Intelligent control method, device and storage medium for building a factory based on cloud
By hierarchical data acquisition and heterogeneous data fusion of cloud construction factories, multi-level equivalent digital characterization data is generated, spatial analysis and security assessment is carried out, and the problem of insufficient data integration and information sharing in the existing technology is solved, and the cloud is realized The intelligent control of the construction factory has improved construction efficiency, safety and resource utilization, and promoted the factory to develop towards green and low-carbon.
Patent Information
- Application Number
- CN202411263989.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-10
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2044-09-10
AI Technical Summary
The existing cloud construction factory management and control methods have shortcomings in data integration, information sharing, construction process planning, security management, resource scheduling and environmental monitoring, making it difficult to achieve comprehensive and real-time awareness of factory operation status and systematic collaborative optimization.
By performing layered data collection and heterogeneous data fusion for cloud construction factories, multi-level equivalent digital characterization data are generated, spatial data analysis, security parameter evaluation, resource status analysis and environmental monitoring are carried out, and management decision-making solutions are generated based on multi-dimensional correlation analysis. .
It significantly improves data utilization efficiency and accuracy, realizes the optimal layout and assembly sequence planning of system components, enhances safety risk prevention and emergency response capabilities, improves resource utilization efficiency, optimizes environmental and energy management, and realizes all aspects of the factory. overall coordination and maximization of overall benefits.
Smart Images

Figure CN118798655B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of data analysis, and in particular to an intelligent management and control method, device and storage medium for building a factory based on the cloud. Background Art
[0002] With the rapid development of the construction industry, traditional construction methods can no longer meet the growing demand for construction. As a new construction model, cloud-based construction factories have emerged, which centralize the production, assembly and construction processes of building components in an intelligent and information-based factory environment. This model combines the efficiency of industrialized production with the flexibility of construction, and can significantly improve construction efficiency, quality and safety. At present, cloud-based construction factories have been initially applied in prefabricated and assembled buildings, modular buildings and other fields, showing great development potential.
[0003] However, the existing cloud-based factory construction management and control methods still have some shortcomings. First, insufficient data integration and information sharing between subsystems make it difficult to form a comprehensive and real-time understanding of the factory's operating conditions. Secondly, key links such as construction process planning, safety management, and resource scheduling are often separated from each other, lacking a systematic collaborative optimization mechanism. Furthermore, environmental monitoring and energy management have not yet been fully integrated into the overall management and control system, making it difficult to achieve refined environmental regulation and energy optimization. Finally, decision support systems generally lack comprehensive consideration of multi-dimensional and multi-objective factors, making it difficult to provide managers with comprehensive and reliable decision-making recommendations. Summary of the invention
[0004] The present application provides a method, device and storage medium for intelligent management and control of a cloud-based construction factory, which are used to improve the efficiency and timeliness of intelligent management and control of a cloud-based construction factory.
[0005] In a first aspect, the present application provides an intelligent management and control method for a cloud-based factory construction, the intelligent management and control method for a cloud-based factory construction comprising:
[0006] S1. Perform hierarchical data collection and heterogeneous data fusion on the cloud-based construction factory to generate multi-level equivalent digital representation data;
[0007] S2, extracting and analyzing the spatial data in the multi-level equivalent digital representation data, determining the spatial layout and mutual relationship of system components, calculating the target assembly sequence and target assembly path between system components, and generating a construction process plan according to the target assembly sequence and the target assembly path;
[0008] S3, perform multi-dimensional safety parameter analysis on the factory structure data and safety protection system data of the multi-level equivalent digital representation data to obtain a safety status parameter set, compare the safety status parameter set with a preset safety threshold, and calculate a safety assessment index; generate safety situation assessment data for a factory-like operating environment based on the safety assessment index, specifically including: perform structural stress analysis on the factory structure data in the multi-level equivalent digital representation data to obtain structural stress distribution data, and perform statistical analysis on the structural stress distribution data to obtain structural stress parameters; perform integrity analysis on the safety protection system data in the multi-level equivalent digital representation data The safety protection system status data is obtained by checking, and the safety protection system status data is classified to obtain safety protection parameters; the structural stress parameters and the safety protection parameters are combined to construct a multi-dimensional safety state vector to obtain an initial safety state parameter set, and the initial safety state parameter set is normalized to obtain a standardized safety state parameter set; each parameter in the standardized safety state parameter set is compared with the corresponding preset safety threshold, and the deviation value of each parameter is calculated to obtain a safety parameter deviation set, and the safety parameter deviation set is weighted summed to obtain a comprehensive safety assessment index, wherein the deviation value is calculated by the following formula:
[0009] ;
[0010] in, For the i The deviation value of the parameter, is the corresponding safety threshold, is the normalized parameter value;
[0011] Performing fuzzy processing on the comprehensive safety assessment index to obtain a safety level assessment result, generating a safety risk distribution map based on the safety level assessment result, performing regional identification and quantitative analysis on the safety risk distribution map, and obtaining safety situation assessment data for a factory-like operating environment;
[0012] S4, extracting resource data from the multi-level equivalent digital representation data and performing resource status analysis to generate a resource scheduling strategy;
[0013] S5. Perform time series analysis and pattern recognition on the real-time collected environmental monitoring data and energy consumption data to generate environmental optimization plans and energy use strategies;
[0014] S6. Conduct multi-dimensional correlation analysis on the construction process plan, the safety situation assessment data, the resource scheduling strategy, the environmental optimization plan and the energy use strategy to generate a management decision plan for the cloud-based construction factory.
[0015] In a second aspect, the present application provides an intelligent management and control device for building a factory based on the cloud, and the intelligent management and control device for building a factory based on the cloud includes:
[0016] The fusion module is used to collect hierarchical data and fuse heterogeneous data of cloud-based construction factories to generate multi-level equivalent digital representation data;
[0017] An extraction module is used to extract and analyze the spatial data in the multi-level equivalent digital representation data, determine the spatial layout and mutual relationship of system components, calculate the target assembly sequence and target assembly path between system components, and generate a construction process plan according to the target assembly sequence and the target assembly path;
[0018] A comparison module is used to perform multi-dimensional safety parameter analysis on the plant structure data and safety protection system data of the multi-level equivalent digital representation data to obtain a safety state parameter set, compare the safety state parameter set with a preset safety threshold, and calculate a safety assessment index; generate safety situation assessment data for a factory-like operating environment based on the safety assessment index, specifically including: performing structural stress analysis on the plant structure data in the multi-level equivalent digital representation data to obtain structural stress distribution data, and performing statistical analysis on the structural stress distribution data to obtain structural stress parameters; performing complete analysis on the safety protection system data in the multi-level equivalent digital representation data The method comprises the steps of: performing an integrity check to obtain safety protection system status data, classifying the safety protection system status data, and obtaining safety protection parameters; combining the structural stress parameters and the safety protection parameters to construct a multidimensional safety state vector, obtaining an initial safety state parameter set, and normalizing the initial safety state parameter set to obtain a standardized safety state parameter set; comparing each parameter in the standardized safety state parameter set with a corresponding preset safety threshold, calculating a deviation value of each parameter, obtaining a safety parameter deviation set, and performing weighted summation on the safety parameter deviation set to obtain a comprehensive safety assessment index, wherein the deviation value is calculated by the following formula:
[0019] ;
[0020] in, For the i The deviation value of the parameter, is the corresponding safety threshold, is the normalized parameter value;
[0021] Performing fuzzy processing on the comprehensive safety assessment index to obtain a safety level assessment result, generating a safety risk distribution map based on the safety level assessment result, performing regional identification and quantitative analysis on the safety risk distribution map, and obtaining safety situation assessment data for a factory-like operating environment;
[0022] An analysis module, used to extract resource data from the multi-level equivalent digital representation data and perform resource status analysis to generate a resource scheduling strategy;
[0023] Identification module, used to perform time series analysis and pattern recognition on real-time collected environmental monitoring data and energy consumption data, and generate environmental optimization plans and energy use strategies;
[0024] The decision-making module is used to perform multi-dimensional correlation analysis on the construction process plan, the safety situation assessment data, the resource scheduling strategy, the environmental optimization plan and the energy use strategy to generate a management decision-making plan for the cloud-based construction factory.
[0025] A third aspect of the present application provides a computer-readable storage medium, in which instructions are stored. When the computer-readable storage medium is run on a computer, the computer executes the above-mentioned intelligent management and control method for cloud-based construction factories.
[0026] In the technical solution provided by this application, a multi-level equivalent digital representation is constructed by collecting, integrating and processing multi-source heterogeneous data, which provides a solid data foundation for subsequent intelligent analysis and decision-making, and greatly improves the efficiency and accuracy of data utilization. This comprehensive data integration enables factory managers to grasp the operation status of the factory in real time and comprehensively, so as to make more timely and accurate decisions. Secondly, through the refined analysis and processing of spatial data, the optimized layout of system components and the intelligent planning of assembly sequence are realized, which significantly improves the construction efficiency and space utilization. This not only shortens the project cycle, but also reduces the rework and waste caused by unreasonable spatial layout. In terms of safety management, multi-dimensional safety parameter analysis and real-time safety situation assessment are introduced, which greatly enhances the risk prevention and emergency response capabilities of the factory and effectively reduces the incidence of safety accidents. Through comprehensive consideration of various factors such as structural stress and safety protection system, potential safety hazards can be identified in advance, and targeted preventive measures can be taken, which fundamentally improves the safety level of the factory. In terms of resource management, through intelligent resource scheduling strategies, resource utilization efficiency is significantly improved, and resource waste and idleness are reduced. Through accurate prediction and dynamic adjustment of resource demand, the factory can achieve the optimal allocation of resources, which not only guarantees production demand but also avoids excessive inventory. Environmental and energy management is another important highlight. Through in-depth analysis of environmental monitoring data and energy consumption data, accurate regulation of environmental parameters and optimization of energy use can be achieved, which not only improves the working environment, but also greatly reduces energy consumption, and promotes the factory to develop in a green and low-carbon direction. In particular, by identifying energy consumption hotspots and formulating differentiated energy-saving strategies, the factory can significantly reduce energy costs without affecting production efficiency. Finally, through multi-dimensional correlation analysis and comprehensive decision-making mechanism, the overall consideration and overall optimization of all aspects of the factory are achieved. This global management perspective enables the factory to find the best balance between multiple goals such as production efficiency, safety management, resource utilization, and environmental protection, thereby maximizing overall benefits. Through scenario analysis and sensitivity analysis, the reliability and adaptability of decision-making are also improved. BRIEF DESCRIPTION OF THE DRAWINGS
[0027] 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.
