A digital and intelligent system applied to collaborative management and control of participating units in hydropower engineering construction

Abstract

The application relates to the technical field of management and control systems, in particular to a digital system applied to collaborative management and control of participating units in hydropower engineering construction, which comprises a project data management module, covers whole life cycle management of hydropower engineering construction, extracts multi-source information data from various subsystems, constructs a project view and a project management information base, a project risk analysis module, constructs an intelligent risk analysis model to perform risk analysis, generates dynamic evaluation results, graded early warning notices and auxiliary decision schemes, a project collaborative management module, processes the graded early warning notices and the auxiliary decision schemes by using a multilayer perception machine model, generates a finalized business result and a management process record, and a project document management module, performs compliance automatic checking and processing on electronic files needing to be archived, dynamically updates and optimizes the project management information base. Through the closed-loop management and control system, the collaborative management and control efficiency of participating units in hydropower engineering construction is improved.

Description

Technical Field

[0001] This invention relates to the field of management and control system technology, specifically to a digital and intelligent system for collaborative management and control of participating units in hydropower engineering construction. Background Technology

[0002] Currently, the administrative management and operational control activities of smart hydropower projects generally rely on multiple information systems with fragmented functions. For example, core management functions such as project planning, resource scheduling, and cost control are typically executed in independent specialized project software; while engineering design documents, administrative approval processes, and daily office collaboration are scattered across various document management and general office automation tools. Each participating unit, including design, construction, supervision, and equipment suppliers, often operates its own internal control system. This results in the difficulty of achieving real-time and effective integration and sharing of engineering data, such as BIM 3D building information models, GIS geographic information data, and emerging IoT real-time sensor data, with core project management information and administrative process data. This widespread information silo phenomenon makes it difficult for data to form a unified view required to support comprehensive management decisions. Although some general-purpose industrial internet platform technologies exist on the market, they often have significant shortcomings in deeply adapting to the complex management scenarios unique to hydropower projects, supporting collaborative control models with efficient multi-stakeholder participation, and realizing intelligent management applications based on domain knowledge.

[0003] Therefore, the main shortcoming of existing technologies in supporting smart hydropower projects to achieve their core management and control objectives is the lack of an industrial internet platform to enable data-driven, intelligent, and collaborative management and control of hydropower projects. This situation leads to unclear overall project situation awareness, delayed risk prevention and control, poor multi-party collaboration, and low management decision-making efficiency, which seriously restricts the improvement of the project's intelligence level and comprehensive benefits.

[0004] To address this, a digital and intelligent system for collaborative management of participating units in hydropower engineering construction is proposed. Summary of the Invention

[0005] The purpose of this invention is to provide a digital and intelligent system for collaborative management of participating units in hydropower engineering construction. Through a closed-loop management system, it improves the efficiency of collaborative management in smart hydropower projects. The system includes: a project data management module, which extracts multi-source information data from the entire lifecycle of the smart hydropower project and its various subsystems, and constructs a project view and project management information database; a project risk analysis module, which constructs an intelligent risk analysis model to perform risk analysis, generating dynamic assessment results, tiered early warning notifications, and auxiliary decision-making schemes; a project collaborative management module, which uses a multilayer perceptron model to process tiered early warning notifications and auxiliary decision-making schemes, generating finalized business results and management process records; and a project document management module, which optimizes the project management information database through knowledge extraction technology.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A digital and intelligent system for collaborative management of participating units in hydropower engineering construction, characterized by comprising: The project data management module extracts multi-source information data from the entire life cycle of hydropower smart projects and various subsystems of smart projects, and builds project views and project management information databases based on the multi-source information data; The project risk analysis module constructs an intelligent risk analysis model to perform risk analysis on the project management information database and project view, and generates dynamic assessment results, hierarchical early warning notifications and auxiliary decision-making solutions. The project collaboration management module uses a multilayer perceptron model to process hierarchical early warning notifications and auxiliary decision-making schemes, calls the project management information database and project view as execution support, and generates finalized business results and management process records. The project document management module receives finalized business deliverables and management process records, performs automated compliance verification and processing on data that needs to be archived, and uses knowledge extraction technology to feed management experience, risk cases and project information back into the project management information database, dynamically updating and optimizing the project management information database.

[0007] Preferably, the process of building the project view includes: This involves integrating data from various subsystems of the hydropower smart project, including project plans and actual progress data, cost accounting and contract performance data, and quality inspection and safety control process data. It also integrates various management process data and performs multi-dimensional correlation and status mapping with the engineering entity information contained in the BIM 3D building information model corresponding to the hydropower smart project, forming a project view that includes time dimension, spatial dimension, project status, management progress, and key indicators.

[0008] Preferably, the project management information database is a professional knowledge graph for hydropower engineering. The knowledge graph describes and stores the relationships between engineering elements, management processes, resource allocation, contract terms, risk factors, and industry standards in a semantic manner, and is used for advanced querying and reasoning.

[0009] Preferably, the risk analysis process performed by the intelligent risk analysis model includes: Based on the integrated contract budget data, terms and conditions data, real-time progress and payment settlement data, and engineering change and claim data obtained from the project management information database and project view, a hybrid analysis method that integrates the experience rules of hydropower industry contract management experts and machine learning algorithms is used to conduct correlation analysis on multi-dimensional performance data and assign hierarchical dynamic weights to risk factors; dynamic assessment results of contract performance risk status under different dimensions, hierarchical early warning notifications, and auxiliary decision-making schemes are generated.

[0010] Preferably, the decision support scheme specifically manifests as outputting decision reference information for a specific risk that has been identified, including multiple alternative countermeasures, as well as expected cost data, expected benefit assessment data, and comprehensive ranking consideration indicators corresponding to each alternative countermeasure.

[0011] Preferably, the process of using a multilayer perceptron model to process tiered early warning notifications and auxiliary decision-making schemes includes: based on the level and type of the tiered early warning notification and the historical similar risk handling cases and management procedures for hydropower projects contained in the project management information database, using a multilayer perceptron model trained on successful historical project cases to dynamically adjust the preset risk response process and administrative approval process. The adjustment includes optimizing the process participation nodes, approval levels, processing time limits, and required supporting material lists; when determining the process participation nodes, further calling the responsibilities and personnel skill profile data of participating units in the project management information database for intelligent recommendation of processing personnel.

[0012] Preferably, the process of generating finalized business deliverables and management process records includes, when handling collaborative project management activities, engineering change approvals, and important scheme reviews, invoking electronic signature services at key approval nodes to ensure the validity of the digital signature of the finalized business deliverables and the compliance of the approval process.

