Fire-fighting smoke exhaust duct laying planning system based on industrial cloud platform

Abstract

This invention discloses a fire-fighting smoke exhaust duct laying planning system based on an industrial cloud platform, belonging to the field of duct laying planning technology. It includes: a building information acquisition module for collecting building data; a cloud-edge collaborative computing module for constructing a smoke exhaust simulation model using the distributed computing resources of the industrial cloud platform and performing real-time parameter iteration through edge nodes; a duct path planning module for generating multiple candidate laying paths using an improved algorithm and optimizing the path based on duct resistance coefficients and construction cost constraints; and a scheme verification and optimization module for performing smoke flow dynamics simulation verification on the optimal path scheme, identifying scheme defects, and iteratively optimizing path parameters using a genetic algorithm. This invention achieves a synergistic improvement in the accuracy, efficiency, and compliance of fire-fighting smoke exhaust duct laying planning through a cloud-edge collaborative intelligent computing architecture, a dual-objective optimization mechanism integrating multi-dimensional constraints, and a dynamic adaptive path planning and verification system.

Description

Technical Field

[0001] This invention relates to the field of pipeline laying planning technology, specifically a fire smoke exhaust pipeline laying planning system based on an industrial cloud platform. Background Technology

[0002] In modern building engineering, the design and installation of fire exhaust ducts are of paramount importance, as they relate to the life and property safety of people inside the building. Most existing fire exhaust duct installation schemes rely on manual experience and traditional design methods. This approach is not only inefficient but also prone to design flaws and fails to fully consider the complex environment and smoke exhaust requirements inside the building. Therefore, how to effectively plan and optimize fire exhaust ducts has become an urgent problem to be solved in building fire protection design.

[0003] Traditional duct design methods lack comprehensive consideration of the three-dimensional structure of buildings and the complex environmental characteristics, often ignoring various complex situations that smoke exhaust ducts may encounter in practical applications, such as airflow dynamics, temperature field distribution, and interference from building components. Due to the neglect of these factors, the duct design scheme cannot achieve optimal results and may even affect personnel evacuation and fire protection effectiveness in the event of a fire. With the development of information technology, industrial cloud platforms and edge computing technology have provided new solutions for the optimization of building fire protection systems. Through the powerful computing capabilities of cloud platforms and the real-time data processing capabilities of edge nodes, more accurate smoke exhaust simulation and path planning can be achieved. Fire smoke exhaust duct laying planning systems based on industrial cloud platforms can provide more scientific and efficient solutions for building fire protection design through big data analysis, smoke flow simulation, and optimization algorithms. Summary of the Invention

[0004] The purpose of this invention is to solve the problems mentioned above by proposing a fire smoke exhaust duct laying planning system based on an industrial cloud platform.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] A fire-fighting smoke exhaust duct laying planning system based on an industrial cloud platform, comprising:

[0007] The module includes a building information acquisition module, a cloud-edge collaborative computing module, a pipeline path planning module, and a solution verification and optimization module.

[0008] Building Information Acquisition Module: Used to collect 3D building model data, fire protection code parameters and environmental feature information. Based on the preprocessing engine of the industrial cloud platform, the data is standardized and the processed information is transmitted to the cloud-edge collaborative computing module.

[0009] Cloud-edge collaborative computing module: Based on the standardized data output by the building information acquisition module, it uses the distributed computing resources of the industrial cloud platform to build a smoke exhaust simulation model, performs real-time parameter iteration through edge nodes, and transmits the calculation results to the pipeline path planning module;

[0010] Pipeline route planning module: Combining the simulation results output by the cloud-edge collaborative computing module, multiple candidate laying paths are generated using an improved algorithm. Based on the pipeline resistance coefficient and construction cost constraints, the optimal path is selected and transmitted to the scheme verification and optimization module.

[0011] Scheme Verification and Optimization Module: The optimal path scheme output by the pipeline path planning module is verified by flue gas dynamics simulation. The big data analysis function of the industrial cloud platform is used to identify the defects of the scheme. The path parameters are iteratively optimized through genetic algorithm to generate the final pipeline laying plan scheme.