[0028] Figure 1 A schematic diagram of an embodiment of an intelligent management and control method for building a factory based on the cloud in an embodiment of the present application;
[0029] Figure 2This is a schematic diagram of an embodiment of an intelligent management and control device for building a factory based on the cloud in an embodiment of the present application. DETAILED DESCRIPTION
[0030] Embodiments of the present application provide a method, device and storage medium for intelligent management and control of a cloud-based construction plant. 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 that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0031] 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 intelligent management and control method of building a factory based on the cloud includes:
[0032] S1. Perform hierarchical data collection and heterogeneous data fusion on the cloud-based construction factory to generate multi-level equivalent digital representation data;
[0033] S2, extracting and analyzing the spatial data in the multi-level equivalent digital representation data, determining the spatial layout and mutual relationship of system components, calculating the target assembly sequence and target assembly path between system components, and generating a construction process plan according to the target assembly sequence and the target assembly path;
[0034] S3, performing multi-dimensional safety parameter analysis on the factory structure data and safety protection system data of the multi-level equivalent digital representation data to obtain a safety status parameter set, comparing the safety status parameter set with a preset safety threshold, and calculating a safety assessment index; and generating safety situation assessment data for a factory-like operating environment based on the safety assessment index;
[0035] S4, extracting resource data from the multi-level equivalent digital representation data and performing resource status analysis to generate a resource scheduling strategy;
[0036] S5. Perform time series analysis and pattern recognition on the real-time collected environmental monitoring data and energy consumption data to generate environmental optimization plans and energy use strategies;
[0037] S6. Conduct multi-dimensional correlation analysis on the construction process plan, the safety situation assessment data, the resource scheduling strategy, the environmental optimization plan and the energy use strategy to generate a management decision plan for the cloud-based construction factory.
[0038] It is understandable that the execution subject of the present application can be a cloud-based intelligent control device for building a factory, or a terminal or a server, which is not limited here. The embodiment of the present application is described by taking the server as the execution subject as an example.
[0039] Specifically, when performing hierarchical data collection on cloud-based construction factories, various sensors and data collection devices can be deployed, such as structural stress sensors, security monitoring cameras, and environmental parameter detectors, to obtain comprehensive raw data. Then, through steps such as data cleaning, standardization, and feature extraction, these heterogeneous data are fused into multi-level equivalent digital representation data. For example, deep learning algorithms are used to extract spatial features from image data, and dynamic change information is obtained by combining time series data analysis, ultimately forming a unified digital representation. Based on these digital representation data, computational geometry and graph theory algorithms are used to analyze the spatial layout and relationships of system components. For example, the Delaunay triangulation algorithm can quickly construct a spatial relationship network between components, while the minimum spanning tree algorithm helps to determine the key connection path. On this basis, combined with topological sorting and path planning algorithms, the optimal assembly sequence and path can be calculated, and a detailed construction process plan can be generated. For example, when assembling a large steel structure, the algorithm may recommend installing the main beam first, then the secondary beam, and finally the floor slab to ensure structural stability and construction efficiency. In terms of safety management, multi-dimensional analysis of factory structure data and safety protection system data is performed. This may include using finite element analysis to assess the structural stress state, while combining fuzzy reasoning techniques to assess the effectiveness of safety protection measures. By comparing the obtained safety state parameters with preset thresholds, safety assessment indicators can be calculated in real time and visual safety situation assessment data can be generated. For example, if the safety indicator of a certain area is below the threshold, the area is immediately marked and improvement suggestions are made. The generation of resource scheduling strategies involves in-depth analysis of resource data. The system may use time series analysis and machine learning algorithms to predict resource requirements, and then generate the optimal resource allocation plan through optimization algorithms (such as genetic algorithms or particle swarm optimization). For example, when a large number of prefabricated components are predicted to arrive, sufficient lifting equipment and human resources are dispatched to the designated location in advance. The formulation of environmental optimization and energy use strategies requires time series analysis and pattern recognition of environmental monitoring data and energy consumption data. This may involve multi-scale analysis using wavelet transforms and clustering algorithms to identify different energy consumption patterns. Based on these analysis results, intelligent environmental control instructions and energy-saving strategies can be generated. For example, when it is identified that certain areas have abnormally high energy consumption during specific periods of time, it is recommended to adjust the production plan of the area or improve the energy efficiency of equipment. Finally, all the above analysis results are subjected to multi-dimensional correlation analysis, which may involve the use of multi-criteria decision-making methods or deep reinforcement learning techniques. By comprehensively considering multiple dimensions such as construction progress, safety risks, resource utilization efficiency, environmental impact, and energy consumption, the system ultimately generates a comprehensive management decision plan. For example, when faced with multiple projects to be built, the system may recommend the priority execution of a specific project based on the comprehensive analysis results, because the project can maximize resource utilization and minimize environmental impact while ensuring safety.
[0040] In the embodiment of the present application, a multi-level equivalent digital representation is constructed by collecting, integrating and processing multi-source heterogeneous data, which provides a solid data foundation for subsequent intelligent analysis and decision-making, and greatly improves the efficiency and accuracy of data utilization. This comprehensive data integration enables factory managers to grasp the operation status of the factory in real time and comprehensively, so as to make more timely and accurate decisions. Secondly, through the refined analysis and processing of spatial data, the optimized layout of system components and the intelligent planning of assembly sequence are realized, which significantly improves the construction efficiency and space utilization. This not only shortens the project cycle, but also reduces the rework and waste caused by unreasonable spatial layout. In terms of safety management, multi-dimensional safety parameter analysis and real-time safety situation assessment are introduced, which greatly enhances the risk prevention and emergency response capabilities of the factory and effectively reduces the incidence of safety accidents. Through comprehensive consideration of various factors such as structural stress and safety protection system, potential safety hazards can be identified in advance, and targeted preventive measures can be taken, which fundamentally improves the safety level of the factory. In terms of resource management, through intelligent resource scheduling strategies, resource utilization efficiency is significantly improved, and resource waste and idleness are reduced. Through accurate prediction and dynamic adjustment of resource demand, the factory can achieve the optimal allocation of resources, which not only guarantees production demand but also avoids excessive inventory. Environmental and energy management is another important highlight. Through in-depth analysis of environmental monitoring data and energy consumption data, accurate regulation of environmental parameters and optimization of energy use can be achieved, which not only improves the working environment, but also greatly reduces energy consumption, and promotes the factory to develop in a green and low-carbon direction. In particular, by identifying energy consumption hotspots and formulating differentiated energy-saving strategies, the factory can significantly reduce energy costs without affecting production efficiency. Finally, through multi-dimensional correlation analysis and comprehensive decision-making mechanism, the overall consideration and overall optimization of all aspects of the factory are achieved. This global management perspective enables the factory to find the best balance between multiple goals such as production efficiency, safety management, resource utilization, and environmental protection, thereby maximizing overall benefits. Through scenario analysis and sensitivity analysis, the reliability and adaptability of decision-making are also improved.
[0041] In a specific embodiment, the process of executing step S1 may specifically include the following steps:
[0042] (1) Collect the structural data of the power support system, steel platform system, rack system and formwork system of the cloud construction factory to obtain the original data of the structural system, and collect the status data of the safety protection system to obtain the original data of the safety system;
[0043] (2) Collecting environmental parameters and equipment operation data of the intelligent monitoring system of the cloud-based factory construction to obtain raw data of the monitoring system, and performing time synchronization processing on the raw data of the structural system, the raw data of the safety system, and the raw data of the monitoring system to obtain time-aligned multi-source heterogeneous data;
[0044] (3) performing data cleaning on the time-aligned multi-source heterogeneous data to obtain cleaned multi-source data, and performing data standardization processing on the cleaned multi-source data to obtain standardized multi-source data;
[0045] (4) extracting features from the standardized multi-source data to obtain a multi-dimensional feature vector, and performing dimensionality reduction processing on the multi-dimensional feature vector to obtain feature data after dimensionality reduction;
[0046] (5) performing spatial registration on the feature data after dimensionality reduction, establishing a spatial correspondence between multi-source data, obtaining spatially aligned feature data, and performing multi-scale fusion on the spatially aligned feature data to obtain multi-scale fused feature data;
[0047] (6) semantically annotating the multi-scale fusion feature data, constructing a semantic association network, obtaining semantically associated multi-level data, and converting the semantically associated multi-level data into a target data format to generate multi-level equivalent digital representation data.
[0048] Specifically, the structural data of the power support system, steel platform system, pylon system and formwork system are collected. These systems are the core components of the cloud-based construction factory, and their structural data include key information such as geometric dimensions, material properties, and load-bearing capacity. For example, the hydraulic cylinder pressure of the power support system, the stress distribution of the steel platform system, the connection status of the pylon system, and the deformation data of the formwork system will be collected in real time. At the same time, the status data of the safety protection system will also be collected, including information such as the tension of the safety net, the integrity of the guardrail, and the location of the warning signs.
[0049] Next, collect environmental parameters and equipment operation data. Environmental parameters include temperature, humidity, wind speed, dust concentration, etc., while equipment operation data covers the operating status, energy consumption, fault information, etc. of various mechanical equipment. These raw data usually come from different sensors and data sources, and there is a problem of inconsistent timestamps. Therefore, time synchronization processing is required to align all data to a unified time axis. For example, align the pressure data of the power support system (collected every second) with the ambient temperature data (collected every minute) to the same time scale. Data cleaning is the next key step, which aims to remove outliers, process missing data, and eliminate noise. For example, a sensor may produce readings that are significantly deviated from the normal range due to a fault. These outliers need to be identified and eliminated or corrected. The cleaned data also needs to be standardized to unify data of different dimensions and ranges to a comparable scale. Commonly used normalization methods include minimum-maximum normalization and Z-score normalization.
[0050] The standardized data is transformed into a multidimensional feature vector that can better characterize the system state through feature extraction. Feature extraction methods include statistical features (such as mean, variance, kurtosis), frequency domain features (such as Fourier transform coefficients), and time-frequency domain features (such as wavelet transform coefficients). The obtained feature vector may have a high dimension, which is not conducive to subsequent processing, so dimensionality reduction is required. Commonly used dimensionality reduction methods include principal component analysis (PCA) and t-SNE. The feature data after dimensionality reduction also needs to be spatially aligned to establish a spatial correspondence between multi-source data. Ensure that data from different systems can accurately correspond in space. For example, the stress data of the power support system is associated with its specific position in three-dimensional space. The feature data after spatial alignment is then fused at multiple scales, comprehensively considering information at different time and space scales to obtain a more comprehensive data representation.
