A communication tower multi-target cooperative system, method and electronic device

Abstract

The application provides a communication tower multi-target cooperative system and method and electronic equipment, relates to the technical field of intelligent design and operation and maintenance of communication infrastructure, and comprises the following steps: a multi-target optimization engine adopts multi-agent reinforcement learning and a preference coding network to generate a Pareto optimal solution set among cost, structural reliability and carbon emission; a three-dimensional visual design module generates a model in real time based on a BIM parameterized component library and a GIS environment model and calls finite element checking, and checking results are returned to the engine as constraints; an Internet of Things integrated module extracts features of tower vibration, strain and other sensing signals through an edge computing gateway to obtain structural deterioration characteristics; a cloud cooperative platform provides lightweight three-dimensional scheme comparison and selection and multi-role workflow cooperation; and a storage system optimization module jointly optimizes storage capacity and charging and discharging strategies, and feeds back standby power constraints to the design engine. The application improves the design efficiency and quality of the communication tower, and reduces the whole life cost and carbon emission.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent design, operation and maintenance of communication infrastructure, and in particular to a multi-objective collaboration system, method and electronic device for communication towers. Background Art

[0002] With the large-scale deployment of 5G networks, the construction density of communication towers has increased sharply, and the limitations of traditional design methods have become increasingly prominent. First, the design process usually relies on the experience of engineers, who adjust the tower height, section size and material selection through trial and error, making it difficult to simultaneously take into account construction cost, structural wind and seismic resistance, and the environmental impact over the full life cycle. Some existing single-objective optimization tools cannot dynamically balance these conflicting objectives. Second, the environmental parameters such as wind pressure and geological conditions adopted in the design stage are mostly regional standard values or static assumptions, which are disconnected from the actual service environment of the towers; the massive Internet of Things monitoring data accumulated on put-into-operation towers has not been fed back to the design stage, resulting in design margins that are either overly conservative or have hidden dangers. Third, the energy storage power supply system supporting the tower is often designed separately from the tower structure, so that the equipment load and power supply demand cannot be jointly optimized, leading to low overall resource utilization. In addition, parties involved in design, construction and operation and maintenance often face "information silos" due to inconsistent information platforms, resulting in slow change response and low collaboration efficiency. Although some tools in the industry have introduced BIM modeling and simple Internet of Things monitoring, there has always been a lack of a systematic solution that deeply integrates multi-objective intelligent optimization, three-dimensional automated design, closed-loop feedback of Internet of Things monitoring data, multi-role cloud collaboration and structure-energy storage coupling optimization. During the long-term construction of communication infrastructure, Ningxia Communication Industry Service Co., Ltd. has accumulated rich experience in dealing with special environments such as drought, sandstorm and strong temperature difference, and has digital foundations such as BIM, GIS and Internet of Things, and has a deep understanding of the existing technical bottlenecks Therefore, a multi-objective collaboration system, method and electronic device for communication towers are proposed. Summary of the Invention

[0003] This specification provides a multi-objective collaboration system, method and electronic device for communication towers, which improves the design efficiency and quality of communication towers, and reduces the full-life cost and carbon emission.

[0004] This specification provides a multi-objective collaboration system for communication towers, comprising: a multi-objective optimization engine, configured to receive design requirements and initial environmental constraints, update an environmental model based on Internet of Things feedback, and automatically generate a Pareto optimal design scheme set through multi-agent reinforcement learning; a three-dimensional visual design module, connected to the multi-objective optimization engine, configured to automatically generate a tower BIM model according to the design parameter vector output by the multi-objective optimization engine, perform structural finite element verification, and feed the verification result back to the multi-objective optimization engine; The Internet of Things (IoT) integration module includes a sensor array and an edge computing gateway deployed on the tower, used to collect structural response and environmental data, extract structural degradation characteristics, and upload them to the cloud; The cloud-based collaborative platform connects with the multi-objective optimization engine and the 3D visualization design module to display multi-dimensional indicators of design schemes, receive selected instructions and drive the generation of construction models and drawings, and realize multi-role collaboration through the workflow engine. The energy storage system optimization module is connected to the Internet of Things integration module. It optimizes the energy storage capacity and charging and discharging strategy based on historical load and new energy data, and feeds back the obtained backup power constraints to the multi-objective optimization engine. The multi-objective optimization engine is also used to periodically obtain structural degradation characteristics and extreme environment data from the IoT integration module, update its internal load and resistance probability model through Bayesian inference, and regenerate design scheme correction suggestions based on the updated model.

[0005] Optionally, the multi-objective optimization engine specifically includes: The state space definition unit has state variables including tower height, moment of inertia of main material section, material yield strength, basic wind pressure return period coefficient, seismic intensity, and peak backup power input by the energy storage system optimization module. An action space definition unit, whose action is a continuous adjustment of the state variable; The multi-agent reinforcement learning network employs a multi-agent deep deterministic policy gradient algorithm. Each agent learns policies corresponding to cost objectives, structural reliability objectives, and carbon emission objectives, and shares information among agents through a centralized Critic network. Preference encoding networks are used to map user-specified preference vectors or randomly sampled preference vectors to weight coefficients in a multi-objective reward function; The reward function is defined as R = -(α1·C) norm +α2·(1-RI norm )+α3·E norm ), where C norm To normalize the total lifecycle cost, RI norm E is a normalized structural reliability index. norm α1, α2, α3 are the normalized life-cycle carbon emission equivalents, and [α1, α2, α3] are the weight coefficients output by the preference coding network.

