A water conservancy construction management process dynamic optimization system

Abstract

The application discloses a water conservancy project construction management process dynamic optimization system, and belongs to the technical field of water conservancy projects, comprising: a multi-source data acquisition module for acquiring real-time internet of things sensing data, image data, environmental data and plan data of a construction site; and a process knowledge graph construction module for establishing a dynamic correlation model with process nodes, resource allocation, environmental factors and risk events as dimensions. The water conservancy project construction management process dynamic optimization system can continuously capture changes in the environment, resources and engineering state by real-time multi-source data fusion sensing and dynamic updating of the process knowledge graph, and generate an adjustment scheme in real time by using a double-layer optimization decision engine. This changes the management process from a fixed script to intelligent navigation, realizes active adaptation and rapid response to a complex engineering environment, and fundamentally improves the agility and resilience of management.

Description

Technical Field

[0001] This invention belongs to the field of water conservancy engineering technology, specifically relating to a dynamic optimization system for water conservancy engineering construction management process. Background Technology

[0002] Traditional water conservancy project construction management processes typically rely on static, pre-defined construction organization designs and schedules. These plans often exhibit rigidity and lag when faced with complex and ever-changing geological conditions, climate and hydrological variations, resource supply fluctuations, and unforeseen risks. They are difficult to adjust and optimize in a timely manner, leading to project delays, cost overruns, and even quality and safety hazards. Existing project management software often focuses on progress tracking and information recording, lacking intelligent analysis and dynamic decision-making capabilities based on real-time multi-source data, thus failing to achieve online, automated, and closed-loop process optimization.

[0003] Therefore, there is an urgent need for an intelligent management system that can sense the environment, make adaptive decisions, and dynamically optimize the execution process.

[0004] The information disclosed in this background section is intended only to enhance the understanding of the overall background of the invention and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0005] This invention aims to overcome the shortcomings of existing technologies and provide a dynamic optimization system for water conservancy project construction management processes that can respond to internal and external changes in real time and dynamically optimize management processes.

[0006] To achieve the above objectives, a specific embodiment of the present invention provides the following technical solution: A dynamic optimization system for water conservancy project construction management process, comprising: The multi-source data acquisition module is used to acquire IoT sensor data, image data, environmental data and planning data from the construction site in real time. The process knowledge graph construction module is used to establish a dynamic association model based on process nodes, resource allocation, environmental factors, and risk events. The dual-layer optimization decision engine includes a process path optimization unit based on global search and a process parameter adjustment unit based on local learning. The closed-loop execution feedback module is used to push optimization instructions to the field terminal and collect execution feedback data to update the knowledge graph.

[0007] In one or more embodiments of the present invention, the multi-source data acquisition module includes: Sensor arrays deployed in key parts of dams, tunnels, and canals; A drone inspection unit that periodically collects data on terrain changes; Interface unit for accessing regional meteorological and hydrological forecast data.

[0008] In one or more embodiments of the present invention, the sensor array includes a piezometer, a displacement gauge, a stress gauge, and a vibration monitor, and the data from each sensor are preprocessed through an edge computing gateway before being uploaded to the cloud.

[0009] In one or more embodiments of the present invention, the process knowledge graph construction module adopts graph neural network technology to dynamically update the association weights between nodes and can automatically identify critical paths and vulnerable links in the process.

[0010] In one or more embodiments of the present invention, in the two-layer optimization decision engine: The process path optimization unit adopts an improved non-dominated sorting genetic algorithm with the optimization objectives of schedule, cost, quality, and safety. The process parameter adjustment unit uses a deep reinforcement learning model based on the Actor-Critic framework to adjust the construction parameters in real time.

[0011] In one or more embodiments of the present invention, the improved non-dominated sorting genetic algorithm introduces a dynamic constraint processing mechanism, which automatically adds constraints when the monitored data exceeds a threshold.

[0012] In one or more embodiments of the present invention, the closed-loop execution feedback module includes: The mobile terminal command push unit supports multiple command formats, including text, images, and 3D animations. Execution confirmation and deviation reporting unit; The optimization effect evaluation unit is used to quantify and analyze the degree of improvement in performance indicators before and after process optimization.

