Climate change is one of the biggest global challenges of our time. It is already affecting us today and we expect that future generations will be even more affected. Fortunately, governments and companies are setting increasingly ambitious climate goals, and work is being done all over the world to find solutions for both counteracting and adapting to climate change using technology.
In AI@SINTEF, we work with artificial intelligence and machine learning (ML) in a number of areas to reduce and manage climate change. A key to success is to combine AI and machine learning expertise with domain knowledge.
Energy mix and AI
The transition from fossil to renewable energy is one of the most important measures to reduce our greenhouse gas emissions. To promote use of emission-free energy, we need to reorganise energy production and distribution in smarter and more flexible ways than before.
Artificial intelligence can, among other tools, contribute to more efficient production planning to promote renewables in the energy mix while ensuring stable power delivery, and to maintain control over costs linked to upgrading and maintenance. Artificial intelligence can also enable higher and more flexible interchangeability between carbon-free energy carriers that is necessary for reaching a zero-emission society.
Mobility and transport
To reduce climate emissions from the transport sector, it is important to find new, more climate-friendly and efficient ways of transporting goods and people. This means not only finding the shortest or fastest route, but also which route will generate the least CO2 emissions. Indeed, decisions in logistics are driven by multiple goals which must balanced: Fastest delivery time, shortest route, cheapest delivery but also social and environmental impact.
By combining machine learning and optimization methods, we research how we can pack pallets with goods as efficiently as possible, so that we need fewer trucks to deliver the same volume. Using less truck for delivery will result in less CO2 emissions.
Both in cities and rural areas, we need efficient transport of people and goods. How do we find the best and most efficient routes, also when there are multiple pick-up points and stops? It may be relevant to switch between different transport modalities such as bus, metro or train, but car rentals and scooters should also be considered. In rural areas, the bus does not run as often, and it is more appropriate to have an on-demand solution. Can we design a flexible minibus without a fixed route taking both people and goods on-demand?
AI could also be used for more intelligent traffic management and for nudging people's behaviour in traffic. This can be used to achieve better flow in traffic, higher utilization of available capacity and transition to transport with less greenhouse gas emissions.
A large fraction of our total greenhouse gas emissions is related to the production of goods, in addition to their use. A crucial measure to reduce emissions is therefore to avoid unusable production, so-called Zero Defect Manufacturing, and to carry out production as resource-efficiently as possible. Here too, using AI methods has proved to be particularly useful.
An example is the production of corrugated cardboard, for which warp is a major problem. Here, artificial intelligence can help to understand the relationship between the application of glue, the drying speed in the production process, and the quality of both the processes and the final product. In turn, this makes it possible to reduce waste, errors, and defects. Production plans can also be optimized for lower energy consumption, together with other input factors such as material consumption, storage, and conversion of cardboard to boxes.
Faster transition to circular economy
A key factor for success is to reduce consumption of raw materials. We need to reuse more materials and repair products rather than throwing them away. To accelerate the transition to a circular economy, there are three areas of particular interest in which AI can make a difference:
- Design of circular products, components, and materials.
- Contribute to increased value creation and profitability for circular business models.
- Optimization of circular infrastructure and production.
Digitization and data management are tools allowing to follow material flows throughout entire value chains. This makes it easier to facilitate the recycling and reuse of materials before a resource ends up as waste. To promote reuse in industry, the ability to document the quality of recycled raw materials is important. This will enable efficient identification of raw materials with proper quality when needed. In addition, production facilities must be as flexible as possible to be able to maximally exploit recycled raw materials. To make this possible, it is essential to use data and artificial intelligence.
Efficient operation and refurbishing of buildings
Buildings and infrastructure are a major source of emissions. This applies to all phases: construction, use, refurbishing and demolition. Refurbishing a building is usually environmentally preferable over demolishing and building a new one. Nevertheless, demolition is often chosen over refurbishing.
AI can, through image analysis, data collection and modelling of the building stock, help to understand the current condition of the building and the financial consequences of different choices. Accurate models for building stock conditions can drive decisions toward refurbishment instead of demolition.
At the same time, we can reduce energy consumption in buildings through smart management and good material choices. Smart indoor climate management uses digital twins of the building and power system to provide the best possible experience for the user with the lowest possible energy consumption.
Reduce the societal risk associated with climate change
It is imperative that we act against climate changes. But at the same time changes are already happening and we need to deal with the consequences.
In some cases, we can estimate the risk of incidents and provide advice on preventive measures through modelling and data capture. Examples of this are monitoring of climate effects on buildings, management of surface water in built-up areas , as well as prediction and warning of landslides.
Artificial intelligence can also play a leading role in helping to get a better understanding of the situation during first response to disasters and other events, and thus increase the likelihood of taking good action.