[0051] Finally, the multi-scale fusion feature data is semantically annotated to construct a semantic association network. The data is associated with its actual physical meaning and engineering background, making it easier to understand and apply. For example, a set of data is annotated as "maximum stress of the main beam" and associated with relevant safety thresholds, design specifications and other information. Finally, these semantically associated multi-level data are converted into a unified target data format to form a multi-level equivalent digital representation of the cloud construction factory. For example, in a large steel structure assembly project, the system may collect the following data: the hydraulic cylinder pressure of the power support system is 150MPa, the maximum stress of the steel platform system is 210MPa, the ambient temperature is 28°C, and the energy consumption of the main lifting equipment is 500kWh / h. These raw data are time-synchronized and unified to one data point every 10 seconds. During the data cleaning process, it was found that a temperature sensor had an abnormal reading (displaying -50°C), and the data was removed. Standardization processing maps all data to the 0-1 interval. The feature extraction step calculates the mean and standard deviation of the pressure data, the maximum value and change rate of the stress data, and the moving average of the temperature data. Dimensionality reduction reduces the original 100-dimensional feature vector to 20 dimensions. Spatial registration matches this data with the 3D model of the factory, accurately locating each component and equipment. Semantic annotation associates this data with high-level concepts such as "structural safety status" and "equipment operating efficiency". Ultimately, these processed data form a digital representation of the current state of the factory.
[0052] In a specific embodiment, the process of executing step S2 may specifically include the following steps:
[0053] (1) extracting features from the spatial data in the multi-level equivalent digital representation data to obtain geometric features and topological relationship data of the components, and performing cluster analysis on the geometric features and topological relationship data to obtain spatial distribution features of the components;
[0054] (2) performing spatial layout planning for the components according to the spatial distribution characteristics to obtain a preliminary layout plan, and performing interference checking on the preliminary layout plan to obtain a component layout plan without interference;
[0055] (3) Based on the non-interference component layout scheme, a connection relationship diagram between components is constructed to obtain a component connection network. , and perform minimum spanning tree analysis on the component connection network to obtain the key connection paths between components, where the component connection network The expression is as follows:
[0056] ;
[0057] (4) According to the critical connection path, the assembly priority of the component is calculated by a topological sorting algorithm to obtain a candidate assembly sequence set, and the assembly feasibility of the candidate assembly sequence set is evaluated to obtain a target assembly sequence;
[0058] (5) Based on the target assembly sequence, the assembly paths between the components are calculated to obtain a candidate assembly path set, and collision detection and path smoothing are performed on the candidate assembly path set to obtain a target assembly path. The path smoothing is calculated using the following formula:
[0059] ;in, is the path point coordinates of the target assembly path, are the path point coordinates of the candidate assembly path, is the smoothing coefficient, 0 ≤ λ ≤ 1, is the time parameter of the waypoint;
[0060] (6) Based on the target assembly sequence and the target assembly path, a preliminary construction flow chart is generated by using a critical path method to obtain candidate construction flow plans, and resource constraint analysis and construction period optimization are performed on the candidate construction flow plans to obtain a construction flow plan.
[0061] Specifically, feature extraction is performed on spatial data to obtain the geometric features and topological relationship data of the components. Geometric features include the size, shape, volume, etc. of the components, while topological relationship data describes the relative position and connection method between components. For example, for a steel structure frame, the geometric features may include the length and cross-sectional dimensions of each beam and column, and the topological relationship includes the coordinates of the connection points and the connection type (such as welding, bolt connection, etc.) between the beams and columns. Cluster analysis is performed on these features and relationship data to obtain the spatial distribution characteristics of the components. Cluster analysis uses the K-means algorithm to classify components with similar features into one category. The objective function of clustering can be expressed as:
[0062] ;
[0063] in, is the objective function, k is the number of clusters, n is the number of samples, For the j The first in the category i samples, For the jThe cluster center of the class is obtained. According to the obtained spatial distribution characteristics, the spatial layout of the components is planned to obtain a preliminary layout plan. The layout is optimized using a genetic algorithm to minimize the interference between components and maximize the space utilization. Subsequently, the preliminary layout plan is checked for interference, and the spatial octree algorithm is used to quickly detect the collision between components to obtain a component layout plan without interference.
[0064] Based on the non-interference layout scheme, a connection relationship diagram between components is constructed to obtain a component connection network. This network can be represented by an adjacency matrix express:
[0065]
[0066] Perform a minimum spanning tree analysis on this connection network and use the Kruskal algorithm to obtain the critical connection path between components. This path ensures that all components are connected and the total connection cost is minimized. Based on the critical connection path, the assembly priority of the components is calculated using a topological sorting algorithm. Topological sorting ensures that the pre-dependencies of each component have been met during the assembly process. After obtaining the candidate assembly sequence set, the assembly feasibility of each sequence is evaluated, considering factors such as stability and operational difficulty, and finally obtaining the target assembly sequence.
[0067] Based on the target assembly sequence, the assembly paths between components are calculated to obtain a set of candidate assembly paths. Collision detection is performed on these paths to ensure that no collision occurs during the assembly process. Then path smoothing is performed using the given formula:
[0068] ;in, is the path point coordinates of the target assembly path, are the path point coordinates of the candidate assembly path, is the smoothing coefficient, 0 ≤ λ ≤ 1, is the time parameter of the waypoint;
[0069] Finally, based on the target assembly sequence and target assembly path, a preliminary construction flow chart is generated through the critical path method. The critical path method uses the following formula to calculate the project completion time:
[0070] ;
[0071] in, The project completion time, CP is the critical path set, For the i The critical path, For the path j The duration of an activity.
[0072] The candidate construction process schemes are subjected to resource constraint analysis and construction period optimization, and finally a construction process scheme is obtained. Specifically, the implementation method is as follows:
[0073] Resource constraint analysis: a) Resource list preparation: List all available resources, including manpower, equipment, materials, etc., and specify the quantity and available time of each resource. b) Task resource requirement quantification: Allocate the specific amount of resources required for each task in the candidate solution. c) Resource load chart generation: Create a time-resource usage chart to show the usage of various resources throughout the project cycle. d) Overload identification: By comparing the resource load chart and the resource list, identify the time period when resource usage exceeds the available amount. e) Resource balancing: Use resource balancing algorithms (such as serial method or parallel method) to adjust the task sequence and time to eliminate resource overload.
[0074] Schedule optimization: a) Critical path identification: Use the critical path method (CPM) to determine the critical task sequence that affects the total schedule. b) Task compression analysis: Perform compressibility analysis on the tasks on the critical path to determine the maximum time and corresponding cost that each task can be compressed. c) Cost-time trade-off analysis: Use linear programming or heuristic algorithms, such as the time-cost trade-off (TCT) method, to find the best compression solution. d) Float time utilization: Rationally allocate the float time of non-critical path tasks to optimize resource utilization. e) Fast follow-up: Identify tasks that can be performed in parallel and adjust the plan to allow these tasks to proceed simultaneously. f) Rushing analysis: Evaluate the impact of adding resources (such as overtime or adding workers) on the schedule and calculate the corresponding cost increase.
[0075] Iterative optimization process: a) Initial solution evaluation: Perform preliminary resource constraint analysis and duration optimization on candidate solutions. b) Solution adjustment: Adjust the task sequence, resource allocation or duration based on the analysis results. c) Re-analysis: Perform resource constraint analysis and duration optimization on the adjusted solution again. d) Multiple iterations: Repeat steps b and c until a solution that meets all constraints and has the best duration is found.
[0076] Sensitivity analysis: a) Impact of parameter changes: Analyze the impact of changes in key parameters (such as the duration of certain tasks or the availability of resources) on the overall solution. b) Scenario simulation: Simulate different project scenarios (such as resource shortages or task delays) to evaluate the robustness of the solution.
[0077] Final plan determination: a) Multi-objective evaluation: Considering multiple objectives such as construction period, cost, resource utilization, etc., use a multi-objective decision-making method such as TOPSIS method to evaluate the optimized plan. b) Plan screening: Select the best construction process plan based on the evaluation results. c) Detail improvement: Make final adjustments and improvements to the selected plan.
[0078] Documentation and visualization: a) Detailed Gantt chart generation: Create a detailed Gantt chart containing all tasks, dependencies and resource allocations. b) Resource usage plan: Generate a detailed usage plan for each resource at each stage of the project. c) Milestone setting: Set key milestones in the optimized process to facilitate project monitoring. d) 3D simulation: Create a 3D visual simulation of the construction process to intuitively display the optimized construction process.
[0079] For example, in a large steel structure plant project, there are 500 steel structure components that need to be assembled. Through feature extraction, the geometric data (such as length, width, height) and connection point coordinates of each component are obtained. Using the K-means clustering algorithm ( k =5), these components were divided into 5 categories: main beams, secondary beams, columns, cross braces and diagonal braces. The spatial layout planning used a genetic algorithm, and after 1000 generations of evolution, a preliminary layout plan was obtained. Interference check found 3 interferences, and after adjustment, a non-interference plan was obtained. A connection relationship diagram was constructed to obtain a 500x500 adjacency matrix. Minimum spanning tree analysis identified 499 key connections. Topological sorting obtained 10 candidate assembly sequences, and the optimal sequence was selected after evaluation: column → main beam → secondary beam → cross brace → diagonal brace. Path planning and smoothing ( λ =0.3) to obtain the final installation path for each component. Critical path analysis shows that the shortest completion time for the project is 45 days, of which the critical path includes 20 main installation steps. After resource constraint analysis and construction period optimization, the final construction plan compressed the project cycle to 42 days while ensuring the balance of daily resource use.
[0080] In a specific embodiment, the process of executing step S3 may specifically include the following steps:
[0081] (1) performing structural stress analysis on the plant structure data in the multi-level equivalent digital representation data to obtain structural stress distribution data, and performing statistical analysis on the structural stress distribution data to obtain structural stress parameters;
[0082] (2) performing an integrity check on the safety protection system data in the multi-level equivalent digital representation data to obtain safety protection system status data, and classifying the safety protection system status data to obtain safety protection parameters;
[0083] (3) combining the structural stress parameter and the safety protection parameter to construct a multi-dimensional safety state vector, obtaining an initial safety state parameter set, and normalizing the initial safety state parameter set to obtain a standardized safety state parameter set;
[0084] (4) Compare each parameter in the standardized safety status parameter set with the corresponding preset safety threshold, calculate the deviation value of each parameter, obtain a safety parameter deviation set, and perform weighted summation on the safety parameter deviation set to obtain a comprehensive safety assessment index, wherein the deviation value is calculated by the following formula:
[0085] ;
[0086] in, For the i The deviation value of the parameter, is the corresponding safety threshold, is the normalized parameter value;
[0087] (5) Fuzzy processing is performed on the comprehensive safety assessment indicators to obtain a safety level assessment result, and a safety risk distribution map is generated based on the safety level assessment result. Regional identification and quantitative analysis are performed on the safety risk distribution map to obtain safety situation assessment data for a factory-like working environment.