[0013] In one or more embodiments of the present invention, the system further includes a risk warning triggering unit that automatically activates the emergency plan process reorganization mechanism and assigns it the highest execution priority when a major risk sign is detected.

[0014] In one or more embodiments of the present invention, the system further includes a case self-learning module, which stores each optimization process, execution effect and context as a case for process optimization recommendation in similar scenarios.

[0015] In one or more embodiments of the present invention, the system interfaces with existing engineering management systems via API interfaces, supporting one-click import and comparative analysis of process optimization schemes.

[0016] Compared with existing technologies, the dynamic optimization system for water conservancy project construction management processes of this invention continuously captures changes in the environment, resources, and project status through real-time multi-source data fusion perception and dynamic updates of the process knowledge graph, and generates adjustment plans in real time using a two-layer optimization decision engine. This transforms the management process from a fixed script into intelligent navigation, enabling proactive adaptation and rapid response to complex engineering environments, fundamentally improving the agility and resilience of management.

[0017] Decisions are based on comprehensive real-time data and structured domain knowledge, rather than just personal experience, reducing subjective misjudgments. By integrating external data such as weather forecasts, the system can predictively adjust plans, transforming passive response into proactive planning. The dual-layer optimization engine enables the coordinated optimization of macro-process and micro-process parameters, ensuring that decision recommendations are both globally optimal and practically operable on-site.

[0018] The system of this invention is not only a one-time optimization tool, but also possesses continuous learning capabilities. The closed-loop execution feedback module feeds the execution effect data of optimization instructions back into the knowledge graph and optimization model, enabling the system to learn from historical experience and actual results, continuously revising its correlation weights and decision-making strategies. This means that the longer the system is used and the more cases it accumulates, the more accurate its optimization suggestions become, and the more they align with the actual situation of this project and subsequent projects for the enterprise, achieving continuous iteration and evolution of management capabilities. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram of the overall architecture of a dynamic optimization system for water conservancy project construction management process according to an embodiment of the present invention; Figure 2 This is a schematic diagram illustrating the construction and updating principle of a dynamic process knowledge graph for a dynamic optimization system for water conservancy project construction management in one embodiment of the present invention. Figure 3 This is a flowchart of the two-layer optimization decision engine of a dynamic optimization system for water conservancy project construction management process in one embodiment of the present invention. Detailed Implementation

[0021] To enable those skilled in the art to better understand the technical solutions in this disclosure, the technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments in this disclosure, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this disclosure.

[0022] like Figures 1-3 As shown in this embodiment of the invention, a dynamic optimization system for water conservancy project construction management processes adopts a cloud-edge-device collaborative architecture. The sensing end includes various sensors, drones, GPS devices on construction machinery, and manual terminals deployed at the water conservancy project site. Edge computing nodes are responsible for filtering, compressing, and performing preliminary analysis on high-frequency, raw data. The cloud platform carries the core functions of data storage, intelligent analysis, optimization decision-making, and model training. On-site management personnel and construction workers receive instructions and interact through mobile application terminals or web-based consoles.

[0023] Specifically, the multi-source data acquisition module is the system's "sensory nerves." It includes: IoT sensor data acquisition: Sensor arrays are deployed at designed density in key locations such as dams, slopes, tunnels, and cofferdams. These include: vibrating wire piezometers to monitor pore water pressure, GNSS displacement monitoring stations and inclinometers to monitor surface and deep displacement, rebar gauges and concrete strain gauges to monitor structural stress and strain, and vibration and temperature sensors. These sensors transmit data to an edge computing gateway via LoRa or NB-IoT wireless networks. The gateway incorporates lightweight algorithms, such as outlier removal and 5-minute average calculation, before packaging and uploading the processed data to a cloud-based time-series database.

[0024] Imagery and Spatial Data Acquisition: Industry-grade drones equipped with RTK modules are deployed to automatically patrol the construction area twice a week along preset routes, collecting high-resolution orthophotos and oblique photography data. Simultaneously, key work surfaces such as the pouring surface of the roller-compacted concrete dam are precisely scanned using a 3D laser scanner to generate high-precision point cloud models. This data is then transferred to cloud object storage via FTP.

[0025] Environmental and external data access: Through standardized API interfaces, the system accesses real-time refined weather forecast data for the work area provided by the National Meteorological Administration, as well as water level and flow data from upstream hydrological stations. This data serves as input to the optimization model as environmental constraints.