[0006] Optionally, after receiving the design parameter vector output by the multi-objective optimization engine, the three-dimensional visualization design module drives the parametric BIM component library to automatically reconstruct the three-dimensional model of the tower assembly, extract the node coordinates, cross-sectional properties and load combination information of the members, and automatically call the cloud-based finite element analysis microservice. If the stress ratio of all members and the displacement at the top of the tower meet the preset specifications, the response index will be returned to the multi-objective optimization engine as part of the positive reward. If any indicator exceeds the limit, the constraint violation degree will be used as a negative reward feedback, and the design parameter vector will be marked as infeasible.

[0007] Optionally, the edge computing gateway in the IoT integration module is specifically configured as follows: Fast Fourier transform is performed on the accelerometer signal to extract the first three natural frequencies, and the percentage of drift from the initial reference frequency is calculated as the overall stiffness degradation characteristic of the structure. Rainfall counting statistics are performed on the strain sensor signals to accumulate the daily fatigue damage increment of each member; By combining daily temperature and humidity data measured by temperature and humidity sensors, the equivalent corrosion depth of steel is updated based on a corrosion electrochemical model. The frequency drift percentage, daily fatigue damage increment, and equivalent corrosion depth are used as a structural degradation feature set, compressed, and uploaded to a cloud database.

[0008] Optionally, the multi-objective optimization engine periodically extracts the structural degradation feature set and corresponding meteorological extreme value data from the cloud database, and performs Bayesian inference: Using the wind pressure return period coefficient and material resistance partial factor adopted in the initial design as the prior distribution, and the bearing capacity attenuation corresponding to the measured extreme wind speed sequence and equivalent corrosion depth as the observation likelihood, the posterior distribution is updated to form a posterior load and resistance probability model that reflects the actual microenvironment of the tower.

[0009] Optionally, the cloud-based collaborative platform includes a lightweight BIM browser based on WebGL and a role-based access control workflow engine. The browser obtains a Pareto solution set and its corresponding cost-reliability-carbon emission three-dimensional scatter plot from the multi-objective optimization engine for decision-makers to select. Upon receiving the selected scheme signal, the corresponding design parameter vector is sent to the 3D visualization design module to generate a construction-level model and engineering drawings. The preset workflow is then initiated, automatically pushing tasks and documents to construction, supervision, and operation and maintenance roles, and supporting online annotation and approval status tracking.

[0010] This specification provides a multi-target cooperative method for communication towers, including: Acquire GIS geographic information, historical meteorological extreme value data and operator equipment load requirements for the target tower construction site, and initialize the state space of the multi-objective optimization engine by combining the initial backup power constraints fed back by the energy storage optimization module. Multi-agent reinforcement learning is performed. In each step of the exploration, the generated design parameter vector is transmitted to the 3D visualization design module to automatically generate a BIM model and perform finite element verification. The verification result is converted into a reward signal and returned until convergence to obtain a set of Pareto optimal design schemes. The performance indicators of each solution are displayed in a visual manner through a cloud-based collaborative platform. After receiving the selected solution, a construction-level model and engineering drawings are generated and distributed to relevant roles through a workflow engine. After the tower is completed and enters service, the edge computing gateway continuously collects structural vibration, strain and environmental data, extracts structural deterioration feature sets, and synchronously drives the energy storage system optimization module to generate energy storage optimization operation strategies. Every preset period, the multi-objective optimization engine retrieves the accumulated measured extreme value environmental data and structural degradation feature set from the cloud database, updates and corrects the internal load and resistance probability model based on Bayesian, and re-executes multi-objective optimization using the updated model, outputting design parameter correction suggestions for subsequent towers in the region, forming a data-driven adaptive design closed loop.

[0011] Optionally, the multi-agent reinforcement learning is performed, in each exploration step, the generated design parameter vector is transmitted to the 3D visualization design module to automatically generate a BIM model and perform finite element verification. The verification result is converted into a reward signal and returned, until convergence is obtained to obtain a set of Pareto optimal design schemes, including: If the finite element verification of the 3D visualization design module finds that the stress ratio or displacement of the member exceeds the limit, a constraint violation degree is generated and the current exploration step is terminated immediately as a high negative reward. Only when the verification is passed are the three target values ​​of total life cost, structural reliability and carbon emissions calculated and combined into a comprehensive reward-driven learning.

[0012] This specification also provides an electronic device, wherein the electronic device includes: A processor; and a memory storing computer-executable instructions, which, when executed, cause the processor to perform any of the methods described above.

[0013] This specification also provides a computer-readable storage medium that stores one or more programs that, when executed by a processor, implement any of the methods described above.

[0014] This invention improves the efficiency of scheme generation and iteration by using multi-agent reinforcement learning for automated multi-objective optimization and real-time finite element verification. The structural reliability and cost error rate of the optimized results are controllable, and it provides a Pareto front for cost, reliability, and carbon emissions that is difficult to obtain using traditional methods, supporting scientific decision-making. By extracting degradation features such as structural inherent frequency drift, fatigue damage, and corrosion depth through an edge computing gateway, and using Bayesian inference to periodically update the design load and environmental model, the design parameters can adapt to the real micro-environment on site, solving the persistent problem of design being out of sync with reality. This extends the service life of the tower and reduces waste caused by excessive conservatism. The peak backup power obtained from the optimized energy storage system is fed back as the structural design load. Joint optimization ensures that the tower structure design and power supply system are no longer separated, reducing life-cycle costs and material waste, and maximizing the local consumption of new energy sources such as photovoltaics, aligning with the "dual-carbon" goals of the communications industry. The lightweight BIM collaboration platform and workflow engine based on WebGL enable all parties, including designers, construction workers, supervisors, and operators, to annotate, approve, and track data in real time on a unified 3D model. This automates task flow and greatly reduces communication costs and change errors. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a schematic diagram of the structure of a multi-target cooperative system for communication towers provided in the embodiments of this specification; Figure 2 A schematic diagram illustrating the principle of a multi-target collaborative method for communication towers provided in the embodiments of this specification; Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this specification; Figure 4 This is a schematic diagram of a computer-readable medium provided for embodiments of this specification. Detailed Implementation

[0017] The following description is intended to disclose the present invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art. The basic principles of the invention defined in the following description can be applied to other embodiments, modifications, improvements, equivalents, and other technical solutions that do not depart from the spirit and scope of the invention.