[0013] Preferably, the process of automatically verifying and processing the data to be archived includes: intelligently constructing a scenario-adaptive compliance judgment benchmark based on the latest effective laws, standards and management systems stored in the project management information database and dynamically associated with the current document business context, combined with the engineering entity status and risk level information obtained from the project management information database; using natural language processing technology to conduct in-depth semantic compliance review of the document content, while generating prompt information containing specific positioning, interpretability analysis and correction guidance, and feeding back the compliance issues identified in the verification and effective handling experience to the project management information database.

[0014] Preferably, the process of feeding back into the project management information database includes: using natural language processing and knowledge engineering technologies to actively identify and extract structured knowledge elements that reflect effective management practices, key risk control points, and decision optimization logic in hydropower engineering scenarios from finalized business deliverables that have passed compliance verification, management process records, and related compliance guidelines and correction suggestions; linking the extracted structured knowledge elements with existing engineering entities, business process nodes, and legal and standard clauses in the project management information database in multiple dimensions and making them explicit knowledge expressions, forming contextualized experiential knowledge units that can be understood and reasoned by computers, and feeding them back into the project management information database.

[0015] Preferably, the dynamically updated and optimized project management information base includes the ability to learn and analyze contextualized experience knowledge units, dynamically update and optimize the recommended execution paths for early warning threshold settings in the intelligent risk analysis model and the management processes of the project collaborative management module.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. The project data management module efficiently extracts and integrates multi-source information data from the entire lifecycle of hydropower projects and various subsystems of smart projects, generating a unified project view and a structured project management information database. This effectively breaks down data silos in the traditional management model, transforming previously scattered and difficult-to-use data into an understandable, analyzable, and decision-supporting management information database. This ensures that all subsequent digital and intelligent collaborative management activities can be based on comprehensive, accurate, and real-time information, truly realizing data-driven management.

[0017] 2. Through the intelligent risk analysis model, it is possible to conduct in-depth analysis and intelligent judgment on the complex engineering and management data that has been integrated from the project management information database and project view. This significantly improves the situation of lagging risk perception and insufficient pre-control capability caused by insufficient data utilization in existing technologies. It can conduct more effective dynamic quantitative assessment and accurate early warning of specific potential risks in hydropower projects, and effectively support the intelligent management and scientific decision-making of risks in digital collaborative management.

[0018] 3. Through the project collaboration management module, based on the level and type of early warning notifications and historical similar risk handling cases and management procedures, machine learning models can be used to dynamically adjust and optimize the preset risk response process and administrative approval process. This effectively solves the technical problems of low collaboration efficiency and slow response caused by complex cross-organizational and cross-departmental collaborative workflows and poor information transmission in hydropower projects. By integrating electronic signature services at key approval nodes, the effectiveness of collaborative results and compliance of processes are ensured. This enables efficient, intelligent and standardized operation of cross-organizational and cross-departmental collaborative project management activities in hydropower projects, and powerfully promotes the implementation of digital and intelligent collaborative management and control.

[0019] 4. Through the project document management module, knowledge extraction technology is used to deeply process the finalized business deliverables, management process records, and related compliance guidelines and correction suggestions generated by the system. This effectively solves the technical challenges in hydropower engineering, such as the difficulty in accumulating and reusing tacit knowledge of project management experience and risk cases, leading to insufficient organizational learning capabilities, and the cumbersome and error-prone nature of traditional document compliance management. Automated verification improves the efficiency and quality of compliance management, and proactive extraction of structured knowledge elements optimizes the project management information database, enhancing the knowledge accumulation, intelligent application, and continuous self-evolution capabilities of digital and intelligent collaborative management in hydropower engineering. Attached Figure Description

[0020] Figure 1 This invention provides a schematic diagram of a digital and intelligent system for collaborative management and control of participating units in hydropower engineering construction. Figure 2 This is a schematic diagram of the dynamic weighted model provided in an embodiment of the present invention; Figure 3 This is a flowchart illustrating the project collaboration management module provided in an embodiment of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Please see Figures 1 to 3 This invention provides a digital intelligent management and control system for smart hydropower projects, the technical solution of which is as follows: Company A is engaged in the construction, operation, maintenance, and management of water conservancy and hydropower projects. To improve management efficiency and intelligent control in complex hydropower projects, Company A has deployed a digital and intelligent system provided by this invention for collaborative management of participating units in hydropower project construction on its industrial internet platform. The specific system flowchart is as follows: Figure 1 As shown.

[0023] The project data management module extracts multi-source information data from the entire life cycle of hydropower smart projects and various subsystems of smart projects, and builds project views and project management information databases based on the multi-source information data; For Company A's smart hydropower engineering project, the project data management module obtains the project's 3D BIM model from the BIM design system through a pre-defined interface adapter. This model contains information on the main structural components and key attributes. Simultaneously, it extracts the project's WBS (Work Breakdown Structure), planned and actual progress data from the project schedule management software. Furthermore, it retrieves construction log summaries, major resource inputs, and feedback on important issues from the daily construction management system. Additionally, the module accesses core design drawings and important technical disclosure documents for this phase from the document management system. The extracted data, such as BIM model data, P6 plan data, construction log text, and drawing files, undergoes processing using the module's built-in ETL tool to correct non-standard records, ensure all progress data uses a consistent date format and percentage representation, identify process names, completion statuses, and problem descriptions from the construction log text, and then aggregates the processed data into the cloud platform data storage layer deployed by this system.

[0024] Furthermore, after data integration, the project data management module aims to generate an intuitive and information-rich project view. The core of this view is to correlate and visualize the status of actual progress data reported by the construction unit and key status information from the construction logs with corresponding engineering entities in the BIM 3D model (e.g., a specific pouring unit of a dam, a load-bearing column of a plant, or an excavation section of a tunnel). For example, the current completion percentage of a WBS task can be directly reflected on its corresponding BIM component, clearly displayed in the 3D model using different colors and labels. Company A's management personnel can access this project view through the unified front-end interface provided by the system's cloud platform. This project view, based on the 3D model for navigation, integrates key progress indicators (such as comparison between planned and actual progress) and the current status of major processes, thereby providing a comprehensive and intuitive understanding of the progress of the core construction phases of the project. To achieve precise association between construction progress data, log status, and other management information and the corresponding engineering entities in the BIM 3D model, two complementary methods are mainly used. The first method is to pre-set the corresponding WBS code in the attribute parameters of the components during the BIM model design phase. After importing data, the system can automatically establish a link between P6 planned tasks and BIM components based on this. The second method, as a supplement to the preset coding method or when the preset coding is unavailable, provides a human-computer interactive mapping tool that allows users to select P6 tasks and corresponding BIM components on the interface, assisting the system in completing the matching and assignment between the two.

[0025] By integrating multi-source information data such as BIM, progress reports, and construction logs, and performing data cleaning, unification, and key information extraction, a visualized project view deeply integrated with the BIM model is generated, effectively breaking down the data silos common in hydropower projects. This significantly improves the accuracy, consistency, and real-time sharing of project data, providing a crucial data foundation and unified cognitive platform for subsequent data-driven decision-making, intelligent analysis, and efficient multi-party collaborative management. It is an important prerequisite for achieving data-driven, intelligent, and collaborative management of hydropower projects.