[0088] Specifically, the structural stress analysis of the plant structure data in the multi-level equivalent digital representation data is performed, and the structural stress distribution data is calculated using the finite element method. This process involves discretizing the complex plant structure into a finite number of elements and then solving the stress balance equation:
[0089] ;
[0090] in, represents the divergence of the stress tensor, is the volume force. The obtained structural stress distribution data is statistically analyzed to calculate the structural stress parameters such as average stress, maximum stress, and stress standard deviation.
[0091] Next, the integrity check of the safety protection system data is performed, including checking the coverage and functional status of the safety protection equipment, to obtain the safety protection system status data. These data are classified to obtain safety protection parameters that reflect different aspects of safety conditions, such as the protection network integrity index and the alarm system response time. The structural stress parameters and safety protection parameters are combined to construct a multidimensional safety state vector ,in Indicates i safety-related parameters. This vector constitutes the initial safety state parameter set. This parameter set is normalized using the min-max normalization method:
[0092] ;
[0093] in, is the normalized parameter value, and are the minimum and maximum values of the parameter respectively. In this way, a standardized safety state parameter set is obtained. Then, each parameter in the standardized safety state parameter set is compared with the preset safety threshold value to calculate the deviation value:
[0094] ;
[0095] in, For the i The deviation value of the parameter, is the corresponding safety threshold. In this way, we get the safety parameter deviation set. We perform weighted summation on the deviation set to get the comprehensive safety evaluation index :
[0096] ;
[0097] in, For the i The weight of the parameters,
[0098] The comprehensive safety assessment indicators are fuzzified, and the continuous indicator values are mapped to discrete safety levels using fuzzy set theory to obtain the safety level assessment results. Based on this result, a safety risk distribution map is generated to show the safety risk level of each area of the factory. This distribution map is used for regional identification and quantitative analysis, including the identification of high-risk areas and the spatial distribution characteristics of risk levels, and finally the safety situation assessment data of the factory-like working environment is obtained.
[0099] For example, in a large steel structure manufacturing plant, the multi-level equivalent digital characterization data includes stress data of 1,000 key structural nodes and status data of 100 safety protection devices. Through finite element analysis, the von Mises stress value of each node is obtained, ranging from 10MPa to 250MPa. Statistical analysis shows that the average stress is 120MPa, the maximum stress is 250MPa, and the stress standard deviation is 40MPa. After the integrity check of the safety protection system data, 90% of the equipment is in normal working condition. The constructed multidimensional safety state vector contains 10 parameters, including maximum stress, average stress, stress standard deviation, safety equipment normal rate, etc. After normalization, the maximum stress parameter value is 1.0, the average stress is 0.458, the stress standard deviation is 0.333, and the safety equipment normal rate is 0.9. Compared with the preset safety threshold (such as the maximum stress threshold of 200MPa), the deviation value is calculated, and the maximum stress deviation is 0.25, indicating that the threshold is exceeded by 25%. The weighted summation yields a comprehensive safety assessment index of 0.15 (assuming the weight distribution is 0.3 for maximum stress, 0.2 for average stress, and 0.1 for the others). Fuzzy processing maps 0.15 to the "medium risk" level. The safety risk distribution map shows that an area in the northwest corner of the factory has the highest risk level and may require priority attention.
[0100] In a specific embodiment, the process of executing step S4 may specifically include the following steps:
[0101] (1) extracting features from the multi-level equivalent digital representation data to obtain a resource feature vector, and performing dimensionality reduction processing on the resource feature vector to obtain resource data after dimensionality reduction;
[0102] (2) performing cluster analysis on the resource data after dimensionality reduction to obtain resource category classification results, and performing labeling processing on the resource category classification results to obtain a resource category label set;
[0103] (3) Grouping resources according to the resource category tag set to obtain resource grouping data, and performing time series correlation analysis on the resource grouping data to obtain resource utilization time series data;
[0104] (4) performing trend forecasting on the resource utilization time series data to obtain resource demand forecasting results, and calculating resource gaps based on the resource demand forecasting results to obtain resource gap data;
[0105] (5) constructing a resource scheduling optimization objective function based on the resource gap data to obtain a resource scheduling objective function, and setting constraints on the resource scheduling objective function to obtain a resource scheduling constraint condition set;
[0106] Among them, the objective function , For the i Class resources are t The gap in time, is the scheduling cost, is the first weight coefficient and is the second weight coefficient, n is the total number of resource types, T is the total number of time units in the planning period;
[0107] (6) Solving the resource scheduling objective function to obtain a set of candidate resource scheduling solutions, and performing multi-objective evaluation on the set of candidate resource scheduling solutions to obtain a resource scheduling strategy.
[0108] Specifically, in the intelligent management and control method of building factories in the cloud, resource scheduling is a key link. First, feature extraction is performed on the multi-level equivalent digital representation data to obtain the resource feature vector. This includes extracting information such as the type, quantity, and performance parameters of the resources. For example, for a large steel structure manufacturing plant, the resource feature vector may contain multiple dimensions such as the number of welding equipment, crane load capacity, and worker skill level. In order to reduce the complexity of the data, this high-dimensional resource feature vector is subjected to dimensionality reduction processing. The commonly used method is principal component analysis (PCA). The core of PCA is to solve the eigenvalue equation:
[0109] ;
[0110] in, is the covariance matrix, is the eigenvalue, is the identity matrix, is the eigenvector. By selecting the eigenvectors corresponding to the largest eigenvalues, we can get the resource data after dimensionality reduction. Next, we perform cluster analysis on the resource data after dimensionality reduction. The commonly used method is the K-means algorithm. The objective function of K-means is to minimize the within-cluster sum of squares (WCSS):
[0111] ;
[0112] in, k is the number of clusters, It is i clusters, It is i The center of the cluster, xis a data point. Through iterative optimization, the resource classification result is obtained. This result is labeled to obtain a resource category label set, for example, the clustering result is labeled as "high-performance welding equipment", "heavy crane", etc. The resources are grouped according to the resource category label set to obtain resource grouping data. The time series correlation analysis is performed on these grouping data, and the time correlation of resource utilization is calculated using the autocorrelation function (ACF):
[0113] ;
[0114] in, is hysteresis τ The autocorrelation coefficient when yes t Resource utilization at any time, μ is the average utilization rate, is the variance. To predict the trend of resource utilization time series data, the ARIMA (Autoregressive Integrated Moving Average) model can be used. The general form of the ARIMA model is:
[0115] ;
[0116] in, is an autoregressive polynomial, is the moving average polynomial, d is the difference order, is white noise. By fitting the ARIMA model, we can get the resource demand forecast results. We can compare the forecast results with the existing resource quantity, calculate the resource gap, and get the resource gap data. Based on the resource gap data, we can build the resource scheduling optimization objective function:
[0117] ;
[0118] in, It is i The gap of class resources at time t, is the scheduling cost, and are the first weight coefficient and the second weight coefficient respectively. n is the total number of resource types, T It represents the total number of time units in the planning period, and sets constraints, such as total resource limit, scheduling time window, etc., to obtain a set of resource scheduling constraints. Finally, a genetic algorithm is used to solve the resource scheduling objective function to obtain a set of candidate resource scheduling solutions. These solutions are evaluated by multiple objectives, considering multiple indicators such as resource utilization efficiency, scheduling cost, project progress, etc., and finally the optimal resource scheduling strategy is obtained.
[0119] For example, in a large steel structure manufacturing plant, the initial resource feature vector contains 50 dimensions, covering 20 different types of resources. Through PCA dimensionality reduction, 15 principal components that explain 95% of the variance are retained. K-means clustering is performed on these 15-dimensional data ( k =5), and 5 resource categories are obtained, which are marked as "Class A welding equipment", "Class B crane", "Class C cutting machine", "Class D testing equipment" and "Class E human resources". Time series correlation analysis shows that the utilization rate of Class A welding equipment has obvious periodicity, with a period of about 7 days. The ARIMA (2,1,2) model is used to predict the demand for the next 30 days. The prediction results show that there will be a shortage of 10 Class A welding equipment on the 15th day. When constructing the objective function, set = 0.6, = 0.4, reflecting that the equipment gap is given more importance than the scheduling cost. After 1000 generations of evolution, the genetic algorithm obtained 10 candidate scheduling schemes. The multi-objective evaluation took into account resource utilization (weight 0.4), scheduling cost (weight 0.3) and project progress (weight 0.3). The optimal scheme finally selected suggested to deploy 8 Class A welding equipment from other factories on the 13th day, and optimize the workflow to improve the utilization of existing equipment. It is expected that the equipment gap on the 15th day can be completely eliminated, while the overall scheduling cost can be controlled within 95% of the budget.
[0120] In a specific embodiment, the process of executing step S5 may specifically include the following steps:
[0121] (1) performing data cleaning on the environmental monitoring data and energy consumption data to obtain a preprocessed environmental energy consumption data set, and performing time series decomposition on the preprocessed environmental energy consumption data set to obtain trend items, periodic items, and random items;
[0122] (2) performing regression analysis on the trend item to obtain long-term change trend data, performing Fourier transform on the periodic item to obtain periodic characteristic data, and performing statistical analysis on the random item to obtain random fluctuation characteristic data;
[0123] (3) performing time series prediction calculations based on the long-term change trend data, the periodic characteristic data, and the random fluctuation characteristic data to obtain environmental parameter and energy consumption prediction data, and performing threshold analysis on the environmental parameter and energy consumption prediction data to obtain abnormal event warning data;
[0124] (4) performing multi-scale analysis on the pre-processed environmental energy consumption data set through wavelet transform to obtain energy consumption characteristic data of different frequency bands, and performing cluster analysis on the energy consumption characteristic data of different frequency bands to obtain energy consumption pattern classification data;
[0125] (5) constructing an environment-energy consumption association matrix based on the energy consumption pattern classification data, obtaining environmental factor impact assessment data on energy consumption, and generating environmental parameter control instructions based on the environmental factor impact assessment data on energy consumption to obtain an environmental optimization plan;
[0126] (6) Performing a spatiotemporal distribution analysis on the energy consumption pattern classification data to obtain energy consumption hotspot area and time period data, and formulating differentiated energy-saving instructions based on the energy consumption hotspot area and time period data to obtain an energy use strategy.
