Background and Objectives

Background

The concept of digital twin (DT) is defined by ISO/IEC 30173:2023 as “a data model that establishes and manages connections between the physical and digital worlds through software and digital technologies,” DT enables real-time monitoring, predictive analytics, and scenario simulation worldwide.

The Strategic Research and Innovation Agenda (SRIA) by the EERA Joint Programme on Hydropower advocates smart operation, predictive maintenance, and DT integration to enhance plant availability. EU-funded initiatives like iAMP-Hydro, D-HYDROFLEX, and Rehydro further underscore this focus, leveraging DT-driven decision support systems (DSS) for vibration analysis, sediment management, and grid optimization. Meanwhile, Gouvães Pumped-Storage Plant in Portugal employs DT for operational optimization, aligning with Europe’s renewable energy transition goals.

In the U.S., the Department of Energy’s Digital Twin for Hydropower Systems – Open Platform Framework (DTHS-OPF), led by Oak Ridge and Pacific Northwest National Laboratories, targets aging infrastructure modernization through modular, data-driven solutions. General Electric’s three-tier “data + models + applications” DT framework exemplifies industry-driven innovation.

China has achieved significant accomplishments in the application of digitalization for large-scale hydropower projects in recent years. The China Three Gorges Corporation, a key player in this field, has developed an intelligent application for the DT of hydraulic turbine generator units. This application delivers intelligent guarantees for the units’ efficient and reliable operation via six core functions: condition monitoring, fault diagnosis, health assessment, digital twin, and digital maintenance.

Additionally, by adopting the combined approach of “high-precision computational fluid dynamics (CFD) simulation model + machine learning”, the prediction speed of the flow field in the whole hydraulic turbine flow passage can be increased to the second level—effectively addressing the limitations of traditional CFD simulations in real-time monitoring and abnormal condition diagnosis.

Similarly, the Huangbai River Digital Twin System employs DT to optimize flood management, resource allocation, and infrastructure safety, demonstrating tangible operational efficiencies.

Collectively, these initiatives highlight digital twin technology’s pivotal role in advancing smart hydropower systems, driving resilience, sustainability, and cross-sector collaboration worldwide.

Introduction 

Digital Twin (DT) technology has emerged as a transformative solution, enabling virtual representations of physical hydropower units for enhanced performance, predictive maintenance, and decision-making. Based on the earlier questionnaire results and research experience, this research task aims to achieve:

• Visual monitoring: Develop a highly realistic 3D interactive visualization interface that displays real-time parameters and
• Prediction and warning: Predict equipment failures and performance degradation, and issue early warnings.
• Operation optimization: Optimize operation strategies to enhance power generation efficiency and other comprehensive
• Intelligent maintenance: Reduce costs and improve efficiency through intelligent planning and resource scheduling for
• Knowledge sharing: Create a platform for direct queries of specific indexes and solutions from past similar

With the task objectives listed above, four major research topics have been established to fulfill these goals:

• Monitoring and perception of hydropower units and its data collection and management
• Model construction related to the entire process of DTs development for hydropower units
• Building and deployment of DTs for hydropower units and their intelligent application development
• Retrofitting of out-of-date equipment and facilities of aging power stations for DTs deployment

The study on monitoring and perception technology for hydropower units explores advanced technical means that go beyond conventional online monitoring systems. The goal is to achieve more extensive and multi-dimensional perception and collection of both static (e.g., geometric features and physical properties) and dynamic (e.g., pressure, temperature, and flow rate) data of the units, thereby supporting the subsequent building and operation of the DTs of the units. For instance, utilizing IoT- based sensors, vibration analysis, and acoustic emission monitoring, the research assesses real-time mechanical and hydraulic performance of turbines, generators, and other structures. Fiber-optic sensing enables precise strain and temperature detection, while LiDAR and computer vision technologies track surface deformations and sediment accumulation.

The study on data transmission and storage technology for DT systems of hydropower units aims to explore the applicability of the high-speed, low-latency, high-performance, and high-secure data transmission protocols (e.g., Fibre Channel Protocol and 5G technology) and their corresponding devices, as well as the application of more advanced database technologies (e.g., distributed file storage (DFS), NoSQL database, NewSQL database, and cloud storage) for storing the massive volumes of multi-source heterogeneous data. This study will support the further processing, management, and analysis of DTs data.

The data processing and integration for DT systems of hydropower units aims to perform necessary preprocessing and integration of data from the monitoring and perception system of the units, as well as other relevant systems (e.g., the dispatching system and maintenance management system). This process lays a solid foundation for the development and application of data analysis models. For instance, data cleaning techniques such as outlier detection, noise filtering, and missing-value imputation are employed to ensure high-quality datasets from heterogeneous sensors (e.g., vibration, pressure, and temperature). Meanwhile, various data fusion methods, including random methods and artificial intelligence methods, can be applied to effectively synthesize, filter, correlate, and integrate multi- source data.

The models are crucial components in the building and application of DTs for hydropower units. These models can be categorized into three dimensions: 3D geometric model, physics and behavior models, and rule model. 3D geometric model describes the hydropower unit in terms of its shape, size, tolerance, and structural relation with appropriate data structures, which are suitable for computer information conversion and processing. Additionally, other related information of the key components, such as material information and assembly information, is also included in the geometric model. Physics and behavior models are advanced simulation models built upon the geometric model. Through sophisticated mathematical algorithms and simulation technologies, such as physics-based simulation and data-driven simulation, they predict the responses and changes of hydropower units under different conditions, thereby describing the dynamic behavior and operating status of the hydropower units. Rule model describes the rules extracted from historical data, expert knowledge, and predefined logic. These rules equip DTs with ability to reason, judge, evaluate, optimize, and predict, thus supporting the development of intelligent applications based on DTs.

The building and deployment of DTs for hydropower units aims to integrate, map, visualize, and synchronize various data and models, ultimately completing the construction, validation and application of the DTs. The technologies used include model lightweighting technology, 3D rendering and visualization technology, and scalable architecture design, among others. For instance, Leveraging computer graphics engines and scientific visualization tools, the research develops high-fidelity 3D models that integrate CFD results, structural stress distributions, and real-time sensor data into dynamic visual outputs. Virtual Reality (VR) and Augmented Reality (AR) platforms enable operators to interact with spatially accurate renderings of turbine flows.

Based on the deployed DTs of hydropower units, technical study will be conducted to develop various intelligent application services, such as intelligent warning and diagnosis, health status assessment, and intelligent maintenance, thereby achieving advanced monitoring, diagnosis, prognosis, maintenance, etc. For instance, machine learning-based anomaly detection can identify turbine wear or structural cracks, thereby reducing downtime. Virtual control rooms can simulate emergency responses, providing training without physical risks.

The retrofitting of aging hydropower infrastructure with DT technology enables legacy plants to achieve modern operational efficiency and predictive capabilities through non-invasive IoT sensor deployment, edge-computing gateways, and hybrid physics-AI modeling. By instrumenting vintage turbines with vibration, strain, and hydroacoustic sensors, real-time data is fed into adaptive digital replicas that compensate for missing historical data through transfer learning and machine learning techniques.

The Task 20 aims to address these goals through several types of research topics over the years. The proposed working plan is described in the next section.