[0018] The following is in conjunction with the appendix Figures 1-4Exemplary embodiments of the invention will be described more fully here. However, exemplary embodiments can be implemented in many forms and should not be construed as limiting the invention to the embodiments set forth herein. Rather, these exemplary embodiments are provided to make the invention more comprehensive and complete, and to facilitate a full communication of the inventive concept to those skilled in the art. The same reference numerals in the figures denote the same or similar elements, components, or parts, and therefore repeated descriptions of them are omitted.

[0019] Subject to the technical concept of this invention, the features, structures, characteristics or other details described in a particular embodiment may be combined in one or more other embodiments in a suitable manner.

[0020] In the description of specific embodiments, the features, structures, characteristics, or other details described in this invention are intended to enable those skilled in the art to fully understand the embodiments. However, it is not excluded that those skilled in the art can practice the technical solutions of this invention without one or more of the specific features, structures, characteristics, or other details.

[0021] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily have to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0022] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0023] The terms “and / or” or “and / or” include all combinations of any one or more of the listed items.

[0024] Figure 1 A schematic diagram of a multi-target cooperative system for communication towers provided in this specification includes: The multi-objective optimization engine 100 is configured to receive design requirements and initial environmental constraints, update the environmental model based on IoT feedback, and automatically generate a set of Pareto optimal design solutions through multi-agent reinforcement learning. The 3D visualization design module 200 is connected to the multi-objective optimization engine and is used to automatically generate the tower BIM model based on the design parameter vector output by the engine and perform structural finite element verification, and feed the verification results back to the multi-objective optimization engine. The IoT integration module 300 includes a sensor array and an edge computing gateway deployed on the tower, used to collect structural response and environmental data, extract structural degradation characteristics and upload them to the cloud; The cloud-based collaborative platform 400 connects with the multi-objective optimization engine and the 3D visualization design module to display multi-dimensional indicators of design schemes, receive selected instructions and drive the generation of construction models and drawings, and realize multi-role collaboration through the workflow engine. The energy storage system optimization module 500 is connected to the Internet of Things integration module. It optimizes the energy storage capacity and charging and discharging strategy based on historical load and new energy data, and feeds back the obtained backup power constraints to the multi-objective optimization engine. The multi-objective optimization engine is also used to periodically obtain structural degradation characteristics and extreme environment data from the IoT integration module, update its internal load and resistance probability model through Bayesian inference, and regenerate design scheme correction suggestions based on the updated model.

[0025] In the specific implementation of this specification, the multi-objective optimization engine 100 is deployed on a cloud GPU server. It receives site selection terrain data, historical meteorological extremes (such as the maximum wind speed and extreme temperature difference once every 50 years) and operator equipment load requirements from the GIS system. Simultaneously, it receives initial backup power constraints from the energy storage system optimization module 500, constructing a state space. During engine operation, it continuously sends design parameter vectors to the 3D visualization design module 200 and receives structural verification results as a reward. The 3D design module 200 has an internal library of standardized components, including single-tube towers and angle steel towers, which can be automatically assembled using the GIS environment model and verified using finite element microservices. After the tower enters service, the IoT integration module 300 collects and processes data through sensors such as acceleration, strain, and tilt angle sensors and the edge gateway 400, extracting degradation characteristics and uploading them to the cloud database. The cloud collaboration platform 400's web interface displays a 3D scatter plot of cost-reliability-carbon emissions from the Pareto solution set generated by the engine 100, allowing decision-makers to select options; selection triggers construction model generation and workflow distribution. The energy storage system optimization module 500 optimizes energy storage based on historical load and wind and solar resource data, and feeds back the maximum backup power under the worst operating conditions to the engine 100 in real time, achieving a coupled design of structure and power supply. Every six months, the engine 100 extracts accumulated measured extreme wind speeds and degradation characteristics from the cloud, updates its internal load probability model through Bayesian inference, and re-executes optimization, outputting design parameter correction suggestions for subsequent towers in the region, completing a design-monitoring adaptive closed loop. The entire system forms a closed-loop data flow from design to operation and maintenance and then back to design feedback, with interfaces between modules implemented through standardized MQTT / HTTP protocols.

[0026] Optionally, the multi-objective optimization engine specifically includes: The state space definition unit has state variables including tower height, moment of inertia of main material section, material yield strength, basic wind pressure return period coefficient, seismic intensity, and peak backup power input by the energy storage system optimization module. An action space definition unit, whose action is a continuous adjustment of the state variable; The multi-agent reinforcement learning network employs a multi-agent deep deterministic policy gradient algorithm. Each agent learns policies corresponding to cost objectives, structural reliability objectives, and carbon emission objectives, and shares information among agents through a centralized Critic network. Preference encoding networks are used to map user-specified preference vectors or randomly sampled preference vectors to weight coefficients in a multi-objective reward function; The reward function is defined as R = -(α1·C) norm +α2·(1-RI norm )+α3·E norm ), where C norm To normalize the total lifecycle cost, RI norm E is a normalized structural reliability index. norm α1, α2, α3 are the normalized life-cycle carbon emission equivalents, and [α1, α2, α3] are the weight coefficients output by the preference coding network.

