Artificial Intelligence Enabling Future Optimal Flexible Biogas Production for Net-Zero

人工智能实现未来最佳灵活沼气生产，实现净零排放

• 批准号: EP/Y005600/1
• 负责人: Michael Short
• 金额: $ 183.04万
• 依托单位: University of Surrey
• 依托单位国家: 英国
• 项目类别: Research Grant
• 财政年份: 2023
• 资助国家: 英国
• 起止时间: 2023 至 无数据
• 项目状态: 未结题
• 来源: https://gtr.ukri.org/projects?ref=EP%2FY005600%2F1
• 关键词: Artificial; Intelligence; Enabling; Future; Optimal

Anaerobic digestion (AD) is a technology where microorganisms break down organic matter to produce biogas, thereby generating renewable energy from waste. Biogas can be combusted to produce electricity or purified and used as a substitute for natural gas (NG). Because it provides a carbon-neutral substitute for fossil fuels, while also preventing methane emissions at landfills by processing organic waste, AD is noted as an important part of the UK Net Zero Strategy: Build Back Greener.This project aims to develop artificial intelligence (AI) tools to enable radical efficiency improvements in AD biogas production. Currently, there are about 650 operational AD sites in the UK, which reduce UK greenhouse gas emissions by an estimated 1%. This contribution is meaningful, but modest in comparison to AD's potential. The fundamental roadblock at present is a lack of flexibility. Due to the complexities of predicting how different waste feedstocks and different microbial communities will interact under varying operating conditions, AD biogas producers must minimise risk by purchasing only the highest-quality, consistent feedstock, which may also be seasonal; any errors could result in long and costly downtimes. Thus, available waste streams are vastly under-utilised; feedstock prices are driven up, weakening the economic viability of AD biogas production; and limited feedstocks may need to be transported longer distances, increasing carbon emissions.AI holds crucial promise for the optimisation and future expansion of AD biogas production. As an industry that does not have the central research capabilities of other large energy sectors, it furthermore presents exceptional challenges due to the complexities and inherent uncertainties across interacting chemical, biological, and - if reductions in total life-cycle emissions are to be achieved - environmental systems. The project team therefore unites expertise in AI, process optimisation, systems microbiology, and life-cycle assessment to develop whole-systems decision-making tools informed by detailed sub-system modelling. The outputs will include decision-making tools, specifically: A) a hybrid machine-learning digital twin of the biodigesters, based on novel mechanistic modelling approaches combined with process data from industrial partners and new experimental data from the project; and B) optimisation-based system models of other components of a site, to perform site-wide real-time optimisation through a multi-layer digital twin that includes economic and environmental indicators. By linking the digital twin of the biodigester to feedstock procurement and downstream processes, it will be possible to quickly determine the impact of different feedstocks, their combinations, and their prices on biogas quality, while also tracking quantified environmental impacts across AD value chains in real-time and assessing negative emissions potential in future.Increasing the flexibility of UK AD industry will expand waste markets and lower prices to grow the sector with more capacity, boost profits and productivity, and enhance the overall attractiveness of AD as an investment. Increasing biogas output will help lower UK dependence on foreign NG sources and lower overall emissions from the energy system. The project is supported by partners from across the UK to ensure the aims and objectives can be met, to result in a step-change in the AD industry and position the UK as a global AD leader. The knowledge, tools, and methods developed will be applicable in wastewater treatment, where AD is also used. Beyond that, our AI approaches to systems biology will have potential for widespread application in bioprocessing sectors more generally, such as biopharmaceuticals, biofuels, food, and fermentation. With our network of partners, we will explore potential commercialisation and licencing of our digital techniques to maximise impact and work across sectors toward the common goal of Net Zero.

厌氧消化（AD）是一种微生物分解有机物质产生沼气的技术，从而从废物中产生可再生能源。沼气可以燃烧发电或净化并用作天然气（NG）的替代品。由于它提供了化石燃料的碳中和替代品，同时还通过处理有机废物来防止垃圾填埋场的甲烷排放，AD被认为是英国净零战略的重要组成部分：重新建设绿色。该项目旨在开发人工智能（AI）工具，以实现AD沼气生产效率的根本提高。目前，英国约有650个运营的AD站点，估计可减少英国温室气体排放量的1%。这种贡献是有意义的，但与AD的潜力相比是适度的。目前的根本障碍是缺乏灵活性。由于预测不同废物原料和不同微生物群落在不同操作条件下如何相互作用的复杂性，AD沼气生产商必须通过仅购买最高质量，一致的原料来最大限度地降低风险，这也可能是季节性的;任何错误都可能导致长时间和昂贵的停机时间。因此，可用的废物流大大未得到充分利用;原料价格被推高，削弱了AD沼气生产的经济可行性;有限的原料可能需要运输更长的距离，增加了碳排放。AI为AD沼气生产的优化和未来扩展提供了至关重要的承诺。作为一个不具备其他大型能源部门的核心研究能力的行业，由于相互作用的化学、生物和-如果要实现减少整个生命周期排放-环境系统的复杂性和固有的不确定性，它还提出了特殊的挑战。因此，项目团队将人工智能、工艺优化、系统微生物学和生命周期评估方面的专业知识结合起来，开发出由详细的子系统建模提供信息的全系统决策工具。产出将包括决策工具，具体而言：A）生物消化池的混合机器学习数字孪生模型，基于新的机械建模方法，结合工业合作伙伴的工艺数据和项目的新实验数据;以及B）站点的其它组件的基于优化的系统模型，通过包括经济和环境指标在内的多层数字孪生模型进行现场范围的实时优化。通过将沼气池的数字孪生模型与原料采购和下游工艺联系起来，可以快速确定不同原料、其组合及其价格对沼气质量的影响，同时还跟踪反倾销价值链中真实的量化的环境影响-增加英国广告行业的灵活性将扩大废物市场，降低价格，这将有助于提高该行业的产能，提高利润和生产率，并增强AD作为投资的整体吸引力。增加沼气产量将有助于降低英国对外国天然气来源的依赖，并降低能源系统的总体排放。该项目得到了英国各地合作伙伴的支持，以确保实现目标和目的，从而实现AD行业的飞跃，并将英国定位为全球AD领导者。开发的知识，工具和方法将适用于废水处理，其中也使用AD。除此之外，我们的系统生物学人工智能方法将有可能广泛应用于生物加工领域，如生物制药，生物燃料，食品和发酵。通过我们的合作伙伴网络，我们将探索我们数字技术的潜在商业化和许可，以最大限度地发挥影响力，并跨部门努力实现净零排放的共同目标。