[0026] Furthermore, based on the project view and multi-source information data, a project management information database is constructed. In Company A's hydropower smart engineering project, this database is specifically implemented as a hydropower engineering knowledge graph. This knowledge graph aims to deeply integrate and link key engineering elements and management information from the core construction phase of the project in a semantic manner, providing structured and networked data support for subsequent intelligent analysis and collaborative work.

[0027] Specifically, the construction of the knowledge graph first relies on the clearly defined correspondence between BIM components and WBS tasks in the project view, and core data extracted from various source systems and processed by ETL. For example, key engineering structural components in the BIM model (such as specific pouring units for the dam, main beams and columns of the plant, and foundations for key equipment), WBS task items in the project schedule management software (including their planned attributes and logical relationships), metadata information from core design drawings and technical disclosure documents, and key process progress descriptions and problem records extracted from construction log summaries are defined as core entities in the knowledge graph. Subsequently, the project data management module establishes associations with clear business semantics between these core entities. For example, a WBS task node is connected to its corresponding specific engineering entity node in the 3D view through precise association with BIM components; the WBS task node can also be connected to the construction log summary node that records its daily progress through construction record relationships. If a specific problem is recorded in the construction log, the problem record node will establish an association with the relevant WBS task and BIM component to report the problem. The metadata nodes of a core design drawing are connected to the corresponding WBS tasks and BIM component nodes described or guided by it through design reference relationships. The establishment of these relationships primarily relies on the data mapping logic already formed in the project view, and standardization methods such as matching the WBS codes in P6 with the preset WBS codes in the BIM component attributes ensure the accuracy of the associations. The constructed knowledge graph is stored and managed on a cloud platform using graph database technology for efficient querying and expansion.

[0028] The project management information repository transforms core project information, which was originally scattered across different systems and views, into an interconnected knowledge network. This provides a crucial data foundation for subsequent project risk analysis modules to identify related risks and for project collaboration management modules to drive context-aware workflows.

[0029] The project risk analysis module constructs an intelligent risk analysis model to perform risk analysis on the project management information database and project view, and generates dynamic assessment results, hierarchical early warning notifications and auxiliary decision-making solutions. The project risk analysis module retrieves and integrates multi-dimensional data related to contract performance from the project management information repository and the real-time updated project view. This data specifically includes contract budget data, detailed terms and conditions, real-time progress data for each relevant WBS task, payment and settlement data, and engineering change and claim records generated during project execution. Because this data has been semantically linked and structured in a knowledge graph, its consistency and integrability are guaranteed.

[0030] Table 1. Example of multi-dimensional data related to contract performance

[0031] Furthermore, after acquiring and integrating multi-dimensional data related to the performance of specific contracts from the knowledge graph, the intelligent risk analysis model immediately initiates its risk analysis and assessment process. The key to this process lies in using a hybrid analysis method that integrates the experience rules of hydropower industry contract management experts with machine learning algorithms. This method utilizes the semantic association capabilities of the knowledge graph to conduct penetrating correlation analysis on complex performance data, and dynamically assigns context-aware weights to key risk factors based on the analysis results.

[0032] Specifically, the hybrid analysis method first uses an expert rule engine to apply pre-defined rules, derived from the experience of senior hydropower contract management experts, to structured entities and their relationships in the knowledge graph. These rules can quickly identify contract performance statuses that conform to known high-risk patterns. For example, key payment milestones and their preceding project progress show a significant mismatch in the knowledge graph, and the associated contract clauses contain explicit penalties. Simultaneously, to compensate for the shortcomings of explicit rules in dealing with complex, unknown, and progressive risks, the process introduces a machine learning algorithm such as Gradient Boosting Decision Tree (GBDT). The GBDT model is pre-trained using historical project data within the platform (structured and accessible in the knowledge graph). It can perform deep learning and pattern recognition on a broader, multi-dimensional set of performance data features extracted from the current project knowledge graph (e.g., the frequency and average impact period of engineering changes in similar historical projects, the correlation between textual features of claims and their processing results, patterns of key resource price fluctuations and cost overruns, and potential collaboration tensions reflected in metadata of multi-party communication records). In this way, the model can uncover complex nonlinear relationships, more subtle risk precursors, or emerging risk evolution trends that are difficult to capture through explicit rules. It can predict the probability of occurrence and the potential impact of specific contractual risks (such as major claim risks and key supplier default risks). The outputs of the GBDT model, such as risk probability scores, anomaly pattern alerts, or key risk factor rankings, will serve as a powerful supplement to expert rule judgments, an important basis for cross-validation, and optimization of expert rule parameters. Finally, the outputs of the expert rule engine and the GBDT model are combined to obtain preliminary risk results.

[0033] Furthermore, after obtaining preliminary risk results through the expert rule engine and GBDT machine learning model, the intelligent risk analysis model performs a deep correlation analysis based on the professional knowledge graph of hydropower engineering. This step is not an isolated, superficial assessment of the identified risk points, but rather fully utilizes the global, multi-dimensional semantic network unique to the knowledge graph to achieve a penetrating insight into the inherent connections and complex transmission mechanisms between risks. This is significantly different from traditional risk assessment methods based on decentralized data sources and limited correlation models. Specifically, for each risk signal indicated by the preliminary risk identification results (e.g., a high-probability risk event output by the GBDT model, or a contract clause warning triggered by expert rules), the intelligent risk analysis model proactively traverses and infers along pre-built (structured connections formed based on the inherent business logic of hydropower engineering and mapped data) and dynamically discovered semantic links in the knowledge graph, starting from the core entity associated with the signal (such as specific WBS tasks, contract clauses, BIM components, and supplier entities). These semantic links span multiple dimensions, including project contract texts, schedule networks, cost breakdown structures, BIM 3D models, technical specifications, resource allocation, and interaction records of all participating parties.