[0012] Preferably, the building information acquisition module specifically includes:

[0013] 3D model acquisition unit: Collects 3D point cloud data of the building's internal structure, generates component models, and obtains a library of standard parameters, including smoke exhaust standards for different areas and duct wind speed limits.

[0014] Environmental feature extraction unit: collects temperature field distribution, airflow disturbance coefficient and combustible material distribution density inside the building, and performs data cleaning through the edge nodes of the industrial cloud platform;

[0015] Data fusion unit: It aligns 3D model data, standard parameters and environmental features in time and space to generate building information maps in a unified coordinate system. The data update frequency is once per hour.

[0016] Preferably, the data fusion unit specifically includes:

[0017] Coordinate transformation subunit: The seven-parameter transformation method is used to transform the local coordinate system of different acquisition devices to the building geodetic coordinate system;

[0018] Feature association sub-unit: Identify the relationship between building components and smoke exhaust requirements through deep learning algorithms, and obtain data on the relationship between building components and smoke exhaust requirements.

[0019] Preferably, the step of identifying the correlation between building components and smoke extraction requirements using deep learning algorithms specifically includes:

[0020] Collect BIM model data of building projects that have passed fire safety inspection, extract the annotation information of evacuation routes and corresponding smoke exhaust outlets, including the three-dimensional coordinate range of evacuation routes, the location coordinates of smoke exhaust outlets, and fire safety code requirements data, and divide the data into training set and validation set in a 7:3 ratio;

[0021] Geometric and environmental features of evacuation routes are extracted from the 3D model. Smoke exhaust demand constraints are converted into feature vectors and encoded. Based on the 3D-CNN+Transformer model, the 3D-CNN extracts spatial features, and the Transformer learns the associated weights. The dual-branch output determines whether smoke exhaust outlets are to be set and their locations. The model is trained with the Adam optimizer until the accuracy on the validation set is ≥95%.

[0022] Input new building information, and output the smoke exhaust vent setting judgment and coordinates through model inference. Mark the locations with a confidence score of ≥90 and associate them with evacuation routes. Collect error samples that are manually verified and update the model through federated learning.

[0023] Preferably, the cloud-edge collaborative computing module specifically includes:

[0024] Distributed computing scheduling unit: Based on the central node of the industrial cloud platform, global simulation of the smoke flow field is performed with a time step of 1 second. The edge nodes are used to process the turbulence model calculation of local areas with a grid size ≤ 0.5m.

[0025] Parameter iteration unit: Based on real-time environmental parameters collected from edge nodes, dynamically corrects the boundary conditions of the smoke exhaust simulation model;

[0026] The calculation results compression unit uses wavelet transform to compress the three-dimensional smoke flow field data, with the compression ratio controlled within 10:1, so that the data transmission delay is ≤200ms.

[0027] Preferably, the pipeline route planning module specifically includes:

[0028] Obstacle Modeling Unit: Converts the components in the 3D building model into obstacles for path planning. The voxel method is used for spatial discretization, and the voxel size is set to 100mm×100mm×100mm.

[0029] Improved algorithm unit: Introducing a pipeline resistance cost factor based on the traditional algorithm;

[0030] Multi-objective decision-making unit: Selects the optimal solution from candidate paths that meets the following conditions: smoke exhaust resistance coefficient ≤ 0.02, path length is more than 5% shorter than the benchmark scheme, and there are 0 conflict points with the building structure.

[0031] Preferably, the improved algorithm unit specifically includes:

[0032] Dynamic cost update sub-unit: When the path passes through a high-temperature area of ​​≥80℃, the cost coefficient of the pipeline insulation layer is automatically increased, the cost increases by 20%, and the resistance calculation model is corrected;

[0033] Path smoothing sub-unit: The B-spline curve is fitted to the polyline path generated by the algorithm to make the pipe turning angle ≤45° and the local resistance loss reduction rate ≥15%.