[0027] In the specific implementation of this specification, the state space definition unit of the multi-objective optimization engine 100 maintains a multi-dimensional vector, including: tower height (range 20-55 meters), moment of inertia of the main material section, material yield strength (such as Q345, Q420 steel or aluminum alloy composite material), basic wind pressure return period coefficient (initial 1.0, corresponding to a 50-year return period), seismic fortification intensity (7 degrees or 8 degrees), and peak backup power input from the energy storage module (unit kW). The action space is the adjustment amount of the above continuous design variables, such as the tower height can be adjusted ±0.2 meters each time. The multi-agent reinforcement learning network adopts the MADDPG algorithm, setting up three Actor networks with the same structure, corresponding to the three objectives of cost, structural reliability, and carbon emissions, respectively. Each Actor outputs an action according to the current state, and the centralized Critic network receives the state-action pairs of all agents and outputs the joint Q value. The preference encoding network is a three-layer fully connected network. Its input is a user-dragdged preference vector (e.g., [0.7, 0.2, 0.1] indicating a greater emphasis on cost) or a vector randomly sampled during training. The output is normalized weights [α1, α2, α3]. In the calculation of the reward function R, C... norm RI is the normalized value of total life-cycle cost (including materials, transportation, installation, operation and maintenance, and recycling costs). norm E is the normalized value of the system reliability index calculated based on the JC method. normThe carbon emission equivalents are normalized values ​​based on the life cycle assessment and scaled to the [0,1] interval. During training, a preference vector is randomly generated for each episode to guide the agent through the entire Pareto front. Once the engine converges, it can output dozens of non-dominated design schemes, each with a specific combination of parameters and an objective value.

[0028] Optionally, after receiving the design parameter vector output by the multi-objective optimization engine, the three-dimensional visualization design module drives the parametric BIM component library to automatically reconstruct the three-dimensional model of the tower assembly, extract the node coordinates, cross-sectional properties and load combination information of the members, and automatically call the cloud-based finite element analysis microservice. If the stress ratio of all members and the displacement at the top of the tower meet the preset specifications, the response index will be returned to the multi-objective optimization engine as part of the positive reward. If any indicator exceeds the limit, the constraint violation degree will be used as a negative reward feedback, and the design parameter vector will be marked as infeasible.

[0029] In the specific implementation described in this specification, after receiving the design parameter vector from the multi-objective optimization engine 100, the 3D visualization design module 200 first starts the parametric BIM engine. This engine is based on a secondary development of Autodesk Revit, and its internal component library covers standard sections of single-tube towers, platforms, antenna supports, ladders, lightning rods, etc. Based on parameters such as tower height, cross-sectional dimensions, and material types, it automatically selects and assembles components to generate a complete 3D model of the tower assembly. Simultaneously, the module automatically extracts the node coordinates, cross-sectional properties, and material information of each member from the model, combines dead load, live load, wind load, and seismic action according to the "Code for Design of Tall Structures," generates a finite element analysis input file, and calls the computing microservices on Huawei Cloud to perform linear and nonlinear analyses. After the analysis is completed, the maximum stress ratio and tower top displacement value of each member are extracted. If the stress ratio of all members is less than 0.95 and the displacement is within the specified limits, the module returns the maximum stress ratio and displacement as positive indicators to the optimization engine, which then calculates a reliability index reward based on these values. If the stress ratio of any main member or web member exceeds the limit, or the displacement at the top of the tower exceeds the limit, the module immediately generates a constraint violation score proportional to the degree of violation and returns it as a negative reward (e.g., -100). Simultaneously, it sends an "infeasible" flag to the engine, terminating the exploration trajectory. This real-time verification mechanism ensures that all surviving solutions in the multi-objective optimization search space are structurally feasible, significantly improving optimization efficiency and avoiding the waste of a posteriori verification.

[0030] Optionally, the edge computing gateway in the IoT integration module is specifically configured as follows: Fast Fourier transform is performed on the accelerometer signal to extract the first three natural frequencies, and the percentage of drift from the initial reference frequency is calculated as the overall stiffness degradation characteristic of the structure. Rainfall counting statistics are performed on the strain sensor signals to accumulate the daily fatigue damage increment of each member; By combining daily temperature and humidity data measured by temperature and humidity sensors, the equivalent corrosion depth of steel is updated based on a corrosion electrochemical model. The frequency drift percentage, daily fatigue damage increment, and equivalent corrosion depth are used as a structural degradation feature set, compressed, and uploaded to a cloud database.

[0031] In the specific implementation described in this specification, the sensor group includes: MEMS triaxial accelerometers, fiber optic strain sensors, and laser inclinometers installed on key parts of the main tower structure; and temperature and humidity sensors installed in the equipment room. The edge computing gateway is based on an ARM Cortex-A72 processor and runs a Linux system. The signal processing program within the gateway performs a Hanning window truncated Fast Fourier Transform on the acceleration signal, continuously tracks the first three natural frequencies (e.g., 0.8Hz, 2.3Hz, and 4.5Hz), and compares them with the reference frequency measured at the time of completion to calculate the percentage of frequency drift. When the drift exceeds 5%, it reflects overall connection loosening or stiffness degradation. For the strain signal, a four-point rainflow counting method is used for real-time statistical analysis to calculate the equivalent stress cycle number and cumulative fatigue damage increment of each major member daily. Simultaneously, combined with temperature and humidity sensor data, based on the Rogowski model of steel corrosion, the corrosion current density is calculated using the daily average temperature and relative humidity to obtain the equivalent corrosion depth. The gateway compresses this characteristic data—frequency drift percentage, daily fatigue damage increment, and equivalent corrosion depth—and uploads it to the cloud IoT platform every 15 minutes via a 5G / NB-IoT module. This solution ensures stable transmission even in remote areas, and edge computing significantly reduces cloud load and bandwidth costs.