[0034] Furthermore, the dynamic weighting of risk factors is achieved through a dynamic weighting model, the flowchart of which is shown below. Figure 2As shown, the model receives three types of data. The first is a quantitative indicator of the preliminary risk results output by the expert rule engine and GBDT model, such as the initial risk level, probability score, or confidence level of the anomaly pattern. The second is the output of deep correlation analysis, regarding the importance of risk factors in the knowledge graph, such as whether the factor is located at the intersection of multiple important influence paths, and the breadth of the factor's potential influence range. The third is the context adjustment coefficient value, which is transformed from the current project context information (such as project stage, recently occurring major events, and their impact levels) obtained in real time from the knowledge graph. These three types of data are standardized. Then, for each type of standardized input data, the model pre-sets a corresponding lookup table. The lookup table transforms each standardized input value into its specific contribution score to the final weight. For example, the higher the correlation keyness score, the greater the keyness contribution score obtained through the lookup table. First, an adjustment rule is selected based on the contribution score of the initial risk result. The initial risk contribution level is then categorized into three levels: low, medium, and high, using a preset warning threshold. When the initial risk contribution level is determined to be low, the contribution score of the initial risk result is directly converted into a weight according to the rule; this weight is a dynamic weight. When the initial risk contribution level is determined to be medium, the contribution score of the initial risk result is converted into a weight according to the rule, and the contribution score of the correlation output result is converted into an adjustment coefficient. Adding 0.5 to the adjustment coefficient yields a dynamic adjustment coefficient, and multiplying the weight by the dynamic adjustment coefficient yields the dynamic weight. When the initial risk contribution level is determined to be high, ... The contribution scores of the initial risk results are converted into weights according to rules. Then, the correlation analysis output results and scenario adjustment contribution scores are converted into correlation coefficients and scenario coefficients respectively using an exponential function. The correlation coefficients and scenario coefficients are multiplied by the weights to obtain dynamic weights. Finally, all dynamic weights are normalized to ensure that the sum of the coefficients is 1. Finally, they are multiplied by the quantitative risk value of the corresponding risk factor to generate dynamic assessment results and early warning signals representing the performance risk status of the contract at different dimensions from the contract clause level, specific bid section level, overall project level to enterprise operation level. The dynamic assessment results will be output in the form of a report for system operators to view.

[0035] The project risk analysis module leverages knowledge graphs for deep correlation and dynamic weighting, transforming multi-source data into precise, multi-level risk insights and early warnings. This directly enhances data-driven decision-making and intelligent analysis capabilities, providing crucial technical support for achieving intelligent and collaborative management of hydropower projects and overcoming the challenges of traditional management models.

[0036] Furthermore, after the intelligent risk analysis model completes a dynamic assessment of a specific risk and issues an early warning, the project risk analysis module immediately initiates the process of generating auxiliary decision-making solutions. For the specific risk that has been identified and assessed, the module retrieves historical experience in handling similar risk cases based on its characteristics recorded in the hydropower engineering professional knowledge graph (such as risk type, cause, and current project stage), and matches it with the system's standardized response strategy library for various common hydropower engineering risks. Through this retrieval and matching process, the system automatically filters or combines several preliminary alternative response measures.

[0037] Subsequently, for each alternative response, the module further utilizes relevant data stored in the knowledge graph (e.g., project resource unit price, labor hour rate, historical cost and benefit records of similar measures, current project plan parameters, and contractual benchmarks) for quantitative analysis. This analysis generates key reference information needed for decision-making. Specifically, this includes the expected cost data for implementing each alternative; the expected benefit assessment data for reducing or eliminating current risks after implementation; and comprehensive ranking indicators for overall evaluation, such as the expected execution cycle, technical feasibility and implementation complexity, and potential secondary impacts on other aspects of the project (e.g., quality and safety). Finally, the project risk analysis module integrates this structured information, containing multiple alternative responses and their corresponding quantitative assessment data (expected costs, expected benefits, and comprehensive indicators), into a decision support plan for project managers to reference.

[0038] The decision support function automatically generates and quantifies various risk response measures through knowledge graph data, transforming hydropower engineering experience and data into clear and comparable decision references. This enhances the data-driven and intelligent level of risk management decisions, provides a scientific basis for managers to select the optimal strategy, thereby enhancing collaborative management efficiency. It also helps alleviate the problems of subjective decision-making and untimely response in traditional decision-making, and improves the efficiency of data-driven intelligent collaborative management of hydropower projects.

[0039] The project collaboration management module utilizes a multilayer perceptron model to process tiered early warning notifications and auxiliary decision-making schemes. It calls upon the project management information database and project view as execution support, generating finalized business deliverables and management process records. The flowchart of the project collaboration management module is shown below. Figure 3 As shown.

[0040] When the project collaboration management module receives the early warning signal and auxiliary decision-making scheme generated by the project risk analysis module, it first initiates dynamic adjustments to relevant administrative approval processes based on this information, combined with successful historical risk management cases and management procedures related to the hydropower project stored in the knowledge graph. This adjustment process utilizes a multilayer perceptron neural network model with an embedding layer integrated at its input for categorical features to make intelligent decisions. The embedding layer is primarily responsible for converting high-cardinality categorical data extracted from input features, such as risk level codes, risk type identifiers, classification labels of historical similar cases, and specific management procedure indexes, into low-dimensional, dense real-valued vector representations. These feature vectors processed by the embedding layer are then fed along with other numerical input features (such as quantified risk impact assessments and project stage parameters) into multiple fully connected hidden layers of the MLP for nonlinear transformation and higher-order feature learning. The MLP neural network model is pre-trained on a large amount of historical project data. This training data covers the characteristics of different risk events, the specific parameters of the risk response and administrative approval processes used at the time (including participating nodes in the process, approval level settings, actual processing time limits, and lists of supporting materials submitted), and the final execution efficiency and actual effectiveness of risk management. During actual operation, when the project collaborative management module receives new tiered early warning notifications and auxiliary decision-making schemes, the MLP neural network model takes the current risk level and type, the characteristics of historical similar risk management cases retrieved from the knowledge graph, and the key elements of the current management procedures as input. Through the complex nonlinear mapping relationship between the learned inputs and the optimal process parameters, it directly outputs a set of optimal process parameter combinations for this scenario. These optimized parameters are specifically reflected in the dynamic adjustment suggestions for process nodes, approval levels, processing time limits, and required supporting materials lists in the preset process. When determining the specific handlers for each participating node in the process, the system further utilizes the knowledge graph to access information on the responsibilities of each participating unit, the skill profiles of key personnel, and historical workload data, achieving intelligent and more suitable handler recommendations. Table 2 shows an example of intelligent handler recommendation. Through this intelligent handler recommendation mechanism, it is ensured that various collaborative tasks can be quickly and accurately assigned to the most suitable executor, thereby improving overall work efficiency and quality.

[0041] Table 2 Examples of Intelligent Recommendation Processors

[0042] Based on risk levels, historical success cases, and procedures, the nodes, levels, time limits, and materials of risk response and approval processes are dynamically optimized through a multi-layer perceptron model. Combined with knowledge graph-based intelligent recommendation of processors, this achieves data-driven and intelligent algorithm-based process self-adaptation and precise resource matching, significantly improving the efficiency, relevance, and decision-making quality of hydropower projects in risk response and administrative approval.

