摘要
现场采集的数据不可避免地包含一些诸如噪声干扰之类的不确定性,由数据驱动建立的模型需要具备较强的处理不确定因素影响的能力.在以往文献的Ⅰ型T-S模糊建模方法的基础上,提出了一种基于数据驱动的Ⅱ型T-S模糊建模方法.其过程是通过分析采集的数据样本计算得到不确定因素的影响程度,在Ⅰ型T-S模糊模型的基础上,前件参数上采用Ⅱ型的模糊集来代替Ⅰ型的模糊集,后件参数上则采用I型模糊集来代替数值,由此拓展得到Ⅱ型T-S模糊模型.最后通过pH中和反应过程对所提出的方法进行仿真验证.仿真结果表明,该方法建立的模型能更好地处理不确定因素的影响,取得更高的准确度.
Data collected from the field inevitably contains uncertainties such as noise or other disturbances; mathematical models established with data-driven approach must possess strong capability to deal with the influence of uncertainties. Following analysis of current methods for type-Ⅰ Takagi-Sugeno （T-S） fuzzy modeling, a method suitable for type- Ⅱ T-S fuzzy modeling was proposed. In the data driven modeling process, the influence of the degree of uncertainty was determined by analysis of the collected data. On the basis of the type-Ⅰ fuzzy model, for antecedent parameters, we employed the fuzzy set of the type- Ⅱ fuzzy model to replace the counterpart from the type- Ⅰ model . But for consequent parameters, we took type-Ⅰ fuzzy sets to replace crisp numbers. This produced an improved type-Ⅱ T-S fuzzy model. Finally, a pH neutralization process was taken as an example to verify the proposed mathematical model. Simulation results showed that this method can handle the influence of uncertainties better and achieves higher accuracy.
出处
《智能系统学报》
2009年第4期303-308,共6页
CAAI Transactions on Intelligent Systems
基金
国家自然科学基金资助项目(60604018)
上海自然科学基金资助项目(06ZR14044)