[0034] Preferably, the scheme verification and optimization module specifically includes:

[0035] Smoke flow simulation unit: FDS software is used to simulate fire scenarios for the optimal path scheme, and output the temperature field and visibility distribution at different times, so that the visibility of the evacuation route is ≥10m within 300s;

[0036] Defect identification unit: Based on historical case data stored on the industrial cloud platform, it identifies potential abnormal datasets of solutions using the random forest algorithm;

[0037] Parameter optimization unit: With the goal of smoke exhaust efficiency ≥90% and construction cost ≤105% of budget, the pipeline path parameters are optimized by genetic algorithm. The number of iterations is set to 50 generations, the population size is 30 in each iteration, and the final pipeline path parameters are output.

[0038] Preferably, the parameter optimization unit specifically includes:

[0039] A dual-objective fitness function is constructed based on smoke extraction efficiency and construction cost as inputs.

[0040] Constraint processing subunit: Apply a penalty factor to parameter combinations that violate the specifications, reduce the fitness by 50%, and make the optimization results meet fire protection requirements.

[0041] Preferably, the constraint processing subunit specifically includes:

[0042] Obtain the mandatory constraints related to pipeline laying in the fire protection code, form a constraint list, obtain the threshold range corresponding to each constraint, set differentiated penalty factors according to the importance of different constraint items, apply a 50% fitness penalty to parameter combinations that violate safety constraints, apply a 20% fitness penalty to non-core constraints, and store the penalty factors in the constraint rule library of the industrial cloud platform.

[0043] In each generation of the genetic algorithm, the current candidate parameter combination is extracted and compared with the constraint list one by one. The spatial relationship between the pipeline and the building components is verified by the three-dimensional collision detection algorithm.

[0044] For parameter combinations that are determined to violate constraints, the corrected fitness value is calculated based on the corresponding penalty factor;

[0045] The corrected fitness directly participates in the selection, crossover and mutation processes of the population. After each generation, the percentage of parameter combinations that satisfy all constraints in the population is counted. When the percentage is ≥80%, the current penalty intensity is maintained. When the percentage is <50%, the penalty factor is increased to guide the population to evolve in a direction that conforms to the norm, until the percentage is ≥80% for three consecutive generations.

[0046] Compared with the prior art, the beneficial effects of the present invention are:

[0047] 1. By combining the global computing power of the industrial cloud platform with the real-time processing advantages of edge nodes, a distributed computing scheduling unit enables collaborative computation of global smoke flow field simulation and local turbulence model, with dynamically corrected formulas supporting real-time iteration of boundary conditions. This architecture ensures both the macroscopic accuracy of fire smoke exhaust simulation and, through localized processing at edge nodes, keeps data transmission latency within 200ms, thus solving the response lag problem of traditional centralized computing in complex building environments.

[0048] 2. Based on the 3D-CNN+Transformer model of the association between building components and smoke exhaust demand, the spatial matching relationship between evacuation routes and smoke exhaust outlets is accurately identified. Combined with the pipe resistance cost factor introduced by the improved algorithm unit, a dual-objective fitness function that takes into account both smoke exhaust efficiency and construction cost is constructed. At the same time, through the differentiated penalty mechanism of the constraint processing sub-unit, it is ensured that the optimization results achieve a balance between performance and economy while meeting fire protection regulations, which greatly improves the efficiency compared with the traditional single-objective planning scheme.

[0049] 3. By combining voxel-based obstacle modeling with improved algorithms, the system can not only automatically avoid building obstacles, but also dynamically adjust the cost coefficient of the insulation layer for high-temperature areas ≥80℃. Furthermore, B-spline curve fitting ensures that pipe turning angles are ≤45°, reducing local resistance loss by ≥15%. In addition, the scheme verification and optimization module integrates FDS smoke flow simulation with iterative optimization using genetic algorithms, achieving a closed-loop process from path generation to defect identification and parameter correction, thus improving the fire safety acceptance pass rate. Attached Figure Description

[0050] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0051] Figure 1 This is a system framework diagram of the present invention;

[0052] Figure 2 This is a diagram showing the internal system framework of the building information acquisition module in this invention.

