Summary
Air pollutants contribute to global warming and climate change. Nitrogen oxides emissions have been increasing by 33% from 1990 until 2021 in upper-middle-income countries and by 11% in low-income countries. In total, in Portugal 40.7% of NOx emissions were due to road transport, 24.3% were from Industrial combustion, 10.7% of emissions were from other mobile sources and 9.1% of NOx emissions were from power stations. Advancements in data learning, optimization of algorithms and predictions enable the development of neural networks and computing. In this work it was developed the monitoring and prediction of nitrogen oxides emissions with neural networks considering experimental data from a natural gas power plant.
Read more Read lessIt was obtained with an accuracy of 0.99 and a RMSE of 3.46, the prediction of NOx emissions with the development of the optimized feed forward algorithm. Furthermore, it was developed a LSTM-RNN algorithm with an accuracy of 0.98 and a RMSE of 0.06. This work was done in line with the goals set in the European Green Deal towards climate neutrality, the Eurpean
Climate Law, the “Fit for 55” legislative package, and the Industrial Emissions Directive.
Results have shown that machine learning and neural networks are useful and efficient for prediction and reduction of emissions, improving sustainability in power systems and mitigating climate change.
Additional informations
| Publication type | Session Materials |
|---|---|
| Reference | C3_11187_2026 |
| Publication year | |
| Publisher | CIGRE |
| Country | Portugal |
| Study committees | |
| File size | 877 KB |
| Price for non member | 30 € |
| Price for member | 30 € |
Authors
PEREIRA Fabiola - UNIVERSIDADE DE LISBOA
