业界声音｜余兵：人工智能技术有望推动能源企业在智慧管理、生态协作等领域实现质的突破

答：技术突破对于可再生能源在未来占据能源主体地位具有决定性影响，我认为以下几方面值得关注：

一是风电、太阳能发电技术。最近十几年风电、光伏发电技术在提效率、降成本上取得革命性突破，目前已在全球大多数地区实现平价上网。随着新技术、新材料、新工艺持续涌现，未来风光发电效率更高、成本更低、应用领域更广泛，成为能源革命的中坚力量。

二是海洋能发电技术。随着海洋工程技术、材料科学、海洋环境科学等多学科交叉融合，更高效、更稳定的潮汐能发电以及更大规模、更低成本的波浪能发电等技术有望取得突破，成为可再生能源领域的重要发展方向。

三是基于储能、氢能、数智化的电力安全稳定技术。可再生能源的间歇性、波动性和季节性问题给电网稳定运行带来严峻挑战。一方面，低成本的长时大容量储能技术亟待突破，主要涉及固态电池、液流电池等新型电池技术，超导储能、飞轮储能等新型物理储能技术，以及氢能与新能源耦合的柔性制氢技术、长距离管道输氢、规模化储氢技术等等。另一方面，电力智能调度技术亟待突破，需要提升水风光功率预测、因地制宜多能多效电源互补等技术，加强能源生产、转化、储备各环节的数据联通技术，推动源网荷储一体化、多能源协同运行、虚拟电厂等先进能源技术和人工智能、大数据等先进信息技术深度融合，促进可再生电力消纳，确保可再生能源电力上网的稳定性与安全性，支撑新型电力系统发展。

问：在这段时间里，大语言模型浪潮掀起，以ChatGPT为代表的生成式人工智能工具席卷全球，将人工智能的可能性带入了一个新的维度。你认为AI与产业结合应用的突破将发生在哪些领域？

答：以ChatGPT为代表的生成式人工智能技术，未来有望推动能源企业在智慧管理、生态协作、智能生产等领域实现质的突破。

一是智慧管理领域。以国家能源集团为例，依托现有资源优势，利用人工智能技术，正在建设煤电路港航、煤电油气化、产运销储用的行业大模型，推动自动化、信息化、智能化、网络化、模型化，牵引实现战略运营、调度、管理、监督的赋能重塑。此外，通过数据画像、智能评估等技术，可以为人力资源画像、经营分析等场景提供智能化辅助决策。

二是生态协作领域。生成式大模型正在加速构建产品3D模型、虚拟主播、虚拟货场，通过与AR、VR等新技术结合，可以增强用户参与度和内容互动性，实现多感官交互的沉浸式体验，有望推广到能源产业链供应链环节，实现供需双方在足不出户的情况下，即可获得能源商品实时状态，推动产业链供应链向全景、共振智能化转变。

三是智能生产领域。在设备智能管理方面，推广应用设备智能设计制造、综合状态评估、高度智能化人机交互等技术。在自动驾驶方面，通过数据、图像增强和标注技术，优化自动驾驶认知决策模型，实现露天煤矿无人矿卡、井工矿无人胶轮车等广泛应用。在安全培训教育方面，深度融合元宇宙、虚拟人和数字培训脚本，模拟真实厂矿生产环境，增强生产作业人员风险防范和应急处突能力。

A: Technological breakthroughs play a decisive role for renewable energy to dominate future energy supplies.

Firstly, wind/solar power technology. The last decade or so saw revolutionary breakthroughs that improved the efficiency and reduced the cost of wind power and PV. Thus far, grid parity has been achieved for wind farms in most parts of the world. As new materials and new processes continued to emerge, wind power will become more efficient and cost-effective, with wider applications, making it the backbone of the energy revolution.

Secondly, marine energy technology. The integration of marine engineering, materials science, and marine environmental science is expected to lead to breakthroughs in tidal energy, which is a more efficient and reliable source of energy, and wave energy characterized by large-scale, cost-effective generation of energy.

Thirdly, power security and stability technology based on energy storage, hydrogen energy and digital intelligence. Intermittency, volatility and seasonality of renewable energy pose serious challenges to the stable operation of power grids. It is simply imperative to make breakthroughs in low-cost, long-term and large-capacity energy storage technologies, which primarily involve flexible hydrogen production, long-distance pipeline hydrogen transportation, and large-scale hydrogen storage technology. On the other hand, we also need to break new grounds in intelligent power dispatching, which involves improving hydro-wind-solar power prediction and multi-energy power complementation adapted to local conditions, further developing data connectivity related to energy production, conversion and storage, integrating generation-grid-load-storage integration, multi-energy coordination and Virtual Power Plants and AI and big data computing, and facilitating on-grid renewable power supply.

Q: We have recently witnessed the rise of large language models. Generative AI tools, such as ChatGPT, have taken the world by storm. In which areas do you expect to see breakthroughs in the integration of AI and industrial applications?

A: Generative AI will motivate energy companies to make meaningful breakthroughs in intelligent management, ecological collaboration, and intelligent production.

First, smart management. China Energy, for example, leverages existing resources and AI to develop an industry-wide large language model (LLM) for coal, power, road, port and shipping businesses. Furthermore, through data portrait and automated assessment, the company provides intelligent decisionmaking assistance for HR profiling, business analysis, etc.

Second, ecological cooperation. Generative LLM is conducive to the modeling of 3D products, virtual anchors and virtual warehouses. With the assistance of AR and VR, we offer an immersive experience of multi-sensory interactions, enhanced user engagement and content interactivity, which are expected to be applied to energy industry chains and supply chains enabling suppliers and their clients to keep track of the status of energy products in real time.

Third, intelligent production. In terms of intelligent equipment management, China Energy encourages the application of technologies including intelligent equipment design and manufacturing, comprehensive status assessment, and highly intelligent human-computer interaction. As regards autonomous driving, China Energy uses data/image enhancement and annotation technology to optimize cognitive decision-making model of autonomous driving and increase the application of unmanned mining trucks in open-pit coal mines and unmanned rubber-tired vehicles. As for safety training, metauniverse, virtual avatars and digital training scripts are deeply integrated to simulate the real factory and mine production environment to help operating personnel enhance risk prevention and emergency response capabilities.