基于符号定向图(SDG)深层知识模型的定性仿真

定性仿真是系统仿真技术中的重要研究和应用方向，同时也是人工智能技术的一个重要方面。近年来，随着现代计算机技术的日新月异，使得基于符号定向图(SDG)深层知识模型的定性仿真研究和应用取得了显著进展。SDG模型能够表达复杂的因果关系，具有包容大规模潜在信息的能力，因此特别适合于过程系统危险评价(PHA)、危险识别与故障诊断。SDG模型结构简单，易于建立和修改。然而高质量的SDG模型必须采用经验数据和定量动态仿真检验和校正。采用SDG模型进行定性仿真的关键技术是高效双向推理“引擎”，由计算机软件自动完成。目前，在计算机辅助过程系统危险评价方面，SDG是最有效的方法；在过程系统故障诊断方面，其完备性较好，但诊断分辨率尚待提高。今后的研究应解决：定性仿真和定量仿真相结合；SDG自动建模；通用推理软件平台；大系统高分辨率故障诊断等问题。

The Qualitative Simulation Based on Deep Knowledge Model of Signed Directed Graph

Qualitative simulation is an important research and application area of the simulation technology, as well as the artificial intelligence. Recently, the highly advanced computational technology has led to the development of the signed directed graph (SDG) deep knowledge model based qualitative approach. SDG model can be used to express the complex cause and effect relations, and has very large capacity of containing process potential information. Therefore, it is especially suitable for process hazard assessment (PHA), hazard identification and fault diagnosis. The structure of SDG model is much simple and easy of establishment and modification. However, the high quality SDG model must be verified and improved by plant information, experience or dynamic quantitative simulation. The key technique of using SDG for qualitative simulation is the high efficiency two-direction inference engine, which is performed by software automatically. Up to now, SDG is the most effective method in computer aided process hazard assessment. In fault diagnosis, SDG has better completeness, but lower resolution. The future directions would aim at the combination of qualitative and quantitative methods, SDG automatic modeling, the generalized inference engine software environment and high resolution fault diagnosis.