MechChem Africa January-February 2025

⎪ Hydraulic, pneumatic and automation solutions ⎪

Transforming utilities with AI Satyajit Dwivedi, regional director for energy utilities, mining & metals and the public sector for the global data analytics and AI specialist, SAS, highlights the value and challenges associated with incorporating AI into utilities management systems, before outlining a transformative pathway to successful adoption.

O ver the past decade, the utilities industry has undergone signifi cant evolution. Methods for gen erating, delivering and utilising energy have transformed from the practices of the early 2000s. This transformation in cludes the completion of enterprise resource planning (ERP) rollouts, which have enhanced data processing and integration capabilities. Smart meters have improved real-time data collection and accuracy in monitoring energy and water usage, and asset management (AM) and customer relationship management (CRM) systems have led to more efficient operations and better customer service. Central to these advancements is the increasing reliance on data, which has signifi cantly enhanced the management of electric ity and water access and improved the bal ancing of supply and demand. Furthermore, the focus on renewable energy and the integration of distributed energy resources (DERs) have become pivotal in the energy security of countries and the transition to sustainable energy systems in industries. This shift supports energy security by diversify ing energy sources and improving resilience against disruptions. In addition, changing customer behav iour driven by economic fluctuations and climate change presents a critical dimension. Economic ups and downs influence consumer energy consumption patterns and demand for services, while climate change introduces new challenges and pressures on utility op erations. Consumers are increasingly seeking sustainable operations and cost-effective solutions, prompting utilities to adapt their power generation and transmission strategies and customer centric programmes. With a data explosion from meters and sensors, the adoption of AI is becoming crucial for addressing changing demand, optimising energy management, enhancing grid reliability, and supporting the overall energy transition, all while responding to a dynamic landscape shaped by economic and environmental factors. Challenges of AI adoption Implementing AI in utilities is fraught with challenges, largely due to the diverse and sometimes conflicting understandings of AI across various departments, such as busi-

ness and IT. Each department may have its own interpretation of what AI can achieve, leading to different dashboard and reporting needs and duplicate demands for AI solutions. These disparate needs and interpretations often prevent the unification of AI efforts under a common strategic context, without which, different departments may select and implement different AI technologies, resulting in fragmented and inefficient deployment. The lack of a dedicated team for digital footprint consolidation and rationalisation exacerbates these challenges, causing utilities to struggle to develop cohesive AI strategies and architecture. This oversight is further compounded by the absence of a multi-year AI transformation roadmap, which hinders the ability to plan and execute AI initiatives in a structured and strategic manner. As a result, there are significant gaps in defining AI use cases and a lack of a comprehensive AI value framework, impeding the realisation of AI’s potential across an organisation. AI transformation roadmap Digital footprint rationalisation should pro vide a clear AI transformation roadmap. It must comprehensively address core initia tives to ensure a robust approach to AI inte gration into the business processes delivering value. Core initiatives are those critical to the utility’s primary operations and include revenue protection efforts such as smart collection analytics, reducing losses and/or minimising energy costs. The focus should be on using AI to enhance short, medium and long-term demand fore casting accuracy and associated peak load management, which are essential for balanc ing supply and demand. Core applications also involve grid maintenance: through predictive analytics and intelligent spare parts manage ment, for example. This involves fuel demand forecasting, fuel supply chain optimisation and health, safety and environment (HSE) analytics, using AI and drones. Advanced machine learning techniques can also be deployed for non-intrusive load disaggregation-based (NILD-based) en ergy management and or demand response management using highly granular meter or datalogger data. Non-core initiatives, while not directly tied to primary operations, are crucial for overall

organisational efficiency and support. These include continuous monitoring for spare parts contracting and procurement, and advanced human resources (HR) analytics, which help streamline processes and improve cost efficiency. Data to decisioning AI Having a data to decisioning platform that runs on modern and scalable architecture is quintessential to maximising on the power of AI, resulting in substantial value realisation across both core and non-core areas. In core areas, a robust platform can provide AI solu tions that deliver highly accurate large scale forecasting, real time process or energy op timisation, intelligent decisioning workflows for deploying departmental resources, in sights from unstructured data, and optimised decisions with scenario planning. In non-core areas, a robust platform sup ports continuous monitoring for processes such as procurement, staff pay and recruit ment. By leveraging a cloud-native AI platform built on Kubernetes architecture, utilities can seamlessly integrate and analyse data from diverse sources, ensuring a unified approach to addressing both strategic and operational challenges. Such a holistic platform empowers utilities to embrace a new era of smart, predic tive, and efficient solutions, driving transfor mative outcomes across their organisations. By adopting a comprehensive AI transfor mation roadmap, utilities can align their AI initiatives with both strategic and operational goals, ensuring a unified approach that maxi mises value across all areas, enabling utilities to adapt to economic fluctuations, climate change, and evolving customer expectations: ultimately fostering a more resilient and sus tainable energy infrastructure. www.sas.com/en_sa

January-February 2025 • MechChem Africa ¦ 27