[0127] Specifically, data cleaning is performed on environmental monitoring data and energy consumption data, a process that includes removing outliers, filling missing values, and eliminating noise. For example, for temperature data, abnormally high or low temperature readings may occur due to sensor failure, and these data need to be identified and eliminated. The cleaned data forms a preprocessed environmental energy consumption data set. Next, the preprocessed environmental energy consumption data set is decomposed into time series to obtain trend items, periodic items, and random items. The trend item reflects the long-term change trend of the data, the periodic item reflects the periodic fluctuation of the data, and the random item represents irregular fluctuations. Regression analysis of the trend item can obtain long-term change trend data, which helps to predict the overall trend in the future. Fourier transform of the periodic item is performed to obtain periodic characteristic data, which can reveal periodic patterns in the data, such as the difference in energy consumption during the day and at night. Statistical analysis of the random item is performed to obtain random fluctuation characteristic data, which helps to understand the uncertainty and degree of fluctuation of the data.
[0128] Based on long-term trend data, periodic characteristic data and random fluctuation characteristic data, time series prediction calculation is performed to obtain environmental parameters and energy consumption prediction data. Time series prediction models such as ARIMA or Prophet models are usually used. The prediction results can provide a reference for future environmental and energy consumption conditions for factory management. Threshold analysis is performed on the predicted data, and a reasonable threshold range is set. Data beyond the range is marked as abnormal event warning data, which helps to timely discover potential environmental or energy consumption problems. Multi-scale analysis is performed on the pre-processed environmental energy consumption data set through wavelet transform to obtain energy consumption characteristic data in different frequency bands. Wavelet transform can analyze signals in both time and frequency domains, which helps to capture the characteristics of energy consumption data at different time scales. Cluster analysis is performed on energy consumption characteristic data in different frequency bands, such as using K-means algorithm to obtain energy consumption pattern classification data. Typical energy consumption patterns can be identified, such as peak period and trough period.
[0129] Based on the energy consumption pattern classification data, an environment-energy consumption association matrix is constructed. This matrix reflects the relationship between environmental factors (such as temperature and humidity) and energy consumption. By analyzing this matrix, the impact assessment data of environmental factors on energy consumption can be obtained. For example, it may be found that for every 1 degree increase in temperature, the air conditioning energy consumption increases by 5%. Based on these impact assessment data, environmental parameter control instructions are generated, such as automatically adjusting the air conditioning settings in hot weather, so as to obtain an environmental optimization plan. Finally, the spatiotemporal distribution analysis of the energy consumption pattern classification data is performed to identify energy consumption hotspot areas and time period data. Areas and time periods with high energy consumption in the factory can be found. Based on these hotspot areas and time period data, differentiated energy-saving instructions are formulated to obtain energy use strategies. For example, for production lines with high energy consumption, it may be recommended to produce during off-peak hours or optimize equipment operating parameters.
[0130] Among them, when constructing the environment-energy consumption association matrix based on energy consumption pattern classification data,
[0131] 1. Build environment-energy consumption correlation matrix:
[0132] a) Data preparation: Organize energy consumption pattern classification data, including different types of energy consumption patterns (such as peak period, valley period, etc.). Collect corresponding environmental factor data, such as temperature, humidity, light, etc.
[0133] b) Matrix structure design: Create a two-dimensional matrix with rows representing different energy consumption patterns and columns representing various environmental factors.
[0134] c) Data filling: For each combination of energy consumption mode and environmental factors, calculate the correlation coefficient. Use Pearson correlation coefficient or Spearman rank correlation coefficient, depending on the data characteristics.
[0135] d) Normalization: Normalize the correlation coefficients in the matrix to ensure that the values are in the range of [-1, 1].
[0136] e) Significance test: Perform a statistical significance test on each correlation coefficient, such as a t-test. Set insignificant correlation coefficients to 0.
[0137] 2. Obtain the impact assessment data of environmental factors on energy consumption:
[0138] a) Regression analysis: For each significant environmental factor-energy consumption pattern pair, regression analysis is performed. Linear regression, polynomial regression or generalized additive model (GAM) can be used, depending on the complexity of the relationship.
[0139] b) Elasticity coefficient calculation: Calculate the elasticity coefficient of each environmental factor on energy consumption. Elasticity coefficient = (percentage change in energy consumption) / (percentage change in environmental factor)
[0140] c) Interaction effect analysis: Use analysis of variance (ANOVA) or structural equation modeling (SEM) to analyze the interaction of multiple environmental factors.
[0141] d) Time lag effect assessment: Use lagged correlation analysis or Granger causality test to assess the lagged impact of environmental factors on energy consumption.
[0142] e) Sensitivity analysis: Monte Carlo simulation is performed to evaluate the sensitivity of energy consumption to changes in environmental factors.
[0143] f) Threshold identification: Use segmented regression or decision tree methods to identify the critical thresholds where environmental factors affect energy consumption.
[0144] g) Impact assessment report generation: Based on the above analysis results, a detailed impact assessment report is generated. The report includes the degree, direction and significance of the impact of each environmental factor on each energy consumption mode.
[0145] It should be noted that when generating environmental parameter control instructions based on impact assessment data, the following are specifically included:
[0146] 1. Impact assessment data analysis:
[0147] a) Rank the impact of each environmental factor and identify the most critical factor.
[0148] b) Determine the optimal value range for each factor, that is, the environmental parameter range with the lowest energy consumption.
[0149] 2. Objective function establishment:
[0150] a) Construct a multi-objective optimization function, including minimizing energy consumption and maximizing production efficiency.
[0151] b) Environmental factors are used as variables and impact assessment data as coefficients.
[0152] 3. Constraint setting:
[0153] a) Set the upper and lower limits of environmental parameters according to equipment specifications and working environment requirements.
[0154] b) Consider the comfort of personnel and add corresponding constraints.
[0155] 4. Optimization algorithm selection:
[0156] a) Depending on the complexity of the objective function, select an appropriate optimization algorithm, such as genetic algorithm or particle swarm optimization.
[0157] b) Use the selected algorithm to solve the objective function and obtain the optimal combination of environmental parameters.
[0158] 5. Control instruction generation:
[0159] a) Convert the optimal parameter combination into specific equipment control instructions.
[0160] b) Design the execution sequence and time interval of control instructions.
[0161] 6. Feedback mechanism design:
[0162] a) Establish a real-time monitoring system to continuously collect environmental data and energy consumption data.
[0163] b) Design feedback algorithms to dynamically adjust control instructions based on actual results.
[0164] 7. Scenario plan formulation:
[0165] a) Develop multiple sets of environmental parameter plans for different production scenarios and seasonal changes.
[0166] b) Design the trigger conditions and switching mechanism for scene switching.
[0167] 8. Formation of environmental optimization plan:
[0168] a) Integrate control instructions, feedback mechanisms and scenario plans into a complete environmental optimization solution.
[0169] b) Develop an implementation plan for the program, including step-by-step implementation and effectiveness evaluation.
[0170] For example, suppose that in a steel structure manufacturing workshop, the following results are obtained through impact assessment data analysis:
[0171] 1. For every 1°C increase in temperature, energy consumption increases by 3%.
[0172] 2. For every 5% increase in humidity, energy consumption increases by 1%.
[0173] 3. For every 100 lux increase in light intensity, energy consumption decreases by 0.5%.
[0174] Based on these data, the following environmental parameter control instructions can be generated:
[0175] 1. Temperature control:
[0176] Set the workshop temperature at 24°C ± 1°C.
[0177] When the outdoor temperature is higher than 28°C, start the air conditioning pre-cooling 2 hours in advance.
[0178] During non-production periods, the temperature is allowed to rise by 2°C to save energy.
[0179] 2. Humidity control:
[0180] Maintain the workshop humidity at 50%±5%.
[0181] When the humidity exceeds 60%, the dehumidification system is activated.
[0182] During dry seasons, use a humidifier to maintain moderate humidity to reduce the effects of static electricity.
[0183] 3. Lighting control:
[0184] Maintain a lighting intensity of 500-600 lux in the working area.
[0185] Install light sensors to automatically adjust artificial lighting based on natural light intensity.
[0186] Use motion-sensing lights in non-work areas to reduce unnecessary lighting.
[0187] 4. Comprehensive regulation:
[0188] According to the production plan, adjust the environmental parameters to the optimal state 1 hour before the production peak.
[0189] Take advantage of the low electricity prices at night to pre-cool or pre-heat the factory.
[0190] Increase local ventilation in equipment-intensive areas to reduce heat accumulation.
[0191] For example, in a large steel structure manufacturing plant, the environmental monitoring system collects data every 5 minutes, including parameters such as temperature, humidity, and dust concentration, and the energy consumption monitoring system records the power consumption of each production line every hour. During the data cleaning process, 2% of the temperature data were found to be abnormal (such as -50°C or 100°C), and these data were removed. Time series decomposition shows that the energy consumption data has obvious periodicity, with each 24 hours as a cycle, and the overall trend is slowly rising. The Fourier transform results show that the main cycles of energy consumption are 24 hours and 168 hours (7 days), reflecting the characteristics of daily and weekly cycles. Time series forecasting calculation predicts energy consumption in the next 7 days, and the results show that the energy consumption peak may occur on the 5th day, exceeding the set threshold by 20%. Wavelet transform analysis found that there are three main energy consumption frequency bands: high frequency band (reflecting instantaneous fluctuations), medium frequency band (reflecting intraday changes) and low frequency band (reflecting weekly changes). Cluster analysis divides energy consumption patterns into 4 categories: trough period (night), stable period (morning and evening), peak period (daytime) and extreme peak period (special working conditions). The environment-energy consumption association matrix shows that when the outdoor temperature exceeds 30°C, the air conditioning energy consumption of the factory increases by 7% for every 1°C increase. Based on this, an environmental parameter control instruction is generated: when the predicted temperature exceeds 30°C, start the factory pre-cooling 2 hours in advance. The spatiotemporal distribution analysis found that workshop A is an energy consumption hotspot from 14:00 to 16:00, and the energy consumption is 30% higher than the average level. In response to this situation, an energy-saving instruction is formulated: adjust the production plan of workshop A and arrange high-energy consumption processes from 9:00 to 11:00, which is expected to reduce peak energy consumption by 15%. Through these measures, the factory expects to achieve the goal of reducing overall energy consumption by 8% while ensuring that production efficiency is not affected.
