conversion factor, Q total is the total workload, and n is the number of energy types.

[0109] The time series decomposition of carbon emission calculation data is processed using the following formula:

[0110]

[0111] Among them, T(t) represents the carbon emission characteristic value at time t, X j is the basic emission in the jth time period, S j is the seasonal factor, L j is the load regulation factor, N j is the time period weight, and m is the number of time periods.

[0112] Peak and valley distribution analysis mainly identifies the peak and trough periods of carbon emissions, and finds out the changing patterns of emissions by segmenting and counting time series data. The classification of emission trends is based on the time scale and the amplitude of change, and similar emission patterns are classified into the same category to form carbon emission characteristic data. The dual carbon target refers to the specific indicator requirements of carbon peak and carbon neutrality. The difference analysis sets a reasonable deviation range by comparing the gap between the actual emission level and the target value. The calculation of the adjustment amount takes into account the technical feasibility and economic efficiency, determines the specific goals of each emission reduction measure, and generates carbon emission optimization data.

[0113] Scenario segmentation is to analyze the carbon emission characteristics under different operation scenarios. Through the analysis of the operation process, high carbon emission links are identified, mainly including the operation of high-energy-consuming equipment, the operation of the temperature control system and transportation operations. The quantitative evaluation of energy-saving space is based on the comparison of the existing process level and the best available technology, and the energy-saving potential of each link is calculated to form carbon emission control data. Environmental management requirements include regulatory standards and internal management objectives of the enterprise. The generation of management recommendations is based on the analysis of carbon emission control data, and specific energy-saving and emission reduction measures are proposed. The control strategy includes technical transformation plan, management optimization plan and emergency disposal plan, which ultimately form the environmental management decision data of green professional warehouses.

[0114] For example, in the process of hazardous waste oil treatment in a professional warehouse of a power supply company, the energy consumption of oil treatment equipment was analyzed. The equipment operation records show that the energy consumption of oil heating and separation equipment in the treatment process accounts for the largest proportion, followed by ventilation and temperature control systems. Oil treatment operations usually need to maintain constant temperature conditions, resulting in a relatively high carbon emission intensity per unit of treatment volume. When conducting time series decomposition analysis, it was found that carbon emissions showed obvious seasonal characteristics, and carbon emissions increased significantly in summer due to increased temperature control load. Through peak and valley distribution analysis, it was identified that 9:00-11:00 and 14:00-16:00 every day are the peak periods of carbon emissions, which corresponds to the concentrated periods of oil treatment operations. Combined with the requirements of the dual carbon goals, phased emission reduction targets were set. For high-carbon links, a series of improvement measures have been taken: improving the insulation process to reduce heating energy consumption, optimizing the operation mode of the ventilation system, and reasonably arranging the operation time to avoid peak electricity consumption. Through the implementation of these measures, the overall carbon emission level of the professional warehouse has been effectively controlled, and a long-term energy-saving and emission reduction mechanism has been established.

[0115] The above describes the professional warehouse control method based on artificial intelligence in the embodiment of the present application. The following describes the professional warehouse control system based on artificial intelligence in the embodiment of the present application. Figure 2 In the embodiment of the present application, an embodiment of a professional warehouse control system based on artificial intelligence includes:

[0116] The acquisition module is used to collect the physical inventory information of the professional warehouse, the physical data of the outbound waiting for use, the physical status of tools and tools, and the location data of the delivery vehicles through multi-source sensing equipment, and perform quality inspection and pre-processing on the collected data. After data fusion processing, a multi-source fusion data set of physical resources is obtained;

[0117] A classification module is used to classify material information by feature extraction based on the multi-source fusion data set of physical resources, and optimize the extracted features by using a transfer learning algorithm to obtain engineering material classification result data;

[0118] A mapping module is used to digitally map the reserve quota standard of the professional warehouse based on the multi-source fusion data set of physical resources and the classification result data of engineering materials, and dynamically monitor the mapping results based on the low-carbonization index to obtain the green professional warehouse status evaluation data;

[0119] A prediction module is used to predict and analyze the physical resource demand of the power supply station based on the green professional warehouse status evaluation data, perform multi-objective optimization calculations on the prediction results, and obtain hazardous waste resource allocation plan data;

[0120] A decomposition module, used to decompose the storage operation process based on the hazardous waste resource allocation plan data, and control and calculate the decomposed temporary storage management unit to obtain hazardous waste resource operation instruction data;

[0121] The calculation module is used to calculate and process the carbon emission data of the professional warehouse according to the hazardous waste material operation instruction data, perform dual-carbon target optimization analysis on the calculation results, and obtain green professional warehouse environmental management decision data.