[0053] Figure 3 This is an internal system framework diagram of the cloud-edge collaborative computing module in this invention;

[0054] Figure 4 This is an internal system framework diagram of the pipeline path planning module in this invention;

[0055] Figure 5 This is an internal system framework diagram of the scheme verification and optimization module in this invention. Detailed Implementation

[0056] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0057] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0058] Please see Figure 1 As shown, a fire-fighting smoke exhaust duct laying planning system based on an industrial cloud platform includes:

[0059] The module includes a building information acquisition module, a cloud-edge collaborative computing module, a pipeline path planning module, and a solution verification and optimization module.

[0060] Building Information Acquisition Module: Used to collect 3D building model data, fire protection code parameters and environmental feature information. Based on the preprocessing engine of the industrial cloud platform, the data is standardized and the processed information is transmitted to the cloud-edge collaborative computing module.

[0061] Cloud-edge collaborative computing module: Based on the standardized data output by the building information acquisition module, it uses the distributed computing resources of the industrial cloud platform to build a smoke exhaust simulation model, performs real-time parameter iteration through edge nodes, and transmits the calculation results to the pipeline path planning module;

[0062] Pipeline route planning module: Combining the simulation results output by the cloud-edge collaborative computing module, multiple candidate laying paths are generated using an improved algorithm. Based on the pipeline resistance coefficient and construction cost constraints, the optimal path is selected and transmitted to the scheme verification and optimization module.

[0063] Scheme verification and optimization module: The optimal path scheme output by the pipeline path planning module is verified by flue gas dynamics simulation. The big data analysis function of the industrial cloud platform is used to identify the defects of the scheme. The path parameters are iteratively optimized through genetic algorithm to generate the final pipeline laying plan scheme.

[0064] This system introduces cloud-edge collaborative computing technology. By leveraging the distributed computing resources of the industrial cloud platform and the real-time data processing of edge nodes, it not only improves computing performance but also ensures the real-time nature and accuracy of the data. This collaborative computing architecture provides strong technical support for complex computational tasks in fire exhaust duct laying, especially smoke flow dynamics simulation and path planning.

[0065] Please see Figure 2 As shown, the building information acquisition module specifically includes:

[0066] 3D model acquisition unit: Collects 3D point cloud data of the building's internal structure, generates component models, and obtains a library of standard parameters, including smoke exhaust standards for different areas and duct wind speed limits.

[0067] Environmental feature extraction unit: collects temperature field distribution, airflow disturbance coefficient and combustible material distribution density inside the building, and performs data cleaning through the edge nodes of the industrial cloud platform;

[0068] Data fusion unit: It aligns 3D model data, standard parameters and environmental features in time and space to generate building information maps in a unified coordinate system. The data update frequency is once per hour.

[0069] The system uses deep learning algorithms combined with 3D building model data to achieve intelligent prediction of the location of smoke exhaust outlets. This method significantly improves the accuracy of smoke exhaust outlet layout, avoids errors caused by manual prediction, and enhances the adaptability and automation of the overall system.

[0070] The data fusion unit specifically includes:

[0071] Coordinate transformation subunit: The seven-parameter transformation method is used to transform the local coordinate system of different acquisition devices to the building geodetic coordinate system;

[0072] Feature association sub-unit: Identify the relationship between building components and smoke exhaust requirements through deep learning algorithms, and obtain data on the relationship between building components and smoke exhaust requirements.

[0073] The specific steps of identifying the correlation between building components and smoke extraction requirements using deep learning algorithms include:

[0074] Collect BIM model data of building projects that have passed fire safety inspection, extract the annotation information of evacuation routes and corresponding smoke exhaust outlets, including the three-dimensional coordinate range of evacuation routes, the location coordinates of smoke exhaust outlets, and fire safety code requirements data, and divide the data into training set and validation set in a 7:3 ratio;

[0075] Geometric and environmental features of evacuation routes are extracted from the 3D model. Smoke exhaust demand constraints are converted into feature vectors and encoded. Based on the 3D-CNN+Transformer model, the 3D-CNN extracts spatial features, and the Transformer learns the associated weights. The dual-branch output determines whether smoke exhaust outlets are to be set and their locations. The model is trained with the Adam optimizer until the accuracy on the validation set is ≥95%.