[0043] When any of the aforementioned collaborative management activities and approval processes reach a predefined key approval node—for example, when final confirmation, approval, and signing are required from key figures such as the project manager, chief engineer, department head, or representative of an external supervision unit—the project collaborative management module will automatically invoke a designated third-party electronic signature service platform within the platform. At this node, the electronic document to be approved (e.g., a PDF file of an engineering change order that has completed all prior reviews, or a PDF version of an important construction plan to be signed) will be pushed to the electronic signature service stage. The designated approver will, upon system prompts, use their personal or organizational digital certificate, supported by the electronic signature service platform, to perform a legally binding electronic signature on the electronic document. After completing the signing action, the electronic signature service platform will return the finalized business deliverable with the approver's valid digital signature to the collaborative management module of this system. Furthermore, the module will verify the validity of the signature. When the signature is verified as valid, it will save the signature information and signing time as an immutable management process record in the approval process.

[0044] The integrated electronic signature service fully digitizes key approval processes in hydropower projects, ensuring the legal validity of electronic signatures on all finalized business deliverables and the compliance and traceability of the approval process. This not only significantly improves the efficiency and convenience of multi-party collaborative approvals, but more importantly, it provides highly reliable and tamper-proof digital credentials and process data for a digital and intelligent collaborative management system. This strengthens the foundation of data-driven decision-making and is a crucial guarantee for achieving closed-loop, standardized, and intelligent management throughout the entire process.

[0045] The project document management module receives finalized business deliverables and management process records, performs automated compliance verification and processing on data that needs to be archived, and uses knowledge extraction technology to feed management experience, risk cases and project information back into the project management information database, dynamically updating and optimizing the project management information database.

[0046] When the project document management module receives the finalized business deliverables and related management process records generated by the project collaboration management module that need to be archived, in addition to performing the four-fold checks of authenticity, completeness, usability, and security, the project document management module will also initiate a more in-depth and intelligent automated content compliance verification and processing process. Specifically, the process involves first, intelligently constructing personalized and context-adaptive compliance judgment benchmarks. For the document to be verified, the module queries the project management information database. It retrieves not only the latest and most effective regulations, industry standards, internal management systems, and quality requirements dynamically linked to the document's business context, but also the current status information of engineering entities related to the solution content and the latest risk level assessment results obtained from the knowledge graph. Based on the document's specific business context, it intelligently selects the most directly relevant clauses and requirements from a general regulatory and standard knowledge base. Furthermore, based on the real-time status data and risk assessment results of the associated engineering entities, and considering the specific risk level and complexity of the engineering characteristics of the document's associated engineering entities, it dynamically adjusts the stringency of the relevant compliance clauses. It transforms the general statements in regulations and standards into specific, quantifiable judgment standards for the current specific context. It also adaptively determines the list of required supporting materials and their required level of detail. Through this refined adaptation process, the system ultimately constructs a context-adaptive compliance judgment benchmark for the document to be verified.

[0047] After obtaining a highly customized and context-adaptive compliance assessment benchmark for the current document to be verified, the project document management module immediately uses Natural Language Processing (NLP) technology to perform a deep semantic compliance review on the actual content of the document against the aforementioned dynamically generated benchmark. The specific review process is as follows: the system's built-in NLP engine performs word segmentation, part-of-speech tagging, and syntactic dependency analysis preprocessing on the input document text. Subsequently, the engine comprehensively uses multiple NLP methods for analysis. Through Named Entity Recognition (NER) technology, it accurately extracts key engineering elements mentioned in the document (such as equipment names, material specifications, geological condition descriptions), quantified technical parameters, and information such as the set responsible parties and time limits. Through Relation Extraction (RE) technology, it understands the logical relationships, the order of action execution, and conditional dependencies between these identified entities. Furthermore, it uses semantic similarity calculation and textual implication judgment to accurately compare and verify the actual content, extracted key information, and logical relationships in the document with each specific, contextualized, and quantified inspection standard in the compliance assessment benchmark generated in the previous step at the semantic level. This process focuses on understanding the true intent and completeness of the text, rather than simply matching keywords superficially, in order to determine whether the document content substantially and comprehensively meets all the compliance requirements stipulated by the dynamic benchmark.

[0048] Furthermore, if the NLP engine identifies any deviations from the compliance judgment benchmark, omissions of information, internal logical contradictions, or statements with potential risks during this in-depth semantic compliance review process, the system will automatically identify these situations as potential risk points. For each identified risk point, the system simultaneously generates structured prompts containing three core components. First, specific location information clearly indicates the exact location of the risk point within the original document, down to the chapter, paragraph, sentence, chart, or footnote number. Second, interpretability analysis, based on the system's judgment logic and relevant compliance benchmarks, briefly explains why the content was deemed potentially non-compliant. For example, it specifies which requirement in the compliance benchmark it violates, which clause of which regulation or standard it conflicts with, and whether it fails to adequately address the specific risk scenarios of the current project. Finally, correction guidance and suggestions provide preliminary, actionable modification suggestions, such as indicating which missing information should be added, which standard clauses should be referenced for correction, and suggestions for clarifying and improving certain expressions. This helps document creators and reviewers quickly locate problems, understand the causes, and efficiently complete compliance modifications to the document content.

[0049] By dynamically utilizing regulations, standards, real-time project status, and risk levels within a knowledge graph, personalized and context-adaptive judgment criteria are constructed, and NLP technology is applied for deep semantic review. This enables a shift in document compliance checks from manual reliance and static standards to intelligent automation and dynamic adaptation, significantly improving document quality and compliance review efficiency. It not only provides data-driven intelligent assurance for the standardized management of hydropower projects and reduces compliance risks, but also provides a high-quality and reliable information foundation for the effective operation of the overall digital and intelligent collaborative management system.

[0050] The system performs in-depth analysis of the text content of these compliance-verified documents and records. In terms of NLP technology applications, a pre-trained Transformer model is used for sequence labeling to achieve high-precision named entity recognition, thereby accurately extracting key engineering objects, participant names, technical terms, and risk descriptions mentioned in the documents. Furthermore, a bidirectional long short-term memory network model is used to identify and understand the complex logical relationships, execution order of actions, and conditional dependencies among the identified entities.

[0051] Furthermore, after acquiring complex logical relationships, the execution order of actions, and conditional dependencies using NLP technology, the system employs knowledge engineering methods for subsequent structured extraction, semantic alignment, and verification. The system references a pre-constructed, detailed hydropower engineering domain ontology. This ontology defines various concepts, entity attributes, standard relationship types, and several knowledge graph patterns customized for specific knowledge types within hydropower engineering. SPARQL queries are used to assist in verifying and reasoning about existing information in the knowledge graph. Based on the hydropower engineering domain ontology and knowledge graph patterns, the system proactively identifies and integrates standardized, structured knowledge elements that clearly reflect effective management practices, key risk control points, and decision optimization logic in specific hydropower engineering scenarios, taking into account the fragmented information extracted by NLP. For example, from a risk management report, the system can structurally extract the core knowledge of how, for risk type X, measures Z are adopted in project phase Y, ultimately reducing the risk level from A to B.