[0122] Through the coordinated cooperation of the above-mentioned components, the physical information of the internal inventory of the professional warehouse, the physical data of the outbound waiting for use, the physical status of the tools and the location data of the delivery vehicles are collected through multi-source sensing equipment, and the collected data is quality tested and pre-processed, so as to achieve the accurate acquisition and effective integration of multi-dimensional data, and provide a reliable data basis for subsequent analysis; the material characteristics are optimized by using the transfer learning algorithm, which significantly improves the accuracy and efficiency of material classification, making the management of different types of materials more accurate; the reserve quota standards of the professional warehouse are digitally mapped, and dynamic monitoring is carried out based on low-carbon indicators, so as to achieve the refinement and scientific management of reserves, and provide a quantitative basis for green operation; through the prediction analysis and multi-objective optimization calculation of the physical resource demand of the power supply station, the accuracy and efficiency of resource allocation are improved, and the waste of resources is reduced; in terms of hazardous waste resource management, the standardization and safety of hazardous waste resource treatment are achieved through the decomposition and processing of the warehousing operation process and the control calculation of the temporary storage management unit; finally, through the calculation and processing of the carbon emission data of the professional warehouse and the optimization analysis of the dual carbon targets, a complete green management system is established, which effectively reduces energy consumption and environmental impact.

[0123] The present application also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are executed on a computer, the computer executes the steps of the artificial intelligence-based professional warehouse control method.

[0124] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0125] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A professional warehouse control method based on artificial intelligence, characterized in that: The professional warehouse control method based on artificial intelligence includes: Through multi-source sensing equipment, the physical information of the internal inventory of professional warehouses, the physical data of outbound waiting, the physical status of tools and tools, and the location data of delivery vehicles are collected. The collected data is quality tested and pre-processed, and after data fusion processing, a multi-source fusion data set of physical resources is obtained. According to the multi-source fusion data set of physical resources, the material information is classified and processed by feature extraction, and the extracted features are optimized by using a transfer learning algorithm to obtain engineering material classification result data; Based on the multi-source fusion data set of physical resources and the classification result data of engineering materials, the reserve quota standard of professional warehouses is digitally mapped, and the mapping results are dynamically monitored based on low-carbonization indicators to obtain green professional warehouse status assessment data; According to the green professional warehouse status assessment data, the physical resource demand of the power supply station is predicted and analyzed, and the prediction results are subjected to multi-objective optimization calculation to obtain the hazardous waste resource allocation plan data; Based on the hazardous waste resource allocation plan data, the storage operation process is decomposed and processed, and the decomposed temporary storage management unit is controlled and calculated to obtain hazardous waste resource operation instruction data; Based on the hazardous waste resource operation instruction data, the carbon emission data of the professional warehouse is calculated and processed, and the calculation results are subjected to dual-carbon target optimization analysis to obtain green professional warehouse environmental management decision data.

2. The artificial intelligence-based professional warehouse control method according to claim 1 is characterized in that: The multi-source sensing equipment is used to collect the physical information of the internal inventory of the professional warehouse, the physical data of the outbound physical objects, the physical status of the tools and the location data of the delivery vehicles, and the collected data is quality tested and pre-processed. After data fusion processing, a multi-source fusion data set of physical resources is obtained, including: The data acquisition interface receives the physical inventory information of the professional warehouse, the physical data of the outbound goods, the physical status of tools and the location data of the delivery vehicles, and generates four types of original data streams; Classify and screen the four types of original data streams, determine the abnormal data judgment boundary through threshold analysis, eliminate the data that exceeds the threshold range, and fill the missing data with adjacent data to obtain preprocessed data; The pre-processed data is serially reorganized according to the time dimension, a data correspondence table is established according to the timestamp information, and a data alignment operation is performed based on the correspondence table to obtain time-series aligned data; The time series alignment data is segmented, a standardized interval is set for each type of data according to the numerical distribution characteristics, and a maximum and minimum value mapping method is used to convert the interval to obtain standardized data; Importing the standardized data into a relationship matrix, establishing a three-dimensional association index based on the material code, storage location number and time series number, performing data cross-matching through the index, and generating multi-dimensional association data; Time series statistics, including mean, variance, and periodic characteristics, are calculated based on the multidimensional correlation data, and hierarchical aggregation is performed in combination with material attribute information to obtain a multi-source fusion data set of physical resources.