[0192] In a specific embodiment, the process of executing step S6 may specifically include the following steps:
[0193] (1) extracting features from the construction process plan, safety situation assessment data, resource scheduling strategy, environmental optimization plan, and energy use strategy to obtain a multi-dimensional feature vector, and normalizing the multi-dimensional feature vector to obtain a standardized feature matrix;
[0194] (2) performing dimensionality reduction analysis on the standardized feature matrix to obtain a key feature set after dimensionality reduction, and constructing a multidimensional association network based on the key feature set to obtain a factor association graph;
[0195] (3) performing association analysis on the factor association graph to obtain association factor clusters, and performing cluster analysis on the association factor clusters to obtain a multi-level decision-making factor system;
[0196] (4) constructing a hierarchical analysis structure based on the multi-level decision-making factor system to obtain initial factor weights, and revising the initial factor weights to obtain revised factor weights;
[0197] (5) calculating a comprehensive score based on the modified factor weights and the original data to obtain a multi-dimensional comprehensive evaluation result, and analyzing the multi-dimensional comprehensive evaluation result to obtain a non-inferior solution set;
[0198] (6) For each solution in the non-inferior solution set, a comprehensive score is calculated based on a preset evaluation index, and each solution in the non-inferior solution set is ranked according to the comprehensive score, and the top P solutions with the highest scores are selected as a candidate decision solution set, where P is a positive integer greater than 1, and scenario analysis and sensitivity analysis are performed on the candidate decision solution set to obtain a solution reliability evaluation result;
[0199] (7) Based on the reliability assessment results and decision objectives of the solutions, a comprehensive evaluation is performed on the candidate decision solution set to obtain a management decision solution for the cloud-based factory construction.
[0200] Specifically, feature extraction is performed on the construction process plan, safety situation assessment data, resource scheduling strategy, environmental optimization plan and energy use strategy to obtain a multi-dimensional feature vector. It involves extracting key information from various aspects, such as extracting features such as construction period and critical path from the construction process plan, and extracting features such as risk level and accident probability from the safety situation assessment data. The extracted features may include numerical, categorical and time series data. Then these feature vectors are standardized to unify data of different dimensions to the same scale to obtain a standardized feature matrix. The standardization method usually adopts z-score standardization or min-max standardization. The standardized feature matrix is subjected to dimensionality reduction analysis to obtain the key feature set after dimensionality reduction. Common methods for dimensionality reduction analysis include principal component analysis (PCA) and t-SNE. The purpose is to reduce the dimension of the data while retaining the most important information. Based on the key feature set after dimensionality reduction, a multi-dimensional association network is constructed to obtain a factor association diagram. This association diagram reflects the mutual influence relationship between different factors, such as the relationship between safety measures and construction efficiency.
[0201] Perform correlation analysis on the factor correlation graph to obtain correlation factor clusters. Use graph theory algorithms, such as community discovery algorithms, to identify closely related factor groups. Then perform cluster analysis on these correlation factor clusters to obtain a multi-level decision factor system. Commonly used methods for cluster analysis include K-means and hierarchical clustering. This multi-level system organizes decision factors according to different levels and categories, providing a structured framework for subsequent decision analysis. Based on the multi-level decision factor system, a hierarchical analysis structure is constructed to obtain the initial value of the factor weight, where the hierarchical analysis structure is constructed:
[0202] a) Determine the goal layer: clarify the ultimate goal of the decision, such as "select the best cloud-based construction factory management solution."
[0203] b) Determine the criterion layer: Based on the multi-level decision-making factor system, the top-level factors are used as the criterion layer. For example: production efficiency, safety, cost-effectiveness, environmental impact, etc.
[0204] c) Determine the sub-criteria layer: The secondary factors under the criterion layer are used as the sub-criteria layer. For example, under the "production efficiency" criterion, there may be sub-criteria such as "equipment utilization rate" and "personnel efficiency".
[0205] d) Determine the solution layer: List all specific solutions to be evaluated.
[0206] e) Establish a hierarchical structure diagram: connect the target layer, criterion layer, sub-criterion layer and solution layer in a tree structure. Get the initial value of the factor weight:
[0207] a) Construct a judgment matrix: Compare the factors at each level in pairs. Use a 1-9 scale, with 1 indicating equal importance and 9 indicating extreme importance. For example, if you compare "production efficiency" and "safety", if you think the former is slightly more important, you can assign a value of 3.
[0208] b) Calculate eigenvectors: Calculate the maximum eigenvalue and the corresponding eigenvector for each judgment matrix. Use the power method or other numerical methods to solve: ,in is the judgment matrix, is the characteristic value, is the feature vector.
[0209] c) Consistency test: Calculate consistency index ,in, is the maximum characteristic value of the judgment matrix, is the order of the judgment matrix.
[0210] Calculating the consistency ratio ,in is a random consistency indicator
[0211] if < 0.1, the judgment matrix has satisfactory consistency
[0212] d) Calculate the comprehensive weight: perform weighted integration of the local weights of each level to obtain the global weight.
[0213] Use the formula:
[0214] in Factor The global weight of For the upper factor j The weight of Factor Factors The local weight of .
[0215] e) Weight normalization: Normalize the obtained weights to ensure that the sum of all weights is 1.
[0216] The analytic hierarchy process (AHP) is a commonly used method that determines the relative importance of each factor through expert scoring and pairwise comparison. The initial value of the factor weight is revised. This revision process may involve the Delphi method, that is, adjusting the weight through multiple rounds of expert consultation to finally obtain the revised factor weight. The comprehensive score is calculated based on the revised factor weight and the original data to obtain a multi-dimensional comprehensive evaluation result, which specifically includes the following steps:
[0217] 1. Data standardization:
[0218] a) Standardize the raw data to make data of different dimensions comparable.
[0219] b) For benefit indicators (the bigger the better), use the formula:
[0220]
[0221] c) For cost indicators (the smaller the better), use the formula:
[0222]
[0223] in, For the Plan No. j The original value of the indicator, and are the maximum and minimum values of the indicator respectively.
[0224] 2. Construct a weighted standardized decision matrix:
[0225] a) Multiply the standardized data with the corrected factor weights.
[0226] b) Get the matrix ,in
[0227] in, For the j The modified weights of the factors.
[0228] 3. Determine the ideal solution and negative ideal solution:
[0229] a) Ideal solution ,in
[0230] b) Negative ideal solution ,in
[0231] 4. Calculate the distance to the ideal solution and the negative ideal solution:
[0232] a) Distance to the ideal solution:
[0233] b) Distance to the negative ideal solution:
[0234] 5. Calculate relative proximity:
[0235]
[0236] 6. Calculate the comprehensive score:
[0237] a) For each solution, multiply its standardized score on each indicator by the corresponding weight and then sum them up.
[0238] b) Overall rating
[0239] 7. Result sorting:
[0240] Based on comprehensive rating or relative proximity , sort the solutions to obtain a multi-dimensional comprehensive evaluation result. It should be noted that in the embodiment of the present invention, the multi-dimensional comprehensive evaluation result includes: Production efficiency dimension: equipment utilization, production cycle time, product qualification rate, personnel efficiency. Safety dimension: accident rate, safety hazard identification rate, emergency response time, safety training coverage. Cost-effectiveness dimension: production cost, energy consumption cost, maintenance cost, return on investment. Quality control dimension: product first-time qualification rate, customer satisfaction, rework rate, quality traceability capability, environmental impact dimension:
[0241] Energy efficiency, waste treatment rate, carbon emissions, noise control level. Resource scheduling dimension: human resource utilization, equipment scheduling efficiency, raw material turnover rate, inventory management level. Supply chain collaboration dimension: supplier management index, raw material supply on-time rate, inventory turnover rate, order fulfillment rate. This evaluation result reflects the comprehensive performance of each possible solution in multiple dimensions. These comprehensive evaluation results are analyzed and the Pareto optimization principle is used to obtain a non-inferior solution set. The solutions in the non-inferior solution set are superior to other solutions in at least one dimension, and there is no situation where they are completely superior to other solutions.
[0242] The non-inferior solution set is optimized to obtain a set of candidate decision solutions. This optimization process may involve multi-criteria decision-making methods, such as TOPSIS (top-up to ideal solution ranking method). Specifically, assume that there is a non-inferior solution set containing 5 solutions (A, B, C, D, E). The preset evaluation indicators include economic benefits (weight 0.4), technical feasibility (weight 0.3) and environmental impact (weight 0.3). The scoring range is 0-100 points. Score each solution: Solution A: Economic benefits 85 points, technical feasibility 90 points, environmental impact 75 points. Solution B: Economic benefits 90 points, technical feasibility 80 points, environmental impact 85 points. Solution C: Economic benefits 80 points, technical feasibility 85 points, environmental impact 90 points. Solution D: Economic benefits 95 points, technical feasibility 75 points, environmental impact 70 points Solution E: Economic benefits 75 points, technical feasibility 95 points, environmental impact 80 points.
[0243] Calculate the comprehensive score: Score of Plan A: 85×0.4 + 90×0.3 + 75×0.3 = 83.5 points. Score of Plan B: 90×0.4 + 80×0.3 + 85×0.3 = 85.5 points. Score of Plan C: 80×0.4 + 85×0.3 + 90×0.3 = 84.5 points. Score of Plan D: 95×0.4 + 75×0.3 + 70×0.3 = 81.5 points. Score of Plan E: 75×0.4 +95×0.3 + 80×0.3 = 82.5 points. Sorting: B(85.5) > C(84.5) > A(83.5) > E(82.5) > D(81.5)
[0244] Select candidate solutions: Assuming P=3, select the top three solutions with the highest scores as the candidate decision solution set, namely {B, C, A}.
[0245] Then, scenario analysis and sensitivity analysis are performed on the candidate decision solution set to evaluate the performance of the solution under different circumstances and the sensitivity to changes in key parameters, thereby obtaining the solution reliability evaluation results. Finally, based on the solution reliability evaluation results and decision goals, a comprehensive evaluation is performed on the candidate decision solution set to obtain the management decision solution for the cloud-based factory construction. Specifically, the evaluation factor set is determined: , which includes the reliability of the solution and various decision-making objectives. For example: U = {reliability, production efficiency, cost control, safety, environmental impact}, determine the comment set: , indicating the evaluation level. For example: = {very bad, bad, average, good, very good}, establish the fuzzy relationship matrix :For each candidate solution, experts give membership scores based on the solution reliability assessment results and decision-making goals to form a fuzzy relationship matrix .
[0246] , in Indicates the membership of the i-th factor to the j-th comment. Determine the weight vector: ,in represents the weight of the i-th factor, and . The weights can be determined using the analytic hierarchy process (AHP). Fuzzy comprehensive evaluation: ,in Indicates fuzzy synthesis operation. Specific calculation: , , according to the maximum membership principle or weighted average method, B is converted into a specific evaluation result. Sensitivity analysis: Adjust the weight vector A, observe the change in the result, and evaluate the stability of the solution. This comprehensive evaluation process uses the fuzzy comprehensive evaluation method to combine qualitative and quantitative factors for a comprehensive evaluation.
