A bi-level dispatching model for load-balancing of distributed-flexible biogas plants in multi-energy complementary microgrid: A hybrid data-driven and mechanism-based approach

This study proposes a bi-level dispatching model for coordinating flexible biogas plants operating to support wind/solar load balancing within a regional microgrid. The upper-layer model employs a dual-objective optimization to generate the optimal biogas dispatching scheme for each biogas plant, while the sub-layer utilizes a BP-PID control model to enable flexible biogas production and adjust feeding schedule accordingly to track the desired biogas demand curve. An LSTM-enhanced Gompertz model is integrated into the controller to capture the relationship between the real-time feeding input and biogas demand response, leveraging a hybrid data and mechanistic-driven approach to combine empirical adaptability with physical interpretability. The experimental results confirm the model’s predictive accuracy and precision, and demonstrate that the system can achieve flexible biogas production while maintaining improved process stability. This approach is expected to enhance the disturbance resilience of flexible biogas production under complex microgrid conditions, thereby maintaining stable system operation, promoting the degradation of organic pollutants, and providing theoretical and technical support for practical applications.