3. The artificial intelligence-based professional warehouse control method according to claim 1 is characterized in that: According to the physical resource multi-source fusion data set, the material information is classified and processed by feature extraction, and the extracted features are optimized by using a transfer learning algorithm to obtain engineering material classification result data, including: Establishing a material feature vector matrix according to the multi-source fusion data set of physical resources, calculating statistical feature values ​​according to different material codes, sorting the feature values ​​according to numerical values, and obtaining an initial sequence of material features; Performing component normalization calculation on the initial sequence of material characteristics, establishing an association mapping through material coding and storage location, and reconstructing feature combination based on the association relationship to obtain material characteristic representation data; The material characteristic characterization data is grouped and screened according to material attributes, weights are assigned to each group of characteristics, and the importance of the characteristics is calculated according to the weight coefficient to obtain key characteristic data of the materials; Performing dimensionality reduction processing on the key feature data of the materials, constructing a feature mapping relationship through a transfer learning algorithm, and performing spatial transformation on the mapping results to obtain feature optimization data; Based on the feature optimization data, a material classification decision tree is constructed, nodes of the decision tree are pruned, branches are selected according to the information gain criterion, and material classification rules are obtained; A correspondence is established between the material classification rules and material characteristics, the material information is classified, and a category identifier is generated through label correspondence to obtain engineering material classification result data.

4. The artificial intelligence-based professional warehouse control method according to claim 1 is characterized in that: According to the physical resource multi-source fusion data set and engineering material classification result data, the professional warehouse reserve quota standard is digitally mapped, and the mapping results are dynamically monitored based on the low-carbonization index to obtain the green professional warehouse status assessment data, including: Merge the physical resource multi-source fusion data set and the engineering material classification result data, establish a material value correspondence table, perform matching calculations based on material codes and reserve quotas, and obtain material reserve mapping data; Divide the material reserve mapping data into numerical intervals according to the professional warehouse reserve quota standard, perform quantitative calculations on the interval values, establish a standard corresponding relationship through numerical conversion, and obtain the reserve quota mapping data; A low-carbonization index system is constructed based on the reserve quota mapping data, the index values ​​are quantitatively analyzed, and data are classified according to energy consumption levels and carbon emissions to obtain low-carbon index analysis data; Based on the low-carbon index analysis data, a monitoring parameter matrix is ​​established, the index change trend is calculated according to the time series, and monitoring results are generated through data comparison to obtain dynamic monitoring data; Performing correlation analysis on the dynamic monitoring data and the reserve quota mapping data, performing data mining on the correlation results, generating state description information according to the mining results, and obtaining state assessment results; A green rating is performed on the operating status of the professional warehouse based on the status assessment results, status evaluation indicators are established based on the rating results, an assessment report is formed through indicator association, and green professional warehouse status assessment data is obtained.

5. The artificial intelligence-based professional warehouse control method according to claim 1 is characterized in that: According to the green professional warehouse status assessment data, the physical resource demand of the power supply station is predicted and analyzed, and the prediction results are subjected to multi-objective optimization calculation to obtain the hazardous waste resource allocation plan data, including: The green professional warehouse status assessment data is sequentially decomposed according to the time dimension, missing values ​​are eliminated through data completion processing, and abnormal points are smoothed and corrected to obtain time series processing data; Correlate and match the time series processing data with the historical records of physical resource demand of the power supply station, construct a time correlation matrix based on the matching results, and obtain a demand feature sequence through data mapping; Performing periodic analysis on the demand feature sequence, extracting trend items and seasonal items through numerical decomposition, calculating the demand change law based on the decomposition results, and obtaining demand forecast data; Compare and calculate the demand forecast data with the physical resource inventory information of the power supply station, calculate the supply demand according to the difference, generate a supply instruction through threshold judgment, and obtain resource allocation data; The resource allocation data is screened for hazardous waste resources, classified and sorted based on hazardous waste grade classification, and an allocation sequence table is established through priority sorting to obtain a hazardous waste allocation sequence; The hazardous waste allocation sequence is constructed according to the time and space constraints, and the plan is evaluated with multiple objectives. The optimal plan is selected based on the evaluation results to obtain the hazardous waste resource allocation plan data.