[0076] Input new building information, and output the smoke exhaust vent setting judgment and coordinates through model inference. Mark the locations with a confidence score of ≥90 and associate them with evacuation routes. Collect error samples that are manually verified and update the model through federated learning.

[0077] The data fusion unit generates a building information map with an update frequency of 1 hour by updating building information in real time and combining it with 3D point cloud data and environmental features. This enables the system to adaptively optimize itself based on the latest building structure and environmental information, ensuring a high degree of fit between the smoke exhaust duct planning and the actual environment.

[0078] Please see Figure 3 As shown, the cloud-edge collaborative computing module specifically includes:

[0079] Distributed computing scheduling unit: Based on the central node of the industrial cloud platform, global simulation of the smoke flow field is performed with a time step of 1 second. The edge nodes are used to process the turbulence model calculation of local areas with a grid size ≤ 0.5m.

[0080] Parameter Iteration Unit: Based on real-time environmental parameters collected from edge nodes, the boundary conditions of the smoke exhaust simulation model are dynamically corrected. The dynamic correction formula is as follows: In the formula, For the updated model parameters, For the current parameter, The learning rate is 0.01. The gradient of the objective function;

[0081] The calculation results compression unit uses wavelet transform to compress the three-dimensional smoke flow field data, with the compression ratio controlled within 10:1, so that the data transmission delay is ≤200ms.

[0082] Please see Figure 4 As shown, the pipeline path planning module specifically includes:

[0083] Obstacle Modeling Unit: Converts the components in the 3D building model into obstacles for path planning. The voxel method is used for spatial discretization, and the voxel size is set to 100mm×100mm×100mm.

[0084] Improved Algorithm Unit: Based on the traditional algorithm, a pipeline resistance cost factor is introduced. The heuristic function formula is as follows: In the formula, The total cost of the node. The actual cost from the starting point to the current node. The estimated resistance from the current node to the target point. This is the weighting coefficient, with a value of 1.2. This is the friction coefficient. This represents the pipe length from the current node to the target node. The inner diameter of the pipe. The flue gas velocity inside the duct. It is the acceleration due to gravity;

[0085] Multi-objective decision-making unit: Selects the optimal solution from candidate paths that meets the following conditions: smoke exhaust resistance coefficient ≤ 0.02, path length is more than 5% shorter than the benchmark scheme, and there are 0 conflict points with the building structure.

[0086] The improved algorithm unit specifically includes:

[0087] Dynamic cost update sub-unit: When the path passes through a high-temperature area of ​​≥80℃, the cost coefficient of the pipeline insulation layer is automatically increased, the cost increases by 20%, and the resistance calculation model is corrected;

[0088] Path smoothing sub-unit: B-spline curve fitting is performed on the polyline path generated by the algorithm to ensure that the pipe turning angle is ≤45° and the local resistance loss reduction rate is ≥15%;

[0089] An improved path planning algorithm is employed, incorporating pipeline resistance and construction cost as constraints to optimize the efficiency issues inherent in traditional path planning. Particularly in dynamic cost updates and path smoothing in high-temperature regions, the improved algorithm significantly reduces path resistance losses and construction difficulty, enhancing the system's practicality and flexibility.

[0090] Please see Figure 5 As shown, the scheme verification and optimization module specifically includes:

[0091] Smoke flow simulation unit: FDS software is used to simulate fire scenarios for the optimal path scheme, and output the temperature field and visibility distribution at different times, so that the visibility of the evacuation route is ≥10m within 300s;

[0092] Defect identification unit: Based on historical case data stored on the industrial cloud platform, it identifies potential abnormal datasets of solutions using the random forest algorithm;

[0093] Parameter optimization unit: With the goal of smoke exhaust efficiency ≥90% and construction cost ≤105% of budget, the pipeline path parameters are optimized by genetic algorithm. The number of iterations is set to 50 generations, the population size is 30 in each iteration, and the final pipeline path parameters are output.

