摘要
基于文献[5]的试验数据,建立了生物质气化气中污染物氨催化脱除过程最小二乘支持向量机(LSSVM)模型。仿真结果表明,该模型预测值与试验值的最大相对误差为3.01%,模型具有较好的泛化能力。在此基础上设计了氨催化脱除过程多目标优化函数,并通过并列遗传算法寻优得到了氨催化脱除过程Pareto最优解集。寻优结果表明,在Pareto最优工况条件下,氨催化分解转化率能达到98.5%以上,催化过程总体性能优于试验值。
Pollutant ammonia catalytic removal process of least squares support vector machines （LS-SVM） model in biomass gasification gas was established based on the test data of document. Simulation results indicated that model predictions and experimental value of maximum relative error was 3.01% which had a better generalization ability.Catalytic removal process of ammonia was designed on the basis of this multi-objective optimization functions and Pareto optimal solution set of ammonia catalytic removal process was produced. Optimization results showed that under the Pareto-optimal working conditions, conversion rate of catalytic decomposition of ammonia could reach more than 98.5% and overall performance of catalytic process was better than testing data.
出处
《可再生能源》
CAS
北大核心
2013年第10期115-119,共5页
Renewable Energy Resources
基金
河北省自然科学基金资助项目(F2011502001)
关键词
生物质气化气
污染物氨
催化脱除
优化
biomass gasification gas
pollutant ammonia
catalytic removal
optimization