[0032] Optionally, the multi-objective optimization engine periodically extracts the structural degradation feature set and corresponding meteorological extreme value data from the cloud database, and performs Bayesian inference: Using the wind pressure return period coefficient and material resistance partial factor adopted in the initial design as the prior distribution, and the bearing capacity attenuation corresponding to the measured extreme wind speed sequence and equivalent corrosion depth as the observation likelihood, the posterior distribution is updated to form a posterior load and resistance probability model that reflects the actual microenvironment of the tower.

[0033] In the specific implementation of this specification, the multi-objective optimization engine 100 includes a Bayesian update module that automatically pulls the structural degradation feature set of the past six months and the meteorological extreme value records of the corresponding period from the cloud database every six months. The specific update process is as follows: For basic wind pressure, the initial prior distribution is the extreme value type I distribution used in the design, with parameters based on a 50-year return period; the daily maximum wind speed sequence within the past six months is obtained through actual measurement, and the extreme value type I distribution is fitted as the likelihood function. The posterior distribution is calculated using the Markov chain Monte Carlo method to obtain the updated wind pressure return period coefficient, which may be greater than 1 (if the actual wind speed is greater than the design value) or less than 1. For material resistance, the initial prior adopts a log-normal distribution, with its mean based on the design strength and the coefficient of variation taken from the specification value; the main material section loss rate is calculated based on the equivalent corrosion depth, and then the bearing capacity attenuation coefficient is deduced. Using this attenuation sequence as observation data, the posterior distribution of the resistance partial coefficient is updated to obtain a degradation model that reflects the local corrosion environment (such as the special impact of drought and sandstorms in Ningxia on coating wear). The engine replaces the original model with the updated posterior load and resistance probability model, reruns multi-objective optimization, and generates an improved design scheme set suitable for the microenvironment. This process achieves a feedback loop from service data to the design model, enabling newly built towers to adapt to the real environmental characteristics of the region.

[0034] Optionally, the cloud-based collaborative platform includes a lightweight BIM browser based on WebGL and a role-based access control workflow engine. The browser obtains a Pareto solution set and its corresponding cost-reliability-carbon emission three-dimensional scatter plot from the multi-objective optimization engine for decision-makers to select. Upon receiving the selected scheme signal, the corresponding design parameter vector is sent to the 3D visualization design module to generate a construction-level model and engineering drawings. The preset workflow is then initiated, automatically pushing tasks and documents to construction, supervision, and operation and maintenance roles, and supporting online annotation and approval status tracking.

[0035] In the specific implementation described in this specification, the cloud-based collaborative platform 400 is based on a B / S architecture. The front-end uses the Three.js library to implement lightweight WebGL BIM model rendering, enabling smooth display of tower models with millions of triangular facets on multiple terminals such as tablets and PCs. Multiple roles are included, such as design engineers, structural reviewers, construction managers, supervisors, and operator representatives. The system uses an RBAC model to strictly define permissions. After the multi-objective optimization engine generates a Pareto solution set, the platform front-end displays the cost, reliability, and carbon emission indicators of each solution in a 3D scatter plot, and links to the lightweight tower model that can be viewed with a click. Decision-makers (such as operators and design managers) compare and select solutions online. After selecting a solution, the platform sends the corresponding design parameter vector to the cloud service of the 3D visualization design module, generating a construction-level BIM model and a complete set of engineering drawings (including a bill of materials and cutting diagrams). Subsequently, the workflow engine automatically initiates the "design-review-construction handover" process, pushing approval tasks to the structural reviewer. The reviewer can directly annotate issues on the model, and after approval, the process automatically flows to the construction and supervision roles. Construction workers can view 3D installation animations and drawings via a mobile app, while supervisors record on-site conditions. All annotations and approval statuses are recorded, forming a traceable collaborative loop. The platform also integrates with WeChat Work API to enable instant messaging, completely resolving the information silo problem in traditional collaboration.

[0036] The energy storage system optimization module 500 is deployed in the cloud. The algorithm employs a non-dominated sorting genetic algorithm (NSGA-II) with an elitist strategy. Decision variables are encoded as follows: the rated capacity of the energy storage lithium battery pack (e.g., 10-50 kWh), the installed capacity of the photovoltaic modules (e.g., 5-20 kWp), and the charging and discharging power values ​​for each time period of the day. The first fitness function is the average annual total cost, including the discounted value of initial equipment investment, maintenance costs, and grid purchase costs; the second fitness function is the local renewable energy consumption rate, i.e., the proportion of local photovoltaic power generation directly utilized by the tower equipment. During optimization, the historical load curves of the towers (including communication equipment, air conditioning, etc.) obtained from the IoT module and the daily solar irradiance predicted by local weather forecasts are used as inputs. Constraints are set as follows: under the extreme scenario of three consecutive days without sunlight, the energy storage system must ensure that the backup power duration of the communication equipment and basic monitoring system is no less than 24 hours. After obtaining the Pareto front through NSGA-II iterations, the system automatically selects the solution corresponding to the worst operating condition, extracts the maximum backup power value under that condition (i.e., the peak power that the energy storage discharge terminal needs to support during mains power outage), and pushes it to the multi-objective optimization engine in real time via the MQTT protocol. The engine adds this power value to the equipment load list and may trigger adjustments to the antenna platform position or structural reinforcement, achieving integrated optimization of structure and energy storage. To address the bottleneck of lithium battery efficiency degradation in the low-temperature environment of Ningxia winters, a temperature control energy consumption sub-model is also embedded in the optimization model to ensure that the optimization results are consistent with reality.

