Virtual data-driven nitrogen element products modelling based on broad series-parallel forest

• Multi-software coupled Nitrogen element products (NxOy) numerical simulation model. • Mechanism modeling framework combining numerical simulation and interpretable AI. • Virtual mechanism data-driven modelling based on broad series-parallel forest (BSPF). • Causality analysis of NxOy concentration in municipal solid waste incineration process. The formation of nitrogen element products (NxOy, including NOx and N 2 O) during municipal solid waste incineration (MSWI) processes poses significant environmental threats. This study proposes a novel virtual data-driven NxOy modeling approach based on a novel broad series-parallel forest (BSPF) algorithm. It aims to develop a data-driven model that can accurately predict NxOy generation and provide insights into the causal relationships between manipulated variables (MVs) and NxOy formation. The proposed BSPF algorithm is novel in its ability to handle complex, non-linear relationships and provide accurate predictions. The methodology involves constructing a coupling numerical simulation model for an actual plant, and employing a multi-factor orthogonal experiment to obtain a virtual simulation mechanism dataset. The results, based on an MSWI power plant in Beijing, indicate that the proposed methodology can effectively assist in elucidating the causal mapping relationship between operational variables and NxOy formation. This work is presented as a feasibility study with limited sample size based on orthogonal experimental design, demonstrating the potential of the integrated numerical simulation and BSPF approach. The established framework provides a foundation for future optimization of the MSWI process, with the aim of reducing pollution emissions and supporting sustainable waste management practices. Further validation with larger-scale, multi-plant operational data is required to advance towards a fully deployable industrial model.