有机朗肯循环（ORC）系统实验是验证或获得系统性能的有效手段，为了在实验工况有限的前提下获得ORC系统最优运行工况，本文在ORC系统变工况实验研究的基础上，提出了基于BP神经网络的ORC系统性能预测方法并建立了仿真模型。预测结果表明：该模型验证最大误差为3.30%，能够预测出ORC热力性能更优的运行工况，其系统净输出功最大值1.47 kW时的运行质量流量为0.15 kg/s，热效率最大值3.71%时的运行质量流量为0.134 kg/s。

Organic Rankine cycle (ORC) system experiment is an effective means to verify and obtain system performance. In order to obtain the optimal operating conditions of ORC system under the premise of limited experimental conditions, in this paper, based on the experimental study of ORC system under variable operating conditions, the performance prediction method of ORC system based on BP neural network is proposed and the simulation model is established. The prediction results show that the maximum validation error of the model is 3.30%, the operating conditions with better ORC thermal performance can be predicted.The maximum net output power of the system is 1.47 kW for the operating mass flow rate of 0.15 kg/s,the maximum thermal efficiency is 3.71% for the operating mass flow rate of 0.134 kg/s.

TK123

国家自然科学基金面上项目（51876043）；国家自然科学基金重点项目（51736005）