[0037] Figure 2 A schematic diagram illustrating the principle of a multi-target cooperative method for communication towers provided in this specification includes: S110: Obtain GIS geographic information, historical meteorological extreme value data and operator equipment load requirements of the target tower construction site, and initialize the state space of the multi-objective optimization engine by combining the initial backup power constraints fed back by the energy storage optimization module. S120: Performs multi-agent reinforcement learning. In each step of exploration, the generated design parameter vector is transmitted to the 3D visualization design module to automatically generate a BIM model and perform finite element verification. The verification result is converted into a reward signal and returned until convergence to obtain a set of Pareto optimal design schemes. S130: Displays the performance indicators of each solution in a visual manner through a cloud-based collaborative platform. After receiving the selected solution, it triggers the generation of construction-level models and engineering drawings, and distributes them to relevant roles through a workflow engine. S140: After the tower is completed and enters service, the edge computing gateway continuously collects structural vibration, strain and environmental data, extracts structural deterioration feature sets, and synchronously drives the energy storage system optimization module to generate energy storage optimization operation strategies. S150: Every preset cycle, the multi-objective optimization engine retrieves the accumulated measured extreme environmental data and structural degradation feature set from the cloud database, updates and corrects the internal load and resistance probability model based on Bayesian, and re-executes multi-objective optimization using the updated model, outputting design parameter correction suggestions for subsequent towers in the region, forming a data-driven adaptive design closed loop.

[0038] In the specific implementation described in this specification, the system obtains GIS data (including DEM elevation, contour lines, and surrounding buildings) of the target location via API, accesses a meteorological database to extract historical extreme values ​​such as the annual maximum wind speed and minimum temperature over the past 30 years, and receives equipment weight, power consumption, and backup power requirements input by the operator. Simultaneously, the energy storage optimization module provides an initial backup power estimate based on data from similar nearby sites, and the multi-objective optimization engine uses this information to initialize the state vector.

[0039] The engine initiates multi-agent reinforcement learning, with the Actor network outputting continuous adjustments to design parameters. The design vector generated at each step is sent instantly to the 3D visualization design module. The module automatically generates a BIM model in the cloud and calls finite element analysis. If a member exceeds limits, a large negative reward is returned and the trajectory is terminated; if successful, the life-cycle cost, reliability indicators based on the analysis results, and carbon emissions are calculated and combined into a comprehensive reward signal, which is then returned. After tens of thousands of iterations of training, the engine converges and outputs a set of Pareto optimal solutions.

[0040] The cloud-based collaborative platform displays all proposals in the form of 3D scatter plots and model previews. Once the operator and the chief designer jointly select a proposal, the platform triggers the generation of construction-level models and drawings, and pushes the tasks to each role according to the preset workflow. The construction unit confirms receipt on the platform and begins material preparation.

[0041] After the tower construction was completed, the construction team installed accelerometers, strain gauges, and inclinometers at key sections of the tower based on the sensor layout map generated by the 3D visualization design module, completed commissioning, and connected them to the edge gateway. Once the tower entered service, the gateway continuously extracted and uploaded features, and the energy storage optimization module dynamically adjusted its charging and discharging strategies based on real-time load and weather forecasts, providing data support for the operation and maintenance phase.

[0042] Every six months, the multi-objective optimization engine automatically performs a closed-loop update task: it retrieves data such as extreme wind speed, frequency drift, and corrosion depth accumulated by the IoT module, and updates and corrects its internal load and resistance models using Bayesian methods. Subsequently, based on the new model, the multi-objective optimization is rerun, and the generated design parameter correction suggestions (such as "increase the wall thickness of the main tower material in this area by 1.5mm" or "upgrade the flange connection bolts by one grade") are pushed to the design team through the cloud platform's knowledge base. This guides the subsequent construction or renovation of towers in the area, completing an adaptive evolutionary closed loop driven by real service data. This method has been validated in multiple scenarios at field test bases in the Helan Mountains of Ningxia and the edge of the Mu Us Desert, demonstrating its reliability and adaptability.

[0043] Optionally, S120 includes: If the finite element verification of the 3D visualization design module finds that the stress ratio or displacement of the member exceeds the limit, a constraint violation degree is generated and the current exploration step is terminated immediately as a high negative reward. Only when the verification is passed are the three target values ​​of total life cost, structural reliability and carbon emissions calculated and combined into a comprehensive reward-driven learning.

[0044] In the specific implementation of this specification, a strict failure termination mechanism is set up during the reinforcement learning exploration process to ensure that all Pareto solutions in the final output are feasible. After the 3D visualization design module completes a finite element analysis, its verification engine checks all members one by one: if the stress ratio (calculated stress / design strength) of any member exceeds the threshold of 0.95, or the maximum displacement at the top of the tower exceeds the limit of H / 75 (H is the tower height), then the design point is considered infeasible. At this time, the module calculates the sum of the excesses as the constraint violation degree and multiplies it by a preset high penalty coefficient (such as -500) as an immediate reward and returns it to the multi-objective optimization engine. After the engine's environment simulator receives this negative reward, it marks the end of this episode and stops exploring in this direction. The relevant Critic network will also be updated to stay away from such infeasible regions. This process greatly improves search efficiency, avoids wasting computational resources in infeasible regions, and naturally guarantees the engineering feasibility of the final Pareto optimal solution set.

[0045] This invention improves the efficiency of scheme generation and iteration by using multi-agent reinforcement learning for automated multi-objective optimization and real-time finite element verification. The structural reliability and cost error rate of the optimized results are controllable, and it provides a Pareto front for cost, reliability, and carbon emissions that is difficult to obtain using traditional methods, supporting scientific decision-making. By extracting degradation features such as structural inherent frequency drift, fatigue damage, and corrosion depth through an edge computing gateway, and using Bayesian inference to periodically update the design load and environmental model, the design parameters can adapt to the real micro-environment on site, solving the persistent problem of design being out of sync with reality. This extends the service life of the tower and reduces waste caused by excessive conservatism. The peak backup power obtained from the optimized energy storage system is fed back as the structural design load. Joint optimization ensures that the tower structure design and power supply system are no longer separated, reducing life-cycle costs and material waste, and maximizing the local consumption of new energy sources such as photovoltaics, aligning with the "dual-carbon" goals of the communications industry. The lightweight BIM collaboration platform and workflow engine based on WebGL enable all parties, including designers, construction workers, supervisors, and operators, to annotate, approve, and track data in real time on a unified 3D model. This automates task flow and greatly reduces communication costs and change errors.