6. The artificial intelligence-based professional warehouse control method according to claim 1 is characterized in that: Based on the hazardous waste resource allocation plan data, the storage operation process is decomposed and processed, and the decomposed temporary storage management unit is controlled and calculated to obtain hazardous waste resource operation instruction data, including: Decomposing the hazardous waste resource allocation plan data into tasks, classifying and summarizing them through job type identification, and prioritizing them according to job requirements to obtain a job decomposition sequence; The operation decomposition sequence is divided into links according to the warehouse operation process, the operation content of each link is arranged in time sequence, and the operation chain is established through process association to obtain process decomposition data; Extracting the temporary storage management unit in the process decomposition data, performing spatial layout analysis based on location information, performing zoning planning based on hazardous waste levels, and obtaining temporary storage area division data; Matching the temporary storage area division data with the operation resources, allocating and calculating the operation personnel and equipment, adjusting through resource constraints, and obtaining resource allocation data; Calculate the operation time of the resource configuration data, determine the time window through the operation standard, and perform scheduling based on the time constraint to obtain the operation scheduling data; The operation scheduling data is converted into specific operation steps, the key operation points are broken down into details, and control signals are generated through instruction coding to obtain hazardous waste resource operation instruction data.

7. The artificial intelligence-based professional warehouse control method according to claim 1 is characterized in that: The carbon emission data of the professional warehouse is calculated and processed according to the hazardous waste resource operation instruction data, and the calculation results are subjected to dual carbon target optimization analysis to obtain the green professional warehouse environmental management decision data, including: Conduct energy consumption analysis on the hazardous waste resource operation instruction data, count the power consumption through equipment operation records, calculate energy consumption based on operation time, and obtain the original carbon emission data of the professional warehouse; The raw carbon emission data of the professional warehouse is classified and counted according to the operation type, the carbon emission intensity is calculated by the unit operation volume, and the numerical conversion is performed based on the emission coefficient to obtain the carbon emission calculation data; Decomposing the carbon emission calculation data in time series, identifying the carbon emission law through peak and valley distribution, classifying the data according to the emission trend, and obtaining carbon emission characteristic data; Perform difference analysis on the carbon emission characteristic data and the dual carbon targets, set the deviation interval by the target value, calculate the adjustment amount according to the interval size, and obtain the carbon emission optimization data; The carbon emission optimization data is segmented by scenario, high-carbon links are identified through operation procedures, and quantitative evaluation is performed based on energy-saving space to obtain carbon emission control data; The carbon emission control data is matched and analyzed with the environmental management requirements, management suggestions are generated through decision rules, control strategies are formed based on the suggestions, and green professional warehouse environmental management decision data are obtained.

8. An artificial intelligence-based professional warehouse control system, used to implement the artificial intelligence-based professional warehouse control method as described in any one of claims 1 to 7, characterized in that: The professional warehouse control system based on artificial intelligence includes: The acquisition module is used to collect the physical inventory information of the professional warehouse, the physical data of the outbound waiting for use, the physical status of tools and tools, and the location data of the delivery vehicles through multi-source sensing equipment, and perform quality inspection and pre-processing on the collected data. After data fusion processing, a multi-source fusion data set of physical resources is obtained; A classification module is used to classify material information by feature extraction based on the multi-source fusion data set of physical resources, and optimize the extracted features by using a transfer learning algorithm to obtain engineering material classification result data; A mapping module is used to digitally map the reserve quota standard of the professional warehouse based on the multi-source fusion data set of physical resources and the classification result data of engineering materials, and dynamically monitor the mapping results based on the low-carbonization index to obtain the green professional warehouse status evaluation data; A prediction module is used to predict and analyze the physical resource demand of the power supply station based on the green professional warehouse status evaluation data, perform multi-objective optimization calculations on the prediction results, and obtain hazardous waste resource allocation plan data; A decomposition module, used to decompose the storage operation process based on the hazardous waste resource allocation plan data, and control and calculate the decomposed temporary storage management unit to obtain hazardous waste resource operation instruction data; The calculation module is used to calculate and process the carbon emission data of the professional warehouse according to the hazardous waste material operation instruction data, perform dual-carbon target optimization analysis on the calculation results, and obtain green professional warehouse environmental management decision data.

9. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the artificial intelligence-based professional warehouse control method as described in any one of claims 1-7 is implemented.

Images