[0026] Data entry for planning and business: The original construction organization design, schedule, BIM model, resource supply plan and other structured data of the project are entered into the system database through the system configuration interface or file import method to form the initial baseline plan.

[0027] like Figure 2 As shown, the process knowledge graph construction module is the system's "knowledge brain," its core function being to transform unstructured engineering experience into a computable and reasonable structured knowledge network. It includes: The graph pattern layer is defined as follows: First, the core entity types are defined, such as "Process: Foundation excavation, formwork support, concrete pouring, curing", "Resources: Excavator, concrete mixer truck, construction team", "Environmental status: Rainfall level, temperature range, water level elevation", and "Risk events: Slope instability warning, material shortage, equipment failure". Next, the relationships between entities are defined, such as "Process A precedes Process B", "Process C consumes resource R", "Environment E constrains Process D", and "Risk W affects Process F".

[0028] Graph Data Layer Construction: Utilizing natural language processing technology, entities and relationships are extracted from historical engineering documents, construction logs, and supervision reports to form an initial graph. For example, from the text "Due to continuous rainfall, earthwork excavation was suspended for 2 days," the relationship "<rainfall, causing suspension, earthwork excavation>" can be extracted. Simultaneously, real-time collected data is mapped to entity instances or attributes in the graph. For instance, the current "rainfall: 25mm / hour" is instantiated as a "heavy rainfall" environmental entity, and an "adverse impact" relationship is established with the currently ongoing "open-air earthwork excavation" process entity.

[0029] Dynamic Updates and Weight Learning: Each relation edge in the graph is assigned a dynamic weight, representing its influence strength or probability. This weight is learned and updated online using a graph neural network. For example, the system records the actual delay time of the "earthwork excavation" process each time "heavy rainfall" occurs. After multiple events, the GNN model automatically adjusts the weight of the relation "<heavy rainfall, adverse effects, earthwork excavation>" to make it closer to the historical statistical average. Simultaneously, the system can discover new potential associations. For instance, when "cement inventory" is below a threshold and "transportation roads are congested," the risk of "concrete pouring interruption" is easily triggered, thus automatically adding or strengthening such relationships in the graph.

[0030] like Figure 3 As shown, the two-layer optimization decision engine is the "decision core" of the system, employing a two-layer collaborative optimization strategy. It includes: upper layer: Process path optimization unit. This unit runs on a daily or weekly cycle, or is activated when triggered by a major risk event. Its inputs are a snapshot of the current process knowledge graph, real-time project status, remaining tasks, and resource pool. The optimization model employs an improved multi-objective genetic algorithm.

[0031] The construction process is encoded as chromosomes, with genes representing procedures and gene order and position representing logical relationships. The goal is to minimize the total construction period and total cost, while maximizing the overall quality score and safety index. This is a multi-objective optimization problem.

[0032] It is worth noting that, in addition to conventional logical constraints and resource constraints, this invention introduces dynamic environmental constraints. The algorithm reads weather forecasts in real time, and if heavy rainfall is predicted for a certain period in the future, it automatically adds a hard constraint of "must be delayed" to the outdoor work procedures scheduled for that period.

[0033] After the algorithm runs, it will obtain a set of Pareto optimal solutions. The system then combines these with the decision-maker's preferences to select an optimal solution and output a new, optimized construction process network plan for the future.

[0034] Lower layer: Process Parameter Adjustment Unit. This unit performs real-time parameter optimization for specific processes in the upper-level plan. For example, for the "dam concrete compaction" process, its parameters include the number of compaction passes, walking speed, and vibration frequency. This unit uses a deep reinforcement learning model.

[0035] The status includes the current concrete consistency, ambient temperature, paving thickness, and the condition of the underlying concrete. The action involves fine-tuning the parameters of the compaction equipment. The reward is given as a positive or negative bonus based on subsequent tests of compaction density and interlayer bond quality.

[0036] By continuously interacting with the "environment," the model learns which actions to take under what conditions to maximize long-term cumulative rewards, i.e., the final construction quality. When a process begins, the model provides optimal construction parameter suggestions based on real-time status input.