[0046] Based on the same inventive concept, embodiments of this specification also provide an electronic device.

[0047] The following describes embodiments of the electronic device of the present invention, which can be considered as specific implementations of the methods and apparatus embodiments of the present invention described above. Details described in the embodiments of the electronic device of the present invention should be considered as supplements to the methods or apparatus embodiments described above; details not disclosed in the embodiments of the electronic device of the present invention can be implemented with reference to the methods or apparatus embodiments described above.

[0048] Figure 3This is a schematic diagram of an electronic device provided as an embodiment of this specification. Refer to the following... Figure 3 The electronic device 300 according to this embodiment of the present invention will be described. Figure 3 The electronic device 300 shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0049] like Figure 3 As shown, the electronic device 300 is presented in the form of a general-purpose computing device. The components of the electronic device 300 may include, but are not limited to: at least one processing unit 310, at least one storage unit 320, a bus 330 connecting different system components (including storage unit 320 and processing unit 310), a display unit 340, etc.

[0050] The storage unit stores program code that can be executed by the processing unit 310, causing the processing unit 310 to perform the steps described in the processing method section of this specification according to various exemplary embodiments of the present invention. For example, the processing unit 310 can perform actions such as... Figure 1 The steps are shown.

[0051] The storage unit 320 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 3201 and / or a cache storage unit 3202, and may further include a read-only memory unit (ROM) 3203.

[0052] The storage unit 320 may also include a program / utility 3204 having a set (at least one) program module 3205, such program module 3205 including but not limited to: an operating system, one or more application programs, other program modules and program data, each or some combination of these examples may include an implementation of a network environment.

[0053] Bus 330 can represent one or more of several types of bus structures, including a memory cell bus or memory cell controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any of the various bus structures.

[0054] Electronic device 300 can also communicate with one or more external devices 400 (e.g., keyboard, pointing device, Bluetooth device, etc.), and with one or more devices that enable viewers to interact with electronic device 300, and / or with any device that enables electronic device 300 to communicate with one or more other computing devices (e.g., router, modem, etc.). This communication can be performed via input / output (I / O) interface 350. Furthermore, electronic device 300 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 360. Network adapter 360 can communicate with other modules of electronic device 300 via bus 330. It should be understood that, although... Figure 3 As not shown, other hardware and / or software modules may be used in conjunction with electronic device 300, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0055] Through the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described in this invention can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this invention can be embodied in the form of a software product, which can be stored in a computer-readable storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, or network device, etc.) to execute the method described above according to this invention. When the computer program is executed by a data processing device, it enables the computer-readable medium to implement the method described above, i.e.: as... Figure 1 The method shown.

[0056] Figure 4 This is a schematic diagram of a computer-readable medium provided for embodiments of this specification.

[0057] accomplish Figure 1The computer program of the method shown can be stored on one or more computer-readable media. A computer-readable medium can be a readable signal medium or a readable storage medium. A readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof.

[0058] The computer-readable storage medium may include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The readable storage medium may also be any readable medium other than a readable storage medium, capable of transmitting, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the readable storage medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.

[0059] Program code for performing the operations of this invention can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can execute entirely on the audience's computing device, partially on the audience's device, as a standalone software package, partially on the audience's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the audience's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0060] In summary, the present invention can be implemented in hardware, or as software modules running on one or more processors, or a combination thereof. Those skilled in the art will understand that in practice, general-purpose data processing devices such as microprocessors or digital signal processors (DSPs) can be used to implement some or all of the functions of some or all of the components according to the embodiments of the present invention. The present invention can also be implemented as a device or apparatus program (e.g., a computer program and computer program product) for performing part or all of the methods described herein. Such programs implementing the present invention can be stored on a computer-readable medium or can be in the form of one or more signals. Such signals can be downloaded from an Internet website, provided on a carrier signal, or provided in any other form.

[0061] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the present invention is not inherently related to any specific computer, virtual device, or electronic device, and various general-purpose devices can also implement the present invention. The above descriptions are merely specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0062] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0063] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A multi-target cooperative system for communication towers, characterized in that, include: The multi-objective optimization engine is configured to receive design requirements and initial environmental constraints, update the environmental model based on IoT feedback, and automatically generate a set of Pareto optimal design solutions through multi-agent reinforcement learning. The 3D visualization design module is connected to the multi-objective optimization engine and is used to automatically generate the tower BIM model based on the design parameter vector output by the engine and perform structural finite element verification, and feed the verification results back to the multi-objective optimization engine. The Internet of Things (IoT) integration module includes a sensor array and an edge computing gateway deployed on the tower, used to collect structural response and environmental data, extract structural degradation characteristics, and upload them to the cloud; The cloud-based collaborative platform connects with the multi-objective optimization engine and the 3D visualization design module to display multi-dimensional indicators of design schemes, receive selected instructions and drive the generation of construction models and drawings, and realize multi-role collaboration through the workflow engine. The energy storage system optimization module is connected to the Internet of Things integration module. It optimizes the energy storage capacity and charging and discharging strategy based on historical load and new energy data, and feeds back the obtained backup power constraints to the multi-objective optimization engine. The multi-objective optimization engine is also used to periodically obtain structural degradation characteristics and extreme environment data from the IoT integration module, update its internal load and resistance probability model through Bayesian inference, and regenerate design scheme correction suggestions based on the updated model.