[0052] Furthermore, these knowledge elements, extracted and structured using NLP and KE methods, will be precisely integrated into and enriched into the project management information base through multi-dimensional semantic links and explicit knowledge expression mechanisms. Specifically, the system will establish links with clear semantic types that conform to the established patterns of the knowledge graph and the ontology definition of hydropower engineering for each newly extracted structured knowledge element and related engineering entities, corresponding business process nodes, and referenced legal and standard clauses in the knowledge graph. For example, when creating a new experience knowledge node, for each previously extracted structured knowledge element, it will connect it with related entities through relation types such as "applied to," "solved," and "optimized." A core instance node representing this experience, i.e., a contextualized experience knowledge unit, will be created in the knowledge graph. Within this newly created unit instance, the key components of the experience will be stored in a structured attribute form according to the predefined knowledge graph pattern, including a clear context description, the management practices and technical measures adopted, and the final implementation effects and lessons learned. In addition, to ensure the contextual nature of this experiential knowledge unit and its availability in the knowledge graph, the system will use the association information obtained in the NLP and KE stages to establish rich, semantically clear, multi-dimensional links between this newly created contextualized experiential knowledge unit instance and a large number of existing engineering entities, process nodes, and specification clauses in the knowledge graph that are closely related to its context.

[0053] By leveraging natural language processing and knowledge engineering technologies, valuable, implicit management experience, risk control points, and decision-making logic from project documents and records are intelligently transformed into structured, contextualized experiential knowledge units within a knowledge graph that are computable, reasonable, and reusable. This achieves automated accumulation, explicit representation, and deep association of project knowledge, significantly enhancing the value density and usability of data. It not only provides high-quality experiential data and a knowledge foundation for hydropower projects to achieve more accurate data-driven decision-making, continuously optimize risk analysis models (such as weights and thresholds), and intelligently recommend collaborative process paths, but also serves as a key enabling link in building learning organizations, achieving iterative upgrades in management intelligence, and deepening overall digital and intelligent collaborative management capabilities.

[0054] After continuously enriching the project management information database with contextualized experiential knowledge units, the system further executes a learning and analysis process for these accumulated units. This learning and analysis process aims to extract patterns and knowledge from historical experience that can be used to improve system effectiveness. Specifically, the system comprehensively utilizes specific data mining and machine learning techniques to conduct in-depth analysis of the structured information contained in these contextualized experiential knowledge units. The system employs the Apriori association analysis algorithm to process the feature data in the experiential units. By calculating the support and confidence of different combinations of situations, measures, and outcomes in historical records, it discovers hidden, statistically significant, and strongly correlated rules. These rules help reveal effective management practice patterns and the evolution and response patterns of specific risks. The system uses case-based reasoning (CBR) logic combined with support vector machines (SVM) for effectiveness evaluation and pattern recognition. Utilizing the similarity matching mechanism of CBR, it retrieves a set of historical cases highly similar to the current analytical focus in terms of project context from a large number of experiential units. Subsequently, the management measures and technical solutions features and their corresponding actual effects in these retrieved similar cases are used as training to build, train, and optimize the SVM supervised learning model. These trained and optimized models can then predictively assess the expected effectiveness of different management measures or technical solutions in specific contexts, quantify and rank the effectiveness of historical decisions, thereby helping the system identify the optimal practice patterns and potential process optimization opportunities under specific conditions.

[0055] Furthermore, the system conducts in-depth statistical learning and pattern analysis on the accuracy of historical early warning events and the actual severity of their corresponding risk events. The system uses receiver operation characteristic curve analysis to dynamically adjust the quantitative thresholds for triggering early warnings at different risk levels. This optimization aims to continuously improve the accuracy of risk warnings, reduce false alarms and missed alarms, and ensure that early warning signals can most effectively guide management actions. Regarding the optimization of the project collaborative management module, the system is also empowered by learning and analyzing the contextualized experience knowledge units accumulated in the knowledge graph. These experience units record detailed actual execution data of various historical risk response processes and administrative approval processes, including the specific triggering context, the actual participating nodes and approval levels in the process, the processing time of each step, the validity and completeness of the submitted supporting materials list, and the final execution results. By using the process mining Alpha algorithm, the system automatically identifies proven efficient process patterns, common bottlenecks, and best practice elements from this rich historical process execution data. These process optimization insights, extracted through process mining and case-based reasoning (CBR) analysis, are then used to optimize recommended execution paths for specific management processes. This optimization is implemented by directly adjusting the internal model parameters and the bias and weight configuration of the recommendation logic of the dynamically adjusting machine learning model (Multilayer Perceptron MLP). When similar project scenarios or specific types of management processes need to be initiated in the future, based on a knowledge base that continuously learns and evolves from successful and failed experiences, it can recommend process execution plans that include more reasonable combinations of process participant roles, more streamlined and efficient approval hierarchy structures, more accurate estimated processing time limits, and more targeted supporting material requirements, thereby achieving continuous iterative improvement in collaborative work efficiency and effectiveness.

[0056] By intelligently learning and analyzing historical experience units, the system achieves dynamic self-optimization of risk warning thresholds and collaborative process paths. This closed-loop learning mechanism enables the system to continuously improve the accuracy of warnings and process efficiency, significantly enhancing the intelligence and adaptability of risk response and collaborative operations in hydropower projects, and providing core impetus for deepening digital and intelligent collaborative management and control.

[0057] By constructing an integrated BIM project view and professional knowledge graph, data silos are broken down, enabling deep information fusion and sharing, laying a solid foundation for data-driven operations. Its hybrid intelligent risk analysis model and decision support functions significantly improve risk foresight and the intelligence level of decision-making. Simultaneously, machine learning-driven dynamic collaborative processes, knowledge graph-based intelligent personnel recommendations, and electronic signature applications greatly enhance multi-party collaboration efficiency and the standardization and agility of management and control. Furthermore, the system's intelligent document compliance verification and knowledge extraction mechanism, combined with the learning and analysis of contextualized experiential knowledge units, achieves dynamic self-optimization of risk warning thresholds and collaborative process recommendation paths. Ultimately, this forms a data-driven, intelligently evolving, and highly collaborative management and control closed loop, comprehensively improving the digital and intelligent management and control efficiency of hydropower projects and effectively addressing the bottlenecks of traditional models.

[0058] Example 2: Company B is a hydropower investment and operation company responsible for the management and operation of several large and medium-sized hydropower stations within the basin, including some aging stations that have been in operation for many years and newly built smart stations. With aging generating units, difficulties in passing on the experience of technical personnel, and increasingly stringent safety and environmental regulations, Company B faces increasingly severe challenges in ensuring the safe, efficient, and economical operation of its power stations. To address these challenges, Company B has introduced the digital and intelligent system provided by this invention, which is applied to the collaborative management of participating units in hydropower engineering construction.