[0094] The parameter optimization unit specifically includes:

[0095] Based on smoke extraction efficiency and construction cost as inputs, a bi-objective fitness function is constructed, where the expression of the bi-objective fitness function is: In the formula, For smoke extraction efficiency, For actual cost, For budget costs, The weighting coefficient for smoke extraction efficiency is 0.6. This is the weighting factor for budget costs, with a value of 0.4. For fitness, when fitness When the value is ≥0.85, the corresponding parameter is considered a better solution and included in the candidate parameter set;

[0096] Constraint processing subunit: Apply a penalty factor to parameter combinations that violate the specifications, reduce the fitness by 50%, and make the optimization results meet fire protection requirements;

[0097] By combining multi-objective decision-making and genetic algorithms, the method optimizes the smoke extraction efficiency and construction cost of pipeline paths, overcoming the limitations of single-objective optimization in traditional methods. The iterative and fitness handling of the genetic algorithm in path optimization, especially the penalty for dynamic constraints of fire protection codes, makes the final optimized scheme more in line with actual engineering needs.

[0098] The constraint processing subunit specifically includes:

[0099] Obtain the mandatory constraints related to pipeline laying in the fire protection code, form a constraint list, obtain the threshold range corresponding to each constraint, set differentiated penalty factors according to the importance of different constraint items, apply a 50% fitness penalty to parameter combinations that violate safety constraints, apply a 20% fitness penalty to non-core constraints, and store the penalty factors in the constraint rule library of the industrial cloud platform.

[0100] In each generation of the genetic algorithm, the current candidate parameter combination is extracted and compared with the constraint list one by one. The spatial relationship between the pipeline and the building components is verified by the three-dimensional collision detection algorithm.

[0101] For parameter combinations deemed to violate constraints, the corrected fitness value is calculated based on the corresponding penalty factor, using the following formula: In the formula, For the corrected fitness, For the original fitness, This is the penalty factor, with a value of 0.2;

[0102] The corrected fitness directly participates in the selection, crossover, and mutation processes of the population. After each generation, the percentage of parameter combinations that satisfy all constraints in the population is counted. When the percentage is ≥80%, the current penalty intensity is maintained. When the percentage is <50%, the penalty factor is increased to guide the population to evolve in a direction that conforms to the norm, until the percentage is ≥80% for three consecutive generations.

[0103] The algorithm employs a constraint processing sub-unit mechanism. By setting different penalty factors for constraints, it effectively avoids design schemes that violate specifications. At the same time, the algorithm modifies the algorithm based on the fitness of each generation of the population and dynamically adjusts the penalty intensity between different generations to ensure that the optimization direction of the system always meets the requirements of fire protection specifications, thereby greatly improving the system's performance.

[0104] In summary, the advantages of this invention are as follows:

[0105] Leveraging the powerful computing capabilities and edge computing resources of the industrial cloud platform, it is possible to process information such as 3D model data of buildings, environmental characteristics, and fire protection code parameters in real time, thereby achieving the standardization, real-time updating, and precise integration of building information. Data from different sources can be optimized through deep learning and algorithms to obtain more accurate smoke exhaust demand analysis, reducing manual intervention and improving design efficiency.

[0106] By leveraging cloud-edge collaborative computing and an improved path planning algorithm, multiple candidate laying paths can be automatically generated based on smoke exhaust simulation results, and the optimal path can be selected by considering factors such as pipe resistance coefficient and construction cost. This process not only reduces the complexity of manual design but also significantly reduces design costs and time, while improving the efficiency of the smoke exhaust system.

[0107] The scheme verification and optimization module performs smoke flow dynamics simulation on the pipeline path, identifies defects by combining historical case data, promptly discovers potential problems in the design scheme, and performs iterative optimization through genetic algorithms to ensure that the final pipeline laying scheme meets fire protection requirements and has high smoke exhaust efficiency.

