Radio access networks must consume energy more efficiently - the Creating a sustainable digital future: low-carbon networks Catalyst shows us how
Base stations of the future: using AI and renewables to create more profitable radio access networks
Few would disagree that the most urgent global challenges we face today include climate change. Industry-wide initiatives and strategies to reduce emissions and lessen the impact of climate change are various, and involve close examination of the many ways in which energy is consumed. The radio access network (RAN) is a fundamental pillar of telecoms infrastructure, and like other systems and equipment, needs to run more efficiently. Much RAN consumption occurs from base stations and their associated passive infrastructure such as air conditioners, inverters, and rectifiers. According to China Mobile, this equipment alone accounts for 70% of direct network emissions, and of these, over 30% is attributable to cooling systems alone.
Improving the energy efficiency of this critical infrastructure is the exact purpose of the Creating a sustainable digital future: low-carbon networks Catalyst. To achieve this, the project has identified various ways in which newer connected technologies can improve base stations’ energy consumption. The first and foremost has been to introduce new energy saving channel functions which use AI and machine learning (ML) to distinguish between broadcast signals and data transmission, and then perform a near-immediate channel shutdown when appropriate. AI and ML have also been incorporated to analyze service patterns and user preferences. In doing so, base stations can allocate resources based on real-time requirements, reducing latency and improving energy-efficiency.
AI is also being used to create intent-driven energy reduction strategies. The project has introduced an intention engine which can identify and translate natural language inputs to create energy-saving strategies which can be executed with automation. Results are then visualized and tailored according to operational and maintenance contexts and requirements. By adhering to best practices from TM Forum documentation IG1218B and IG1218C, and adopting TM Forum assets such as IG1230, IG1253, TR290, IG1305, and 921A, the solution has achieved an intent-driven, closed-loop system to achieve energy reduction.
Aside from AI, the Catalyst has improved the efficiency of existing infrastructure within the equipment room. For example, by syncing humidity sensors and thermometers with meteorological data and time-domain forecasting algorithms like long short-term memory (LSTM), the temperature can be accurately predicted and cooling systems engaged more efficiently, improving the power usage effectiveness (PUE) of equipment. To reduce emissions further, the project team also sought to power more base stations using renewable energy sources. To do this, the project team needed to ensure the energy source was stable and reliable – a common pitfall when adopting renewable solutions. To make solar a viable source of power, the Catalyst introduced a maximum power point tracking (MPPT) controller, which optimizes the power output of solar panels by continuously tracking and adjusting the operating point to the maximum power in variable conditions.
Application and wider value
Through the combination of these energy efficiency methods, the Catalyst has successfully reduced energy consumption by 25% in 5G base stations, and achieved a PUE reduction of 13%. The project has also demonstrated improved photovoltaic energy efficiency, which is now achieving 98% efficiency, and ready to be deployed more widely. These impressive early results are a testament to how effectively AI and ML can be used to reduce energy, but also to genuinely deliver increased performance. This of courses, makes much easier for CSPs of any size to adopt similar approaches to energy reduction.
According to Yun Li, a Manager at China Mobile and Project Champion for the Catalyst, “this project adopts innovative energy-saving technologies and strategies that can significantly reduce the energy consumption of base stations. As well as this, CSPs could experience significant efficiencies, reduce operational costs and increase profitability by maximizing radio resource use and reducing waste. Crucially, the solution can be scaled, replicated and supports various vendor configurations, allowing for seamless adoption for CSPs operating similar spectrum, which has been the case for AIS Thailand. The project's innovative technologies and strategies have brought new opportunities and prospects to the telecommunications industry. It gives network operators more effective, adaptable, and intelligent solutions, driving technological progress in the industry and helping to achieve global sustainable development goals.”