[0059] In Company B's operation and maintenance management, the core task of the project data management module has shifted to the integration of data from the entire operation and maintenance lifecycle of each power station and a comprehensive understanding of asset health status. Through interface adapters, the module efficiently integrates multi-source information data from various power stations, including real-time operating parameters from the SCADA system, vibration and oil temperature status data from the online monitoring system (CMS), work orders and historical maintenance records from the equipment maintenance management system (CMMS), manual inspection and periodic test reports, equipment ledgers and technical specifications from the BIM model and digital asset information database.

[0060] After cleaning, transformation, and semantic processing, this data was used to construct an operational knowledge graph centered on hydropower station equipment assets. Simultaneously, the system generated a multi-dimensional, visualized view of the power station's health status and operational performance, geared towards operation and maintenance management. This view, based on a digital twin model of the power station, dynamically displays the main operating parameters, health scores, warning statuses, maintenance history, and related operation and maintenance document links for each key piece of equipment in real time, providing Company B's operation and maintenance management team with an intuitive and comprehensive grasp of the entire power station group's asset status.

[0061] Multi-dimensional data related to equipment health, operating conditions, maintenance history, and external environment are retrieved from the operational knowledge graph and real-time project view. Subsequently, a hybrid analysis method integrating expert experience rules for hydropower station equipment operation and maintenance with machine learning algorithms is applied for processing. For example, the expert rule base contains a large amount of diagnostic logic for typical fault modes of specific equipment (such as hydro-generator units, main transformers, and high-voltage switchgear). For instance, if the temperature of a phase winding of a generator continuously exceeds the limit by X degrees Celsius, accompanied by an abnormal increase in vibration value Y, it is judged as a Class Z fault risk, with a high level. Simultaneously, machine learning algorithms learn patterns from a large amount of historical operation and monitoring data to identify gradual equipment performance degradation trends or early fault characteristics that are difficult to define through simple rules, and predict the probability of specific fault occurrences or remaining useful life (RUL).

[0062] Deep correlation analysis helps to analyze how changes in the inflow of an upstream reservoir affect the operating conditions of the generating unit, and thus how these changes relate to the wear rate of key components of the unit. Ultimately, it assesses the impact on the unit's overhaul cycle and potential power generation loss risks. Dynamic weighting of risk factors ensures that the contribution of various risk factors to the overall risk assessment can be adaptively adjusted under different operating modes, seasons, and equipment aging stages.

[0063] Ultimately, this module generates multi-level quantitative assessment data and tiered early warning signals for Company B's operational risk status, ranging from key equipment component level, single unit level, entire hydropower station level, to the operational level of the cascade hydropower group in the river basin. When the system detects a deterioration in the health status of a piece of equipment, a predicted failure probability exceeding a threshold, or a significant risk indicated by an operational indicator, it immediately issues an early warning notification to Company B's central control center, power station operation and maintenance department, and relevant management personnel, and initiates the generation of auxiliary decision-making schemes. These auxiliary decision-making schemes provide various alternative response measures for specific equipment failure warnings or operational risks, such as "immediately arrange inspections or precise diagnostics," "adjust operating modes to reduce risks," "plan maintenance in advance and prepare spare parts," and "activate emergency response plans." These measures are accompanied by comprehensive assessment information on the expected costs, expected benefits, urgency, and resource requirements of each measure, assisting Company B in making optimal operation and maintenance decisions. After receiving the above operational risk warnings and auxiliary decision-making schemes, the project collaboration management module will drive Company B's internal and necessary external collaborative units to efficiently execute operation and maintenance-related collaborative workflows.

[0064] Upon receiving an alert that "the vibration of a critical unit's bearing exceeds the standard, predicting a major failure risk within one week" and a recommended decision-making plan of "immediately arranging precise diagnosis and preparing an emergency repair plan," the system dynamically adjusts the preset "equipment failure diagnosis and emergency repair process" based on the alert's level and type, and referring to Company B's historical successful handling cases and current equipment maintenance management procedures stored in the knowledge graph. This adjustment may include automatically compressing certain non-core approval steps and prioritizing the process based on the urgency of the risk; intelligently recommending the optimal diagnostic execution team and necessary external experts based on the current workload, personnel skill profiles, and availability of the equipment to be diagnosed for each maintenance team recorded in the knowledge graph; and dynamically generating a preliminary list of required supporting materials to assist in the diagnosis based on the initial characteristics of the failure. During these maintenance collaboration processes, at key approval or confirmation nodes, the system calls and integrates Company B's unified electronic signature service platform. Relevant personnel complete electronic signatures using their digital certificates through the system, ensuring the compliance, authenticity, and traceability of maintenance decisions and execution records. All collaborative work process records, approval opinions, electronic signature information, and final operation and maintenance reports will be fully preserved as finalized business deliverables and management process records.

[0065] The project document management module is responsible for receiving and managing various finalized business deliverables and process records generated during the aforementioned collaborative operation and maintenance process. The module first performs intelligent, automated content compliance verification on these documents. Based on the latest laws, regulations, industry standards, Company B's internal operation and maintenance management systems, and equipment operation SOPs stored in the knowledge graph related to power plant operation and maintenance, and combined with the real-time status and risk level information of the currently operated equipment, the system dynamically constructs personalized, context-adaptive compliance judgment benchmarks. Subsequently, NLP technology is used to conduct a deep semantic compliance review of the document content, identifying potential violations such as non-standard data entry and missing safety measure descriptions, and generating prompts containing specific locations, explanations, and correction suggestions.

[0066] Furthermore, the project document management module utilizes knowledge extraction technology to proactively identify and extract structured knowledge elements from these compliance-verified operation and maintenance documents. These elements reflect effective management practices, key fault diagnosis logic, efficient emergency response plans, and equipment performance optimization techniques in the hydropower station operation and maintenance field. These extracted operation and maintenance experience knowledge elements are transformed into contextualized experience knowledge units, deeply linked to corresponding equipment entities, fault codes, operation and maintenance process nodes, and relevant procedures and clauses in the knowledge graph, thereby continuously enriching and optimizing Company B's operation and maintenance professional knowledge graph.

[0067] Finally, these continuously accumulating contextualized experience knowledge units will optimize the system's intelligent capabilities. For example, through learning and analyzing a large number of successful and unsuccessful fault diagnosis and repair cases, the system can dynamically optimize the parameters and feature weights of the machine learning model used for equipment fault prediction in the intelligent risk analysis module, improving the accuracy of fault warnings. Simultaneously, it can also optimize the setting of warning thresholds to better adapt to the characteristics of different equipment at different stages of their lifecycle. For the project collaboration management module, by analyzing the execution data of historical efficient operation and maintenance processes, the system can optimize recommended execution paths and resource allocation suggestions for specific operation and maintenance tasks, thereby enabling Company B's overall operation and maintenance management system to form a virtuous cycle of continuous learning, self-evolution, and efficiency improvement based on data and knowledge. At the same time, all important operation and maintenance documents and records will be generated into electronic archives as required, supporting Company B's long-term knowledge management and compliance audit needs.

