Machine learning for process improvement and resource management in the oil and gas sector

用于石油和天然气行业流程改进和资源管理的机器学习

• 批准号: 561264-2021
• 负责人: Lasserre, PatriciaP
• 金额: $ 5.83万
• 依托单位: University of British Columbia
• 依托单位国家: 加拿大
• 项目类别: Alliance Grants
• 财政年份: 2022
• 资助国家: 加拿大
• 起止时间: 2022-01-01 至 2023-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=761678
• 关键词: Machine; learning; process; improvement; resource

Canada is the fourth largest producer of oil and gas combined and the sector represents 5.6% of Canada's GDP with British Columbia being the second largest producer. Oil and gas operators apply to the provincial regulator authorities (the BC Oil and Gas Commission - BCOGC - in British Columbia) for approvals before beginning work (e.g. environmental permits or consultation with Indigenous Nations among others). Efficacy of the Commission is critical for the public, the Indigenous communities affected, and the economic dynamism of the sector. This research partnership proposes to explore the use of data science and machine learning techniques to look at two challenges in regards to improving services for all BCOGC stakeholders groups: (1) benchmarking the application processing timelines, including improving the estimated time for consultation with the impacted indigenous communities, and (2) forecasting application volumes to support resource management at the BCOGC. This research responds to a need to evaluate a new application process introduced by the commission, which allows operators to combine multiple activities (e.g. wells, road...) in one application rendering predictions and their interpretability challenging. Other challenges lie in the amount of data available and the absence of deterministic relationships, the knowledge base being mostly made up of rules and regulations that are not necessarily comparable in mainstream machine learning tools. In this research, we will build a unified framework that can incorporate the non-commensurable information for improved predictions. The factors affecting the timeline can be found in similar processes in other provinces, and with operators' internal processes. Standardizing expectations around application timelines across jurisdictions will allow operators to better plan their investments in Canada. It also provides an opportunity for regulators to compare service levels and performances, and collaborate on finding opportunities to improve. Forecasting the number of applications received by BCOGC will have a direct impact on BCOGC resource allocations, thus supporting higher quality responses, and ensuring the identification of environmental risks.

加拿大是第四大石油和天然气生产国，该部门占加拿大GDP的5.6%，不列颠哥伦比亚省是第二大生产国。石油和天然气运营商在开工前向省监管当局(不列颠哥伦比亚省不列颠哥伦比亚省石油和天然气委员会--BCOGC)申请批准(如环境许可证或与土著民族协商等)。委员会的效力对公众、受影响的土著社区以及该部门的经济活力至关重要。这一研究伙伴关系提议探索使用数据科学和机器学习技术来研究在改善对BCOGC所有利益攸关方群体的服务方面面临的两项挑战：(1)确定申请处理时间表的基准，包括改善与受影响土著社区协商的估计时间；(2)预测申请量，以支持BCOGC的资源管理。这项研究回应了对委员会引入的新申请程序进行评估的需要，该程序允许运营商结合多种活动(例如油井、道路……)在一个应用程序中，呈现预测及其可解释性具有挑战性。其他挑战在于可用的数据量和缺乏确定性关系，知识库主要由规则和法规组成，而主流机器学习工具不一定具有可比性。在这项研究中，我们将建立一个统一的框架，可以结合不可通约的信息来改进预测。影响时间表的因素可以在其他省份的类似流程中找到，也可以在运营商的内部流程中找到。对跨司法管辖区的申请时间表进行标准化预期将使运营商能够更好地规划他们在加拿大的投资。它还为监管机构提供了一个比较服务水平和表现的机会，并合作寻找改进的机会。预测BCOGC收到的申请数量将对BCOGC的资源分配产生直接影响，从而支持更高质量的响应，并确保识别环境风险。