[0108] The system comprehensively considers smoke extraction efficiency and construction cost through a multi-objective decision-making algorithm. While meeting fire protection requirements, it effectively reduces construction costs. Through optimization techniques such as dynamic cost updates and path smoothing, it further reduces the impact of high-temperature areas on pipelines and improves pipeline performance and reliability.

[0109] By strictly handling regulatory constraints and implementing a penalty mechanism, the optimized pipeline route not only meets the requirements for smoke extraction efficiency but also complies with fire safety regulations, preventing design schemes that violate regulations and ensuring the safety and legality of the construction process.

[0110] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to specific implementations. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A fire-fighting smoke exhaust duct laying planning system based on an industrial cloud platform, characterized in that, include: The module includes a building information acquisition module, a cloud-edge collaborative computing module, a pipeline path planning module, and a solution verification and optimization module. Building Information Acquisition Module: Used to collect 3D building model data, fire protection code parameters and environmental feature information. Based on the preprocessing engine of the industrial cloud platform, the data is standardized and the processed information is transmitted to the cloud-edge collaborative computing module. Cloud-edge collaborative computing module: Based on the standardized data output by the building information acquisition module, it uses the distributed computing resources of the industrial cloud platform to build a smoke exhaust simulation model, performs real-time parameter iteration through edge nodes, and transmits the calculation results to the pipeline path planning module; Pipeline route planning module: Combining the simulation results output by the cloud-edge collaborative computing module, multiple candidate laying paths are generated using an improved algorithm. Based on the pipeline resistance coefficient and construction cost constraints, the optimal path is selected and transmitted to the scheme verification and optimization module. Scheme verification and optimization module: The optimal path scheme output by the pipeline path planning module is verified by flue gas dynamics simulation. The big data analysis function of the industrial cloud platform is used to identify the defects of the scheme. The path parameters are iteratively optimized through genetic algorithm to generate the final pipeline laying plan scheme. The building information acquisition module specifically includes: 3D model acquisition unit: Collects 3D point cloud data of the building's internal structure, generates component models, and obtains a library of standard parameters, including smoke exhaust standards for different areas and duct wind speed limits. Environmental feature extraction unit: collects temperature field distribution, airflow disturbance coefficient and combustible material distribution density inside the building, and performs data cleaning through the edge nodes of the industrial cloud platform; Data fusion unit: It aligns 3D model data, standard parameters and environmental features in time and space to generate building information maps in a unified coordinate system. The data update frequency is once per hour. The data fusion unit specifically includes: Coordinate transformation subunit: The seven-parameter transformation method is used to transform the local coordinate system of different acquisition devices to the building geodetic coordinate system; Feature association subunit: Identifies the relationship between building components and smoke exhaust requirements through deep learning algorithms, and obtains data on the relationship between building components and smoke exhaust requirements; The specific steps of identifying the correlation between building components and smoke extraction requirements using deep learning algorithms include: Collect BIM model data of building projects that have passed fire safety inspection, extract the annotation information of evacuation routes and corresponding smoke exhaust outlets, including the three-dimensional coordinate range of evacuation routes, the location coordinates of smoke exhaust outlets, and fire safety code requirements data, and divide the data into training set and validation set in a 7:3 ratio; Geometric and environmental features of evacuation routes are extracted from the 3D model. Smoke exhaust demand constraints are converted into feature vectors and encoded. Based on the 3D-CNN+Transformer model, the 3D-CNN extracts spatial features, and the Transformer learns the associated weights. The dual-branch output determines whether smoke exhaust outlets are to be set and their locations. The model is trained with the Adam optimizer until the accuracy on the validation set is ≥95%. Input new building information, and output the smoke exhaust vent setting judgment and coordinates through model inference. Mark the locations with a confidence score of ≥90 and associate them with evacuation routes. Collect error samples that are manually verified and update the model through federated learning. The pipeline route planning module specifically includes: Obstacle Modeling Unit: Converts the components in the 3D building model into obstacles for path planning. The voxel method is used for spatial discretization, and the voxel size is set to 100mm×100mm×100mm. Improved algorithm unit: Introducing a pipeline resistance cost factor based on the traditional algorithm; Multi-objective decision-making unit: Selects the optimal solution from candidate paths that meets the following conditions: smoke exhaust resistance coefficient ≤ 0.02, path length is more than 5% shorter than the benchmark solution, and there are 0 conflict points with the building structure. The improved algorithm unit specifically includes: Dynamic cost update sub-unit: When the path passes through a high-temperature area of ​​≥80℃, the cost coefficient of the pipeline insulation layer is automatically increased, the cost increases by 20%, and the resistance calculation model is corrected; Path smoothing sub-unit: The B-spline curve is fitted to the polyline path generated by the algorithm to make the pipe turning angle ≤45° and the local resistance loss reduction rate ≥15%.