[0037] When the data acquisition module detects that the slope displacement rate exceeds the red warning threshold, the warning triggering unit will immediately interrupt the regular optimization process. It first quickly infers the affected subsequent process chains in the graph, and then calls the process path optimization unit. However, at this time, the weight of the "safety" objective in the optimization objective function is set to infinity, and "emergency slope reinforcement" is forcibly inserted as the new highest priority process, and the process is replanned.

[0038] The closed-loop execution feedback module ensures that optimization decisions are effectively implemented. It includes: The results of the optimized decision engine are automatically converted into executable instructions. For example, process change instructions are generated into illustrated construction task sheets with the changed parts highlighted; parameter adjustment instructions are converted into equipment control instructions or operation guides. These instructions are pushed to the mobile apps of the relevant responsible persons via message queues, requiring confirmation of receipt.

[0039] On-site personnel carry out construction according to instructions and report the start and completion status of the work via an app, or report unexpected problems, such as "insufficient stock of the required steel bar type." Meanwhile, sensors and drones continuously collect post-construction data, such as the structural temperature field after pouring.

[0040] The system compares and optimizes expected results with actual performance. For example, it compares predicted and actual project durations to analyze the reasons for discrepancies. The evaluation results are fed back to the process knowledge graph construction module to update the weights of entity relationships, such as "the effect of a new process in shortening the project duration has been confirmed." Simultaneously, complete "scenario-decision-outcome" cases are stored in a case library for future retrieval and recommendations in similar scenarios.

[0041] It will be apparent to those skilled in the art that this disclosure is not limited to the details of the exemplary embodiments described above, and that this disclosure can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of this disclosure is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within this disclosure. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0042] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A dynamic optimization system for the construction management process of water conservancy projects, characterized in that, include: The multi-source data acquisition module is used to acquire IoT sensor data, image data, environmental data and planning data from the construction site in real time. The process knowledge graph construction module is used to establish a dynamic association model based on process nodes, resource allocation, environmental factors, and risk events. The dual-layer optimization decision engine includes a process path optimization unit based on global search and a process parameter adjustment unit based on local learning. The closed-loop execution feedback module is used to push optimization instructions to the field terminal and collect execution feedback data to update the knowledge graph.

2. The system according to claim 1, characterized in that, The multi-source data acquisition module includes: Sensor arrays deployed in key parts of dams, tunnels, and canals; A drone inspection unit that periodically collects data on terrain changes; Interface unit for accessing regional meteorological and hydrological forecast data.

3. The system according to claim 2, characterized in that, The sensor array includes a piezometer, a displacement gauge, a stress gauge, and a vibration monitor. Data from each sensor is preprocessed through an edge computing gateway and then uploaded to the cloud.

4. The system according to claim 1, characterized in that, The process knowledge graph construction module uses graph neural network technology to dynamically update the association weights between nodes and can automatically identify critical paths and vulnerable links in the process.

5. The system according to claim 1, characterized in that, In the aforementioned two-layer optimization decision engine: The process path optimization unit adopts an improved non-dominated sorting genetic algorithm with the optimization objectives of schedule, cost, quality, and safety. The process parameter adjustment unit uses a deep reinforcement learning model based on the Actor-Critic framework to adjust the construction parameters in real time.

6. The system according to claim 5, characterized in that, The improved non-dominated sorting genetic algorithm introduces a dynamic constraint processing mechanism, which automatically adds constraints when the monitored data exceeds a threshold.

7. The system according to claim 1, characterized in that, The closed-loop execution feedback module includes: The mobile terminal command push unit supports multiple command formats, including text, images, and 3D animations. Execution confirmation and deviation reporting unit; The optimization effect evaluation unit is used to quantify and analyze the degree of improvement in performance indicators before and after process optimization.

8. The system according to claim 7, characterized in that, The system also includes a risk warning triggering unit, which automatically activates the emergency plan process reorganization mechanism and assigns it the highest execution priority when a major risk sign is detected.

9. The system according to claim 1, characterized in that, The system also includes a case self-learning module, which stores each optimization process, execution effect, and context as a case for recommending process optimizations in similar scenarios.

10. The system according to any one of claims 1-9, characterized in that, The system interfaces with existing engineering management systems via API, enabling one-click import and comparative analysis of process optimization solutions.

Images