[0247] For example, in a large steel structure manufacturing plant, the initial feature extraction obtained 100 features, including average daily output, safety accident rate, energy efficiency, etc. After standardization, the number of features was reduced to 20 key features through PCA dimensionality reduction. The constructed factor association diagram showed that there was a strong negative correlation between safety measures and production efficiency (correlation coefficient -0.6), while the degree of automation was positively correlated with energy efficiency (correlation coefficient 0.7). Cluster analysis divided these factors into 5 main categories: production efficiency, safety management, resource scheduling, environmental control and energy management.
[0248] The initial weights obtained by the hierarchical analysis showed that the weight of production efficiency was 0.3, safety management was 0.25, resource scheduling was 0.2, environmental control was 0.15, and energy management was 0.1. After expert revision, the weight of safety management increased to 0.3, while production efficiency dropped to 0.25, reflecting a higher emphasis on safety. The comprehensive score calculation resulted in 10 potential solutions, of which 3 solutions constituted a non-inferior solution set. These three solutions have their own advantages in different dimensions, such as solution A has the highest production efficiency, solution B has the best safety, and solution C has the leading energy efficiency. The scenario analysis considered three situations: normal, high demand, and low demand, and the sensitivity analysis focused on the impact of raw material price fluctuations on the performance of the solutions. The final comprehensive evaluation results showed that solution B scored the highest after comprehensive consideration of various factors, because it performed stably in different scenarios and was less sensitive to raw material price fluctuations. Therefore, solution B was selected as the final management decision solution, which recommended increasing safety investment by 20%, optimizing production processes to improve efficiency by 15%, and reducing energy consumption by 10% through intelligent scheduling.
[0249] The above describes the intelligent control method for building a factory based on the cloud in the embodiment of the present application. The following describes the intelligent control device for building a factory based on the cloud in the embodiment of the present application. Figure 2 In the embodiment of the present application, an embodiment of the intelligent control device for building a factory based on the cloud includes:
[0250] The fusion module 201 is used to collect hierarchical data and fuse heterogeneous data of the cloud construction factory to generate multi-level equivalent digital representation data;
[0251] An extraction module 202 is used to extract and analyze the spatial data in the multi-level equivalent digital representation data, determine the spatial layout and mutual relationship of system components, calculate the target assembly sequence and target assembly path between system components, and generate a construction process plan according to the target assembly sequence and the target assembly path;
[0252] Comparison module 203, used to perform multi-dimensional safety parameter analysis on the plant structure data and safety protection system data of the multi-level equivalent digital representation data to obtain a safety status parameter set, compare the safety status parameter set with a preset safety threshold, and calculate a safety assessment index; and generate safety situation assessment data for a factory-like operating environment based on the safety assessment index;
[0253] An analysis module 204 is used to extract resource data from the multi-level equivalent digital representation data and perform resource status analysis to generate a resource scheduling strategy;
[0254] Identification module 205, used to perform time series analysis and pattern recognition on the real-time collected environmental monitoring data and energy consumption data, and generate environmental optimization solutions and energy use strategies;
[0255] The decision module 206 is used to perform multi-dimensional correlation analysis on the construction process plan, the safety situation assessment data, the resource scheduling strategy, the environmental optimization plan and the energy use strategy to generate a management decision plan for the cloud-based construction factory.
[0256] Through the collaborative cooperation of the above components, through the collection, fusion and processing of multi-source heterogeneous data, a multi-level equivalent digital representation is constructed, which provides a solid data foundation for subsequent intelligent analysis and decision-making, and greatly improves the efficiency and accuracy of data utilization. This comprehensive data integration enables factory managers to grasp the operation status of the factory in real time and comprehensively, so as to make more timely and accurate decisions. Secondly, through the refined analysis and processing of spatial data, the optimized layout of system components and the intelligent planning of assembly sequence are realized, which significantly improves the construction efficiency and space utilization. This not only shortens the project cycle, but also reduces the rework and waste caused by unreasonable spatial layout. In terms of safety management, multi-dimensional safety parameter analysis and real-time safety situation assessment are introduced, which greatly enhances the risk prevention and emergency response capabilities of the factory and effectively reduces the incidence of safety accidents. Through comprehensive consideration of various factors such as structural stress and safety protection system, potential safety hazards can be identified in advance, and targeted preventive measures can be taken, which fundamentally improves the safety level of the factory. In terms of resource management, through intelligent resource scheduling strategies, resource utilization efficiency is significantly improved, and resource waste and idleness are reduced. Through accurate prediction and dynamic adjustment of resource demand, the factory can achieve the optimal allocation of resources, which not only guarantees production demand but also avoids excessive inventory. Environmental and energy management is another important highlight. Through in-depth analysis of environmental monitoring data and energy consumption data, accurate regulation of environmental parameters and optimization of energy use can be achieved, which not only improves the working environment, but also greatly reduces energy consumption, and promotes the factory to develop in a green and low-carbon direction. In particular, by identifying energy consumption hotspots and formulating differentiated energy-saving strategies, the factory can significantly reduce energy costs without affecting production efficiency. Finally, through multi-dimensional correlation analysis and comprehensive decision-making mechanism, the overall consideration and overall optimization of all aspects of the factory are achieved. This global management perspective enables the factory to find the best balance between multiple goals such as production efficiency, safety management, resource utilization, and environmental protection, thereby maximizing overall benefits. Through scenario analysis and sensitivity analysis, the reliability and adaptability of decision-making are also improved.
[0257] The present application also provides an intelligent management and control device for a cloud-based construction factory, which includes a memory and a processor. The memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the processor executes the steps of the intelligent management and control method for the cloud-based construction factory in the above-mentioned embodiments.
[0258] 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 intelligent management and control method for building a factory based on the cloud.
[0259] 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.
[0260] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program codes.
[0261] 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. An intelligent management and control method for building a factory based on the cloud, characterized in that: The intelligent management and control method for building a factory based on the cloud includes: S1. Perform hierarchical data collection and heterogeneous data fusion on the cloud-based construction factory to generate multi-level equivalent digital representation data, including: collecting the structural data of the power support system, steel platform system, rack system and formwork system of the cloud-based construction factory to obtain the original data of the structural system, and collecting the status data of the safety protection system to obtain the original data of the safety system; collecting the environmental parameters and equipment operation data of the intelligent monitoring system of the cloud-based construction factory to obtain the original data of the monitoring system, and performing time synchronization processing on the original data of the structural system, the original data of the safety system and the original data of the monitoring system to obtain time-aligned multi-source heterogeneous data; performing data cleaning on the time-aligned multi-source heterogeneous data to obtain the cleaned multi-source heterogeneous data. data, and perform data standardization on the cleaned multi-source data to obtain standardized multi-source data; perform feature extraction on the standardized multi-source data to obtain a multi-dimensional feature vector, and perform dimensionality reduction on the multi-dimensional feature vector to obtain reduced-dimensional feature data; perform spatial registration on the reduced-dimensional feature data to establish a spatial correspondence between multi-source data to obtain spatially aligned feature data, and perform multi-scale fusion on the spatially aligned feature data to obtain multi-scale fused feature data; perform semantic annotation on the multi-scale fused feature data to construct a semantic association network to obtain semantically associated multi-level data, and convert the semantically associated multi-level data into a target data format to generate multi-level equivalent digital representation data; S2, extracting and analyzing the spatial data in the multi-level equivalent digital representation data, determining the spatial layout and mutual relationship of system components, calculating the target assembly sequence and target assembly path between system components, and generating a construction process plan according to the target assembly sequence and the target assembly path; S3, perform multi-dimensional safety parameter analysis on the factory structure data and safety protection system data of the multi-level equivalent digital representation data to obtain a safety status parameter set, compare the safety status parameter set with a preset safety threshold, and calculate a safety assessment index; generate safety situation assessment data for a factory-like operating environment based on the safety assessment index, specifically including: perform structural stress analysis on the factory structure data in the multi-level equivalent digital representation data to obtain structural stress distribution data, and perform statistical analysis on the structural stress distribution data to obtain structural stress parameters; perform integrity analysis on the safety protection system data in the multi-level equivalent digital representation data The safety protection system status data is obtained by checking, and the safety protection system status data is classified to obtain safety protection parameters; the structural stress parameters and the safety protection parameters are combined to construct a multi-dimensional safety state vector to obtain an initial safety state parameter set, and the initial safety state parameter set is normalized to obtain a standardized safety state parameter set; each parameter in the standardized safety state parameter set is compared with the corresponding preset safety threshold, and the deviation value of each parameter is calculated to obtain a safety parameter deviation set, and the safety parameter deviation set is weighted summed to obtain a comprehensive safety assessment index, wherein the deviation value is calculated by the following formula: ; in, For the i The deviation value of the parameter, is the corresponding safety threshold, is the normalized parameter value; Performing fuzzy processing on the comprehensive safety assessment index to obtain a safety level assessment result, generating a safety risk distribution map based on the safety level assessment result, performing regional identification and quantitative analysis on the safety risk distribution map, and obtaining safety situation assessment data for a factory-like operating environment; S4. Extract resource data from the multi-level equivalent digital representation data and perform resource status analysis to generate a resource scheduling strategy, specifically including: extracting features from the multi-level equivalent digital representation data to obtain resource feature vectors, and performing dimensionality reduction processing on the resource feature vectors to obtain resource data after dimensionality reduction; performing cluster analysis on the resource data after dimensionality reduction to obtain resource category division results, and labeling the resource category division results to obtain resource category label sets; grouping resources according to the resource category label sets to obtain resource grouping data, and performing time series correlation analysis on the resource grouping data to obtain resource utilization time series data; performing trend forecasting on the resource utilization time series data to obtain resource demand forecasting results, and calculating resource gaps according to the resource demand forecasting results to obtain resource gap data; constructing a resource scheduling optimization objective function based on the resource gap data to obtain a resource scheduling objective function, and setting constraints on the resource scheduling objective function to obtain a resource scheduling constraint set; wherein, the objective function , For the i Class resources are t The gap in time, is the scheduling cost, is the first weight coefficient and is the second weight coefficient, n is the total number of resource types, T is the total number of time units in the planning period; solving the resource scheduling objective function to obtain a set of candidate resource scheduling solutions, and performing multi-objective evaluation on the set of candidate resource scheduling solutions to obtain a resource scheduling strategy; S5. Perform time series analysis and pattern recognition on the real-time collected environmental monitoring data and energy consumption data to generate environmental optimization plans and energy use strategies; S6. Conduct multi-dimensional correlation analysis on the construction process plan, the safety situation assessment data, the resource scheduling strategy, the environmental optimization plan and the energy use strategy to generate a management decision plan for the cloud-based construction factory.