2. The fire-fighting smoke exhaust duct laying planning system based on an industrial cloud platform according to claim 1, characterized in that, The cloud-edge collaborative computing module specifically includes: Distributed computing scheduling unit: Based on the central node of the industrial cloud platform, global simulation of the smoke flow field is performed with a time step of 1 second. The edge nodes are used to process the turbulence model calculation of local areas with a grid size ≤ 0.5m. Parameter iteration unit: Based on real-time environmental parameters collected from edge nodes, dynamically corrects the boundary conditions of the smoke exhaust simulation model; The calculation results compression unit uses wavelet transform to compress the three-dimensional smoke flow field data, with the compression ratio controlled within 10:1, so that the data transmission delay is ≤200ms.

3. The fire-fighting smoke exhaust duct laying planning system based on an industrial cloud platform according to claim 2, characterized in that, The scheme verification and optimization module specifically includes: Smoke flow simulation unit: FDS software is used to simulate fire scenarios for the optimal path scheme, and output the temperature field and visibility distribution at different times, so that the visibility of the evacuation route is ≥10m within 300s; Defect identification unit: Based on historical case data stored on the industrial cloud platform, it identifies potential abnormal datasets of solutions using the random forest algorithm; Parameter optimization unit: With the goal of smoke exhaust efficiency ≥90% and construction cost ≤105% of budget, the pipeline path parameters are optimized by genetic algorithm. The number of iterations is set to 50 generations, the population size is 30 in each iteration, and the final pipeline path parameters are output.

4. The fire-fighting smoke exhaust duct laying planning system based on an industrial cloud platform according to claim 3, characterized in that, The parameter optimization unit specifically includes: A dual-objective fitness function is constructed based on smoke extraction efficiency and construction cost as inputs. Constraint processing subunit: Apply a penalty factor to parameter combinations that violate the specifications, reduce the fitness by 50%, and make the optimization results meet fire protection requirements.

5. A fire-fighting smoke exhaust duct laying planning system based on an industrial cloud platform according to claim 4, characterized in that, The constraint processing subunit specifically includes: Obtain the mandatory constraints related to pipeline laying in the fire protection code, form a constraint list, obtain the threshold range corresponding to each constraint, set differentiated penalty factors according to the importance of different constraint items, apply a 50% fitness penalty to parameter combinations that violate safety constraints, apply a 20% fitness penalty to non-core constraints, and store the penalty factors in the constraint rule library of the industrial cloud platform. In each generation of the genetic algorithm, the current candidate parameter combination is extracted and compared with the constraint list one by one. The spatial relationship between the pipeline and the building components is verified by the three-dimensional collision detection algorithm. For parameter combinations that are determined to violate constraints, the corrected fitness value is calculated based on the corresponding penalty factor; The corrected fitness directly participates in the selection, crossover and mutation processes of the population. After each generation, the percentage of parameter combinations that satisfy all constraints in the population is counted. When the percentage is ≥80%, the current penalty intensity is maintained. When the percentage is <50%, the penalty factor is increased to guide the population to evolve in a direction that conforms to the norm until the percentage is ≥80% for three consecutive generations.

Images