2. The communication tower multi-target cooperative system as described in claim 1, characterized in that, The multi-objective optimization engine specifically includes: The state space definition unit has state variables including tower height, moment of inertia of main material section, material yield strength, basic wind pressure return period coefficient, seismic intensity, and peak backup power input by the energy storage system optimization module. An action space definition unit, whose action is a continuous adjustment of the state variable; The multi-agent reinforcement learning network employs a multi-agent deep deterministic policy gradient algorithm. Each agent learns policies corresponding to cost objectives, structural reliability objectives, and carbon emission objectives, and shares information among agents through a centralized Critic network. Preference encoding networks are used to map user-specified preference vectors or randomly sampled preference vectors to weight coefficients in a multi-objective reward function; The reward function is defined as R = -(α1·C) norm +α2·(1-RI norm )+α3·E norm ), where C norm To normalize the total lifecycle cost, RI norm E is a normalized structural reliability index. norm α1, α2, α3 are the normalized life-cycle carbon emission equivalents, and [α1, α2, α3] are the weight coefficients output by the preference coding network.

3. The communication tower multi-target cooperative system as described in claim 2, characterized in that, After receiving the design parameter vector output by the multi-objective optimization engine, the 3D visualization design module drives the parametric BIM component library to automatically reconstruct the 3D model of the tower assembly, extract the node coordinates, section properties and load combination information of the members, and automatically call the cloud-based finite element analysis microservice. If the stress ratio of all members and the displacement at the top of the tower meet the preset specifications, the response index will be returned to the multi-objective optimization engine as part of the positive reward. If any indicator exceeds the limit, the constraint violation degree will be used as a negative reward feedback, and the design parameter vector will be marked as infeasible.

4. The communication tower multi-target cooperative system as described in claim 1, characterized in that, The edge computing gateway in the IoT integration module is specifically configured as follows: Fast Fourier transform is performed on the accelerometer signal to extract the first three natural frequencies, and the percentage of drift from the initial reference frequency is calculated as the overall stiffness degradation characteristic of the structure. Rainfall counting statistics are performed on the strain sensor signals to accumulate the daily fatigue damage increment of each member; By combining daily temperature and humidity data measured by temperature and humidity sensors, the equivalent corrosion depth of steel is updated based on a corrosion electrochemical model. The frequency drift percentage, daily fatigue damage increment, and equivalent corrosion depth are used as a structural degradation feature set, compressed, and uploaded to a cloud database.

5. The communication tower multi-target cooperative system as described in claim 4, characterized in that, The multi-objective optimization engine periodically extracts the structural degradation feature set and corresponding meteorological extreme value data from the cloud database, and performs Bayesian inference: Using the wind pressure return period coefficient and material resistance partial factor adopted in the initial design as the prior distribution, and the bearing capacity attenuation corresponding to the measured extreme wind speed sequence and equivalent corrosion depth as the observation likelihood, the posterior distribution is updated to form a posterior load and resistance probability model that reflects the actual microenvironment of the tower.

6. The communication tower multi-target cooperative system as described in claim 1, characterized in that, The cloud-based collaborative platform includes a lightweight BIM browser based on WebGL and a role-based access control workflow engine. The browser obtains a Pareto solution set and its corresponding cost-reliability-carbon emission three-dimensional scatter plot from the multi-objective optimization engine for decision-makers to select. Upon receiving the selected scheme signal, the corresponding design parameter vector is sent to the 3D visualization design module to generate a construction-level model and engineering drawings. The preset workflow is then initiated, automatically pushing tasks and documents to construction, supervision, and operation and maintenance roles, and supporting online annotation and approval status tracking.

7. A multi-target cooperative method for communication towers, characterized in that, include: Acquire GIS geographic information, historical meteorological extreme value data and operator equipment load requirements for the target tower construction site, and initialize the state space of the multi-objective optimization engine by combining the initial backup power constraints fed back by the energy storage optimization module. Multi-agent reinforcement learning is performed. In each step of the exploration, the generated design parameter vector is transmitted to the 3D visualization design module to automatically generate a BIM model and perform finite element verification. The verification result is converted into a reward signal and returned until convergence to obtain a set of Pareto optimal design schemes. The performance indicators of each solution are displayed in a visual manner through a cloud-based collaborative platform. After receiving the selected solution, a construction-level model and engineering drawings are generated and distributed to relevant roles through a workflow engine. After the tower is completed and enters service, the edge computing gateway continuously collects structural vibration, strain and environmental data, extracts structural deterioration feature sets, and synchronously drives the energy storage system optimization module to generate energy storage optimization operation strategies. Every preset period, the multi-objective optimization engine retrieves the accumulated measured extreme value environmental data and structural degradation feature set from the cloud database, updates and corrects the internal load and resistance probability model based on Bayesian, and re-executes multi-objective optimization using the updated model, outputting design parameter correction suggestions for subsequent towers in the region, forming a data-driven adaptive design closed loop.

8. The multi-target cooperative method for communication towers as described in claim 7, characterized in that, The multi-agent reinforcement learning process involves transmitting the generated design parameter vectors to the 3D visualization design module at each exploration step to automatically generate a BIM model and perform finite element verification. The verification results are converted into reward signals and returned until a set of Pareto optimal design schemes is obtained, including: If the finite element verification of the 3D visualization design module finds that the stress ratio or displacement of the member exceeds the limit, a constraint violation degree is generated and the current exploration step is terminated immediately as a high negative reward. Only when the verification is passed are the three target values ​​of total life cost, structural reliability and carbon emissions calculated and combined into a comprehensive reward-driven learning.

9. An electronic device, wherein, The electronic device includes: A processor; and a memory storing computer-executable instructions, which, when executed, cause the processor to perform the method according to any one of claims 7-8.

10. A computer-readable storage medium, wherein, The computer-readable storage medium stores one or more programs that, when executed by a processor, implement the method of any one of claims 7-8.

Images