[0068] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A digital and intelligent management and control system applied to smart hydropower projects, characterized in that, include: The project data management module extracts multi-source information data from the entire life cycle of hydropower smart projects and various subsystems of smart projects, and builds project views and project management information databases based on the multi-source information data; The construction of the project view includes: integrating engineering plan and actual progress data, cost accounting and contract performance data, and quality inspection and safety control process data related to hydropower smart engineering from various smart engineering subsystems; and performing multi-dimensional association and status mapping between various management process data and the engineering entity information contained in the BIM 3D building information model corresponding to the hydropower smart engineering to form a project view containing time dimension, spatial dimension, engineering status, management progress, and key indicators; the project management information database is specifically a hydropower engineering professional knowledge graph, which describes and stores the relationship between engineering elements, management processes, resource allocation, contract terms, risk factors, and industry standards in a semantic manner for advanced query and reasoning. The project risk analysis module constructs an intelligent risk analysis model to perform risk analysis on the project management information database and project view, generating dynamic assessment results, tiered early warning notifications, and auxiliary decision-making schemes. The intelligent risk analysis model's risk analysis process includes: based on integrated contract budget data, clause agreement data, real-time progress and payment settlement data, and engineering change and claim data obtained from the project management information database and project view, a hybrid analysis method integrating expert experience rules in hydropower industry contract management and machine learning algorithms is used to perform correlation analysis on multi-dimensional performance data and assign hierarchical dynamic weights to risk factors; dynamic assessment results, tiered early warning notifications, and auxiliary decision-making schemes are generated for the contract performance risk status under different dimensions; the hybrid analysis method combines the output results of the expert rule engine with the output results of the GBDT model to obtain preliminary risk results; for each risk signal indicated by the preliminary risk results, starting from the core entity associated with the risk signal, multi-path and multi-depth traversal and reasoning are performed along the semantic links pre-constructed in the hydropower engineering professional knowledge graph and dynamically discovered during the analysis. The core entities include WBS tasks, contract clauses, BIM components, and supplier entities; The hierarchical dynamic weighting is implemented through a dynamic weighting model. The dynamic weighting model receives the quantitative indicators of the preliminary risk results, the importance and potential impact range of the risk factors output by the deep correlation analysis in the knowledge graph, and the situation adjustment coefficient value converted from the current project context information obtained in real time from the knowledge graph. Based on the contribution score of the initial risk results and the preset warning threshold, the initial risk contribution level is divided into three levels: low, medium, and high. Different dynamic weighting adjustment methods are used for the low, medium, and high levels. The obtained dynamic weights are normalized and multiplied by the quantitative risk value of the corresponding risk factor to generate a dynamic assessment result and warning signal of the performance risk status. The project collaboration management module uses a multilayer perceptron model to process hierarchical early warning notifications and auxiliary decision-making schemes, calls the project management information database and project view as execution support, and generates finalized business results and management process records. The project document management module receives finalized business deliverables and management process records, and performs automated compliance verification and processing on data to be archived. Through knowledge extraction technology, it feeds management experience, risk cases, and project information back into the project management information database, dynamically updating and optimizing the database. This dynamic updating and optimization includes the ability to learn and analyze contextualized experience knowledge units, dynamically updating and optimizing the early warning threshold settings used in the intelligent risk analysis model and the recommended execution paths for the management processes of the project collaborative management module. Specifically, through statistical learning and pattern analysis of the accuracy of historical early warning events and the actual severity levels of corresponding risk events, it uses receiver operation characteristic curve analysis to dynamically adjust the quantitative thresholds for triggering early warnings at different risk levels.

2. The intelligent management and control system for hydropower smart engineering as described in claim 1, characterized in that, The process of building the project view includes: This involves integrating data from various subsystems of the hydropower smart project, including project plans and actual progress data, cost accounting and contract performance data, and quality inspection and safety control process data. It also integrates various management process data and performs multi-dimensional correlation and status mapping with the engineering entity information contained in the BIM 3D building information model corresponding to the hydropower smart project, forming a project view that includes time dimension, spatial dimension, project status, management progress, and key indicators.

3. The intelligent management and control system for hydropower smart engineering as described in claim 1, characterized in that, Specifically, the decision support scheme is characterized by outputting decision reference information for specific risks that have been identified, including multiple alternative countermeasures, as well as expected cost data, expected benefit assessment data, and comprehensive ranking indicators corresponding to each alternative countermeasure.

4. The intelligent management and control system for hydropower smart engineering as described in claim 1, characterized in that, The process of using a multilayer perceptron model to process graded early warning notifications and auxiliary decision-making schemes includes: based on the grade and type of the graded early warning notification and the historical similar risk handling cases and management procedures for hydropower projects contained in the project management information database, using a multilayer perceptron model trained on successful historical project cases to dynamically adjust the preset risk response process and administrative approval process. The adjustment includes optimizing the process participation nodes, approval levels, processing time limits, and the list of required supporting materials; when determining the process participation nodes, further calling the responsibilities and personnel skill profile data of the participating units in the project management information database for intelligent recommendation of processing personnel.

5. The intelligent management and control system for hydropower smart engineering as described in claim 1, characterized in that, The process of generating finalized business deliverables and recording management processes includes, when handling collaborative project management activities, engineering change approvals, and important scheme reviews, invoking electronic signature services at key approval nodes to ensure the validity of the digital signatures of the finalized business deliverables and the compliance of the approval process.

6. The intelligent management and control system for hydropower smart engineering as described in claim 1, characterized in that, The automated compliance verification and processing process for the data to be archived includes: intelligently constructing a scenario-adaptive compliance judgment benchmark based on the latest effective laws, standards, and management systems stored in the project management information database and dynamically associated with the current document business context, combined with the engineering entity status and risk level information obtained from the project management information database; using natural language processing technology to conduct in-depth semantic compliance review of the document content, while generating prompt information containing specific positioning, interpretability analysis, and correction guidance; and feeding back the compliance issues identified in the verification and effective handling experience to the project management information database.

7. The intelligent management and control system for hydropower smart engineering as described in claim 1, characterized in that, The process of feeding back into the project management information database includes: using natural language processing and knowledge engineering technologies, actively identifying and extracting structured knowledge elements that reflect effective management practices, key risk control points, and decision optimization logic in hydropower engineering scenarios from finalized business deliverables, management process records, and related compliance guidelines and correction suggestions that have passed compliance verification; linking the extracted structured knowledge elements with existing engineering entities, business process nodes, and legal and standard clauses in the project management information database in multiple dimensions and making them explicit knowledge expressions, forming contextualized experiential knowledge units that can be understood and reasoned by computers, and feeding them back into the project management information database.

Images