Intelligent fuzzy neural network modeling for flexible operation of combined heat and power plant with heat-power decoupling technology

With large-scale renewable energy integrated into the grid, it is urgent to improve the flexibility of thermal power plants to ensure the safe and stable operation of power grid. Heat-power decoupling can effectively improve the flexibility of combined heat and power (CHP) plant. However, the introduction of heat-power decoupling technology makes the dynamic characteristics of CHP plant more complex and modeling more difficult. In response to this issue, an intelligent fuzzy neural network modeling method is proposed in this paper. The modeling method includes identification of antecedent parameters and consequent parameters. For the identification of antecedent parameters, an improved entropy clustering algorithm is presented to automatically partition the data space. Space constraint factor and decision-making constant are introduced to automatically obtain initial cluster centers and clusters number. And to avoid cluster overlap and maintain data integrity, operations such as clusters merging and cluster centers modification are adopted to make clustering more reasonable. An adaptive chaotic nutcracker optimization algorithm (ACNOA) with dual exploration and dual exploitation is presented to acquire consequent parameters. In ACNOA, the trade-off factors are redefined to enhance and balance global exploration and local exploitation for higher parameters convergence speed and search accuracy. Finally, the proposed modeling method is applied to a 350 MW CHP plant with absorption heat pump. In model identification, the ranges of prediction error for four sub-models are ±0.1 MPa, ±0.15 kJ/kg, ±0.08 MW, ±2E-03 MPa, which are less than 1 % of the actual outputs. The simulation results demonstrate the effectiveness and feasibility of the proposed method and provides a basis for the design of controller under flexible operation.