【摘要】 通过对炼油厂分馏塔侧线质量进行机理分析， 并根据实际流程利用ＡＳＰＥＮ对过程进行模拟得到各控制变量与侧线质量的对应结果， 利用数理统计方法， 得到带有可信度ＣＦ量度的规则。利用正向推理， 构置了一个基于规则的在线监测炼油厂分馏塔侧线质量指标的模糊专家系统ＳＱＰＥＳ， 并利用工厂采集的实时数据对所建立的专家系统进行了验证。结果证明， 这种规则获取手段对建立炼油厂分馏塔质量指标监测的专家系统是适用的，它的建立， 既可作为油品质量在线监控的参考， 又可为在线计算部分提供计算结果判断变化趋势正确与否的依据， 在工业生产中有比较广阔的应用前景。更多还原

【Abstract】 The first, distillation tower side draw products’ qualities and the influencing factors are analyzed with mechanism, then, with a certain flow sheet, ASPEN was used to get highly reliable qualities result. After that, those results are processed with orthogonal experiment method to find those most important factors that influence the final side draw products’ qualities, statistics principles are used to extract the rules with a certainty factor (CF). Based on the rules, a fuzzy expert system SQPES is built up with forward chaining method. At last, the fuzzy expert system SQPES is tested with simulation data and DCS data collected from the factory. The test results show that, this kind of rule extracting method is suitable for getting the rules of distillation tower side draw products’ qualities. This expert system can provides a reliable means for predicting the qualities of side draw products and can used not only as a separate on\|line predicting software for on\|line control, but also as one part of on\|line simulation as an assistant part to judge if the simulation results change correctly.更多还原