IBM Research India     

links

IBM Research India - Climate and Sustainability

Climate and Sustainability 

Enterprises are under significant pressure from investors, consumers, and policymakers to act on climate change mitigation by disclosing their GHG emissions and committing to reduction of emissions from their industrial activities including operations, manufacturing, logistics, and supply chains. In fact, over 20% of the world’s largest companies have set long term net-zero targets but need technology help to measure, track, and reduce their emissions while building operational resiliency to adapt to the ill-effects of climate change[1]. IBM has a unique opportunity to address this opportunity through a highly differentiated technology offering that will support enterprises with enterprise-grade support and integration. Towards this goal, IBM Research spear-headed the work on developing the newly released Environmental Intelligence Suite (EIS) to manage an enterprise's carbon performance leveraging AI, optimization, and remote sensing capabilities. The technologies now available within the EIS also help companies predict, prepare for and adapt to the increasingly severe risks of global warming due to to climate-change. This unique integration of AI with environmental, climate, and weather data is aimed at adapting to and mitigating climate change with an accurate analysis of climate risk at scale and precise accounting of carbon emissions. 

 

Based on IBM's AI and Hybrid cloud capabilities, IBM Research is working on developing a suite of solutions (as shown in Fig.1) for climate mitigation and adaptation as follows: 

• Foundational technologies: IBM Research has designed, developed, and productized an enterprise scale carbon accounting engine (compliant with the Green House Gas Protocol) under the Environmental Intelligence Suite (EIS):https://www.ibm.com/products/environmental-intelligence-suite. Over the years, IBM research has also developed a global geo-spatial data platform called PAIRS embedded with state of the art AI, data management, and remote sensingcapabilities. Based on these foundational technologies[2,3], IBM research is developing the next generation of decarbonization technologies to accelerate the net-zero journey of enterprises.     
• Data & AI: The accuracy of carbon footprint estimation heavily relies on data quality. However, data quality issues e.g., due to duplicate and / or incomplete / missing data are a scommon pain-point across enterprises. Existing approaches to resolve these issues are fairly basic. and are limited by the availability of domain specific data. To address this challenge, we have developed a generalized, scalable transfer learning-based imputation framework which learns the forward and backward temporal dependencies in time-series data. To address the lack of sufficient enterprise data, IBM Research is also developing an inverse modelling approach of linking limited emission data with atmospheric concentration using atmospheric transport and chemistry models. Inverse modelling is used to check the consistency between bottom-up emission inventories and GHG concentrations measured in the atmosphere using satellite observations. Thus, it provides estimates of total emission, both anthropogenic and from natural sources.  
• Decarbonization Accelerators: These are solution 'recipes' that abstract out common software components across a chosen class of problems in the context of carbon performance estimation and optimization. We are developing the following decarbonization accelerators to help enterprise to optimize their own as well as their supply chain emissions as follows.: 

• Carbon Accounting for Hybrid Cloud: Enterprises are heavily dependent on Information Technology for digitizing and automating their operations. Much of these enterprise IT workloads are either already deployed on the public cloud or private data centers or expected to migrate to a data center in the near future. The estimated electricity consumption of data centers is of the order of 200 terawatt-hours, which is approximately 1% of global electricity consumption. Although the grim energy predictions of the past for data centers have not come to bear, with the increase in AI-powered workloads and other trends, ensuring that the energy consumption at data centers still remains contained requires continued and significant investments to identify and eliminate inefficiencies at its various elements  and operations that consume and hog power. At IBM Research, we address the above problem using an ambitious, comprehensive, and multi-pronged strategy that includes carbon quantification for tenants and workloads on IBM Cloud and on-prem data centers, AI-infused sustainability transformations for enterprise customers, and multi-disciplinary sustainable computing research spanning the areas of infrastructure, hardware systems, platform, software, and AIOps to improve manufacturing processes , design specialized hardware and cooling systems, build improved software, and develop innovative run-time algorithms to manage systems and software to mitigate environmental cost over the complete lifecycle. 
• Emission Performance Accelerator: We have developed a generalized, scalable, and easy to use Emission Performance Accelerator to identify carbon hotspots in Enterprise assets and operations. This asset would help enterprises to compare and contrast emissions and associated contextual parameters such as weather, asset metadata, operations related data to find the factors behind the regulatory risk of assets. Using explainable AI the tools would also help to identify quantitative and qualitative opportunities for carbon reduction. 

• Land Based Emissions Accelerator: We are developing a novel, configurable Land based Emission Estimator and Reduction Accelerator (LEERA) which is an ensemble framework of static and dynamic GHG estimation approaches which can provide hyperlocal, yet scalable spatio-temporal GHG maps for the region and period of interest to derive data-driven insights to reduce the emissions. Using the spatio-temporal GHG maps, the framework can perform hotspot identification, and sustainable supplier selection and procurement, fugitive emission detection, localization and quantification which will allow the stakeholders to take actions based on data-driven insights to reduce GHG emissions.  
• Emission Optimization Accelerator: The goal of this work is to conceptualize, implement and demonstrate a generic 'emission optimization accelerator' framework that can accurately and efficiently address the cost-emissions trade-off that is commonly observed in many industry constructs like supply chains, asset management, cloud computing etc. To start with, the focus will be on the following 2 use-cases of interest in supply chains: i) order fulfilment optimization and ii) inventory optimization; and potentially for future business use-cases of interest.  

 

 

 

 

As mentioned earlier, many companies are committing to net-zero goals, but do not know exactly what targets to set year on year so they reach net-zero by year X (e.g., 2030). To address this gap, we are working on following scientific challenges: 

• How to develop a rigorous, science-based carbon target advisory that can recommend the ideal future carbon trajectory to any company, irrespective of industry type, size, location etc. so they can feasibly reach their net-zero goals. 
• What decisions must the company be taking every day / month / year, so that they are still on track with respect to the committed net-zero goals? And, if they do go off-track, what should they be doing to come back on track with little cost over-runs? 

 

References: 

[1] https://eciu.net/analysis/reports/2021/taking-stock-assessment-net-zero-targets 

[2] ibm.biz/environmental-intelligence  

[3] https://research.ibm.com/science/deep-search/