基于二次规划的炼化设备泄漏诊断技术研究

Leakage fault detection and identification of chemical equipment based on Quadratic programming method

石油化工生产过程的设备泄漏会带来严重的安全隐患和环境污染问题,而现有的设备泄漏检测主要是应用专业的泄漏监视仪器,增加了生产投资成本和维护成本。应用基于人工智能的生产数据分析技术对炼化设备泄漏诊断进行了研究,在不增加专业设备检测仪器的前提下,根据对炼化过程数据及过程约束条件的分析,将泄漏诊断问题转化为一个连续非线性规划问题,应用二次规划法对过程设备泄漏进行监测,并对泄漏项惩罚权重提出了自适应赋值式,从而可准确的对泄漏点进行定位,最后通过一个仿真实例验证了算法的有效性。

Process equipment leakage introduces danger to chemical production safety and serious environment pol- lution, and the current methods to monitor the leakage is to use the special equipment for leakage. To decrease the cost of leakage monitoring, a quadratic programming method to detect and identify the leakage was proposed based on the analysis of process constraint and data, an equation to determine the value of penalizing bias item adaptively was also proposed. A simulation case showed the affectivitv of the proposed method.