The energy industry faces great challenges in light of climate change, the increasing demand for energy, and the urgent need to shift to sustainability and reduce carbon emissions. Can artificial intelligence play a role in achieving these goals?
In an interview with ESG Mena, Ron Beck, Senior Director of Product Marketing at Aspen Technology Inc., explains how his company uses industrial AI to achieve energy efficiency and emissions reduction in heavy industry, focusing on successful experiences in the region.
- How is Aspen Tech’s Industrial AI technology specifically contributing to reducing carbon emissions in industries? Can you provide specific examples of successful implementations?
AspenTech Industrial AI solutions are already significant contributors to reducing carbon emissions. Aspen Hybrid Models combine AI algorithms with first principles modeling systems Aspen Plus and HYSYS. Hybrid models are increasing the speed and accuracy of digital twin on and offline models; hybrid models are resulting in up to18% energy efficiency savings in economic units in plants, hybrid heat exchanger (HX) train models are providing guidance for better HX cleaning, saving up to 15% energy use. Hybrid reactor models in chemical plants are reducing waste materials and water use. Industrial AI-powered adaptive process control is improving energy use an incremental 5-10%. AI neural network powered renewable energy forecasting for power distribution with the AspenTech ADMS system are optimising utilisation of wind and solar power to reduce carbon intensity of electricity provided to electrical powered equipment. Machine-learning powered prescriptive maintenance, the Aspen Mtell product, provides 10-60 days advance warning of equipment degradation, reducing unplanned downtime by 50-80% and therefore avoiding over 50% of site flaring. Each unplanned shutdown can force flaring, emitting as much carbon as up to six months of normal operations. Eliminating these unplanned shutdowns has huge impact.
A number of companies have reported on the significant contributions AspenTech Industrial solutions have provided in reducing carbon emissions. Let me share several examples:
Socar, Azerbaijani Energy Company, has implemented hybrid models achieving 36% waste heat recovery in certain chemical units.
Tupras, Turkish refining giant, has implemented hybrid models for refining heat exchanger networks, for 1-2% CO2 emissions and associated energy savings of over $1M per year.
Sardeolica, Italian utility scale wind farm operator, has implement Aspen Mtell AI-based machine learning that is extending the lifetime of their wind turbine systems by an estimated 8% life extension, with associated incremental renewable electricity production.
Saudi Aramco and AspenTech are collaborating on the release and further development of the Aspen Strategic Planning for Sustainability Pathways (SPSP), this helps energy and chemical enterprises to map 10-20 year plans for incorporating low carbon process technology and process units into energy and production operations.
These are just four of many examples of industrial AI strategically reducing carbon emissions.
2. What are the challenges facing the use of AI and technology in carbon reduction processes, especially since there are accusations against technology that it contributes significantly to increasing the production of carbon emissions?
Asset-intensive industries, for instance, companies with refining, LNG, hydrogen chemical, mineral processing, and power generation assets, have been slower to embrace AI in the key processing and optimization business functions. The reasons are logical. These industries have a “safety first” mindset closely followed with focus on operational excellence. To overcome that AI reluctance, AspenTech has focused on delivering AI within the framework of our highly trusted first principles engineering model-based systems, such as Aspen HYSYS and Aspen Unified PIMS and we call that INDUSTRIAL AI.
One of the key principles that AspenTech follows is what we called “purpose built AI”. What we mean by that is to apply the right digital technology solution to the right problem. Some problems are inherently linear, and not difficult to model and make effective decisions.
The more powerful AI algorithms, including generative AI and large language models, should only be applied to the right problems that they are best suited to solve. Yes, large language models are electric consumers; but AspenTech (and others) are also pursuing more efficient approaches, including small language models (SLMs) and hybrid models based on simulation data.
Such approaches can be extremely efficient in their use of data and compute power, and when applied to sustainability problems such as increasing energy efficiency have an extremely positive net benefit toward sustainability outcomes.
Our solutions are interestingly, also being applied by organisations we partner with such as Microsoft to model and develop solutions for conserving and recycling the waste heat generated by massive data centers, improving their carbon footprint.
3. How can Industrial AI and other digital technologies assist the MENA region in meeting the increasing demand for energy while also reducing carbon emissions?
Industrial AI and associated digital technologies are already creating significant and measurable value, both in making better decisions around the growth and optimisation of increasingly complex assets, as well as in reducing emissions.
A key challenge in the region is to master turning these AI solutions into business value. There is a lot of staff upskilling required. There is also a crucial role for executive leadership to embrace and guide the business and organisation change process required to embrace AI.
As key Middle East players expand their energy networks into other regions, especially SE Asia and the Pacific Rim, the value and supply chains become more complex and stretched. The opportunity is huge, but it comes with the challenges of understand and manage the increased complexity. Here is where applying industrial AI, to increase organisational agility, to be able to understand and make optimal decisions across business complexity is crucial. We believe industrial AI plays a necessary and strategic role in making this happen.
4. You have a program named “Innovate for sustainability”, how could this program help companies and organisations to use AI to achieve sustainability?
Some of the key sustainability trends and forces, such as circularity and the energy transition, require not incremental improvements but step-change improvements. This requires embracing innovation and change and the companies that can be agile and creative will win.
We are embracing a new way of co-innovating with customers to support the need for faster innovation and for companies to work together to achieve that. One very concrete outcome of this is our partnership with Saudi Aramco on digital innovation to support strategic planning. We are working on others in the region and globally.
5. Why is upskilling the workforce crucial for successful decarbonisation, and how can AI technologies support this process?
Successful decarbonisation of an enterprise requires leadership from the top – namely executive vision and commitment. But the companies that are being most successful are the ones where every person in the workforce is made to feel that their role and impact matters.
The right digital tools are absolutely crucial here, to provide the visibility and transparency such that each worker can see the impact that their actions, on a daily basis, have on a company’s operational excellence and decarbonisation journey.
Industrial AI software is designed to be easy to adopt and use. At the same time, explainability and transparency are crucial concepts. For that to work in an organisation, all technical, knowledge, and operational workers need to have the new skills needed to understand and interpret how AI software is interpreting data and recommending action.
The upskilled workforce will be the key players in making sure the full value of industrial AI innovation can be captured by an organisation.
AspenTech has introduced, and is continuing to introduce, a wide range of ways for providing education and “just in time” advice to the workers who are striving to get the most from these new tools.
6. What differentiates Industrial AI from general AI?
Industrial AI from AspenTech combines the speed and power of AI algorithms with the efficiency and guardrails of real-world domain expertise (i.e., engineering fundamentals, asset operational insight and industry knowledge). These Industrial AI solutions address the risks of AI via guardrails, robustness and trusted results. The unique advantage of AspenTech’s engineering models in the AI journey, is that these models will generate simulated data that leverage laws of chemistry and physics while filling gaps in the range of scenarios needed to extrapolate insights. They augment historical and real time asset data. Also, by seamlessly integrating AI applications into existing optimisation software already in wide use, industrial AI applications become easy to adopt by both experienced users and the next-generation workforce.