ML (Machine Learning) Approach to control Thermal Strategies and mitigate Sensor Failure penalty on Emissions

As vehicle emission standards are becoming stringent worldwide because of the looming climate crisis, it is important to control the pollutants that vehicles emit. To achieve the stringent emission target, it has become a priority to enhance the capability of emission control system (ECS) which consist of Diesel Oxidation Catalyst (DOC), Diesel Particulate Filter (DPF) and Selective Catalytic Reduction (SCR) sub-systems. One of the bottlenecks is the limited operating temperature range of the after-treatment system. In modern emission control systems, the temperature characteristics should always be optimized to have the best efficiency. To achieve the optimal operating temperature, different thermal control strategies are followed in the Engine and emission control unit. Temperature sensor values are one of the primary inputs for thermal management strategies. In the event of temperature sensor malfunction, the ECS performance is affected due to incorrect temperature input, resulting in higher emissions leading to performance limitations. To mitigate this issue, it is important to predict the exhaust gas temperature precisely. In this paper, studies are carried out to show Machine Learning (ML) based digital sensors can be instrumental in maintaining ECS functionality and performance. This paper focuses on developing Machine Learning (ML) Model to replicate the sensor prediction based on dependent parameters. A Multi-Layer Perceptron (MLP) neural network is explored and implemented to predict the SCR inlet temperature. The predicted temperature is used to control various thermal strategies to improve the SCR performance. The selected model is trained and tested with actual vehicle data for real time correlation. The model’s performance is improved through evaluation metrices such as R2-Score, Mean Squared Error, and Mean Absolute Error. These metrices provide a thorough evaluation of the algorithm’s performance compared to the actual observed values. The high R2 Score indicates strong predictive capability, while the low errors demonstrate the model’s reliability.