- 期刊
石油污染場址整治技術篩選模糊專家系統之研究
A Web-Enabled Fuzzy Expert System for Selecting Remediation Technologies at Petroleum Contaminated Sites
摘要
石油産品洩漏導致土瓖及地下水污染，在台灣已成爲主要環保議題之一，急需有效的整治。石油污染場址整治技術，可概略分爲現地整治技術及離地整治技術等二類，由於在篩選最適當的整治技術過程中，若面臨專業知識的不足，將導致整治經費大量超支。有鑑於此，本研究發展一套石油污染場址整治技術篩選的Web-Enabled模糊專家系統，以期爲石油業者提供符合成本效益的篩選石油污染整治技術決策輔助工具。本研究以口語草稿分析法進行石油污染整治技術專家的知識擷取，將其知識以問題行爲圖，描述關鍵决策因子的概念，接著以决策表，表達篩選適合整治技術的配對模式，幷將此决策模式轉換爲模糊知識法則，再以Java元件化的封裝技術，結合Matlab建構一完整的模糊專家系統。經由測試個案的實證，本研究所提出之模糊專家系統，確實可以輔助决策者，選擇符合成本效益的石油污染整治技術。
並列摘要
The pollution of soil and groundwater resulting from leak of petroleum products has been one of the main issues of environment protection in Taiwan. Recently, a variety of remediation technologies for petroleum-contaminated sites have been developed from advanced countries, and of which can be classified into two categories, on-situ and off-site. Deciding appropriate remediation technology is a knowledge-intensive task and the lack of comprehensive domain knowledge usually leads to budget overrunning. In this study, we develop a Web-enabled fuzzy expert system for decision making in the selection of cost-effective remediation technologies for petroleum-contaminated sites. We conduct a verbal protocol analysis in order to identify the key decision factors from senior experts and decision process, which is organized as problem behavior graph. In the following, the graph is refined as decision tables showing the matching principals of screening technologies under various site conditions. The resulting tables are transformed into fuzzy knowledge rules in the expert system. The proposed system is implemented by Java component technology and Matlab fuzzy toolbox. From the result of system validation and verification, it is shown that the system really helps the decision maker, according to cost and effectiveness, for selecting the right remediation technologies on petroleum-contaminated sites.