2. The intelligent management and control method for building a factory based on the cloud according to claim 1 is characterized in that: Step S2 comprises: Extracting features from the spatial data in the multi-level equivalent digital representation data to obtain geometric features and topological relationship data of the components, and performing cluster analysis on the geometric features and topological relationship data to obtain spatial distribution features of the components; Performing spatial layout planning for components according to the spatial distribution characteristics to obtain a preliminary layout plan, and performing interference checking on the preliminary layout plan to obtain a component layout plan without interference; Based on the non-interference component layout scheme, a connection relationship diagram between components is constructed to obtain a component connection network , and perform minimum spanning tree analysis on the component connection network to obtain the key connection paths between components, where the component connection network The expression is as follows: ; According to the critical connection path, the assembly priority of the components is calculated by a topological sorting algorithm to obtain a candidate assembly sequence set, and the assembly feasibility of the candidate assembly sequence set is evaluated to obtain a target assembly sequence; Based on the target assembly sequence, the assembly paths between the components are calculated to obtain a candidate assembly path set, and collision detection and path smoothing are performed on the candidate assembly path set to obtain the target assembly path. The path smoothing is calculated by the following formula: ;in, is the path point coordinates of the target assembly path, are the path point coordinates of the candidate assembly path, is the smoothing coefficient, 0 ≤ λ ≤ 1, is the time parameter of the waypoint; According to the target assembly sequence and the target assembly path, a preliminary construction flow chart is generated by the critical path method to obtain candidate construction flow plans, and resource constraint analysis and construction period optimization are performed on the candidate construction flow plans to obtain a construction flow plan.
3. The intelligent management and control method for building a factory based on the cloud according to claim 1 is characterized in that: Step S5 includes: Performing data cleaning on the environmental monitoring data and energy consumption data to obtain a preprocessed environmental energy consumption data set, and performing time series decomposition on the preprocessed environmental energy consumption data set to obtain trend items, periodic items, and random items; Performing regression analysis on the trend item to obtain long-term change trend data, performing Fourier transform on the periodic item to obtain periodic characteristic data, and performing statistical analysis on the random item to obtain random fluctuation characteristic data; Perform time series prediction calculations based on the long-term change trend data, the periodic characteristic data, and the random fluctuation characteristic data to obtain environmental parameters and energy consumption prediction data, and perform threshold analysis on the environmental parameters and energy consumption prediction data to obtain abnormal event warning data; Performing multi-scale analysis on the pre-processed environmental energy consumption data set through wavelet transform to obtain energy consumption characteristic data of different frequency bands, and performing cluster analysis on the energy consumption characteristic data of different frequency bands to obtain energy consumption mode classification data; Based on the energy consumption mode classification data, an environment-energy consumption association matrix is constructed to obtain evaluation data on the impact of environmental factors on energy consumption, and environmental parameter control instructions are generated according to the evaluation data on the impact of environmental factors on energy consumption to obtain an environmental optimization plan; The energy consumption pattern classification data is subjected to a spatiotemporal distribution analysis to obtain energy consumption hotspot area and time period data, and differentiated energy-saving instructions are formulated based on the energy consumption hotspot area and time period data to obtain an energy use strategy.
4. The intelligent management and control method for building a factory based on the cloud according to claim 1 is characterized in that: Step S6 comprises: Extracting features of the construction process plan, safety situation assessment data, resource scheduling strategy, environmental optimization plan and energy use strategy to obtain a multi-dimensional feature vector, and standardizing the multi-dimensional feature vector to obtain a standardized feature matrix; Performing dimensionality reduction analysis on the standardized feature matrix to obtain a key feature set after dimensionality reduction, and constructing a multidimensional association network based on the key feature set to obtain a factor association graph; Performing association analysis on the factor association graph to obtain association factor clusters, and performing cluster analysis on the association factor clusters to obtain a multi-level decision-making factor system; Constructing a hierarchical analysis structure based on the multi-level decision-making factor system to obtain initial factor weights, and revising the initial factor weights to obtain revised factor weights; Calculating a comprehensive score according to the modified factor weights and the original data to obtain a multi-dimensional comprehensive evaluation result, and analyzing the multi-dimensional comprehensive evaluation result to obtain a non-inferior solution set; For each solution in the non-inferior solution set, a comprehensive score is calculated based on a preset evaluation index, each solution in the non-inferior solution set is sorted according to the comprehensive score, and the top P solutions with the highest scores are selected as a candidate decision solution set, where P is a positive integer greater than 1, and scenario analysis and sensitivity analysis are performed on the candidate decision solution set to obtain a solution reliability evaluation result; Based on the reliability assessment results and decision-making objectives of the scheme, a comprehensive assessment is performed on the candidate decision scheme set to obtain a management decision scheme for building a factory in the cloud.
5. An intelligent control device for a cloud-based factory construction, used to execute the intelligent control method for a cloud-based factory construction as described in any one of claims 1 to 4, characterized in that: The intelligent control device for building a factory based on the cloud includes: The fusion module is used to perform hierarchical data collection and heterogeneous data fusion on the cloud construction factory to generate multi-level equivalent digital representation data, specifically including: collecting the structural data of the power support system, steel platform system, rack system and template system of the cloud construction factory to obtain the original data of the structural system, and collecting the status data of the safety protection system to obtain the original data of the safety system; collecting the environmental parameters and equipment operation data of the intelligent monitoring system of the cloud construction factory to obtain the original data of the monitoring system, and performing time synchronization processing on the original data of the structural system, the original data of the safety system and the original data of the monitoring system to obtain time-aligned multi-source heterogeneous data; performing data cleaning on the time-aligned multi-source heterogeneous data to obtain cleaned Multi-source data, and performing data standardization processing on the cleaned multi-source data to obtain standardized multi-source data; performing feature extraction on the standardized multi-source data to obtain a multi-dimensional feature vector, and performing dimensionality reduction processing on the multi-dimensional feature vector to obtain feature data after dimensionality reduction; performing spatial registration on the feature data after dimensionality reduction to establish a spatial correspondence between multi-source data to obtain spatially aligned feature data, and performing multi-scale fusion on the spatially aligned feature data to obtain multi-scale fused feature data; performing semantic annotation on the multi-scale fused feature data to construct a semantic association network to obtain semantically associated multi-level data, and converting the semantically associated multi-level data into a target data format to generate multi-level equivalent digital representation data; An extraction module is used to extract and analyze the spatial data in the multi-level equivalent digital representation data, determine the spatial layout and mutual relationship of system components, calculate the target assembly sequence and target assembly path between system components, and generate a construction process plan according to the target assembly sequence and the target assembly path; A comparison module is used to perform multi-dimensional safety parameter analysis on the plant structure data and safety protection system data of the multi-level equivalent digital representation data to obtain a safety state parameter set, compare the safety state parameter set with a preset safety threshold, and calculate a safety assessment index; generate safety situation assessment data for a factory-like operating environment based on the safety assessment index, specifically including: performing structural stress analysis on the plant structure data in the multi-level equivalent digital representation data to obtain structural stress distribution data, and performing statistical analysis on the structural stress distribution data to obtain structural stress parameters; performing complete analysis on the safety protection system data in the multi-level equivalent digital representation data The method comprises the steps of: performing an integrity check to obtain safety protection system status data, classifying the safety protection system status data, and obtaining safety protection parameters; combining the structural stress parameters and the safety protection parameters to construct a multidimensional safety state vector, obtaining an initial safety state parameter set, and normalizing the initial safety state parameter set to obtain a standardized safety state parameter set; comparing each parameter in the standardized safety state parameter set with a corresponding preset safety threshold, calculating a deviation value of each parameter, obtaining a safety parameter deviation set, and performing weighted summation on the safety parameter deviation set to obtain a comprehensive safety assessment index, wherein the deviation value is calculated by the following formula: ; in, For the i The deviation value of the parameter, is the corresponding safety threshold, is the normalized parameter value; Performing fuzzy processing on the comprehensive safety assessment index to obtain a safety level assessment result, generating a safety risk distribution map based on the safety level assessment result, performing regional identification and quantitative analysis on the safety risk distribution map, and obtaining safety situation assessment data for a factory-like operating environment; The analysis module is used to extract resource data from the multi-level equivalent digital representation data and perform resource status analysis to generate a resource scheduling strategy, which specifically includes: extracting features from the multi-level equivalent digital representation data to obtain resource feature vectors, and performing dimensionality reduction processing on the resource feature vectors to obtain resource data after dimensionality reduction; performing cluster analysis on the resource data after dimensionality reduction to obtain resource category division results, and labeling the resource category division results to obtain resource category label sets; grouping resources according to the resource category label sets to obtain resource grouping data, and performing time series correlation analysis on the resource grouping data to obtain resource utilization time series data; performing trend forecasting on the resource utilization time series data to obtain resource demand forecasting results, and calculating resource gaps according to the resource demand forecasting results to obtain resource gap data; constructing a resource scheduling optimization objective function based on the resource gap data to obtain a resource scheduling objective function, and setting constraints on the resource scheduling objective function to obtain a resource scheduling constraint set; wherein, the objective function , For the i Class resources are t The gap in time, is the scheduling cost, is the first weight coefficient and is the second weight coefficient, n is the total number of resource types, T is the total number of time units in the planning period; solving the resource scheduling objective function to obtain a set of candidate resource scheduling solutions, and performing multi-objective evaluation on the set of candidate resource scheduling solutions to obtain a resource scheduling strategy; Identification module, used to perform time series analysis and pattern recognition on real-time collected environmental monitoring data and energy consumption data, and generate environmental optimization plans and energy use strategies; The decision-making module is used to perform multi-dimensional correlation analysis on the construction process plan, the safety situation assessment data, the resource scheduling strategy, the environmental optimization plan and the energy use strategy to generate a management decision-making plan for the cloud-based construction factory.
6. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, an intelligent management and control method for building a factory based on the cloud as described in any one of claims 1 to 4 is implemented.
Citation Information
Patent Citations
Intelligent construction site safety management platform system based on data interaction
CN116911759A
Road concrete construction quality management system
CN117495210A
Intelligent construction method and system based on EMPC mode
CN117495316A
Construction intelligent management method and system for project supervision
CN117993866A
Building machine operation modal monitoring system based on cloud intelligent construction
CN118443086A
