變壓器是供電系統的心臟,升壓變壓器為提升電壓以減少線路損失,便利長距離輸送;降壓變壓器為降低電壓以提供各電壓等級之用戶使用,故扮演著極重要角色。 變壓器是一種十分貴重的供電設備,發生事故對於影響供電系統相當嚴重,因此維持變壓器正常運轉是現今運轉維護人員的重大責任。 變壓器內部所使用的絕緣材料有絕緣紙、絕緣木、壓紙板、環氧樹脂、玻璃纖維、絕緣油,在長期的運轉過程中受到溫度及濕度的影響,導致絕緣逐漸劣化,最終因內部絕緣破壞而形成事故;因此如何及早發掘變壓器內部異常的徵兆及時做適當應變處理乃是本研究的目的。 變壓器運轉時產生的油中氣體是一項相當重要的診斷依據資料,目前主要檢測方式是以週期性離線方式抽取油樣進行溶解氣體分析,藉以判斷變壓器之性能,其缺點在檢測期間形成一空窗期,不易即時掌握油中氣體的變化動態。 本研究為變壓器故障早期偵測診斷,實際監控台電台南變電所運轉中的一台電力變壓器,結合個人電腦及線上油中氣體偵測器,除即時擷取資料及監控變壓器的油中氣體成份外,同時利用類神經網路理論進行變壓器的故障原因診斷。在本論文內,所裝置的油中氣體偵測器可有效檢測出氣體變化,再配合類神經網路的正確分析診斷,可及早發掘變壓器潛在故障並可即時處置,以防範事故發生於未然。
It is not overstated that transformer is the heart of electrical system. A step-up transformer is used to reduce the power losses, while the step-down transformer provides the various levels of voltage to the customers. Transformer is one of the most expensive and important electrical equipments installed in a power system. Any fault occurs in a main transformer will cause in the malfunction of whole power system. Those materials used in the interior parts of a transformer include insulated paper, stereotype paperboards, wood, dioxin resin, optical fiber, and insulation oil. The property of their insulation level could be deteriorated due to the change of ambient temperature and humidity in the long time operation. Thus, an inevitable fault may occur owing to the poor insulation. To discover and analyze those abnormal symptoms in the earlier stage and takes remedies in time is the main purpose of this study. In addition to the periodic examination of the transformers regular testing of the isolation oil is the widely adopted technique to observe the operating situations and performance of the insulation oil inside the power transformers. The disadvantage of the above method is that it cannot grasp the dynamic state of transformer that a possible danger would take place anytime. This paper presents a diagnostic system includes a PC computer and an on-line gas detector to monitor a transformer located at Tainan Primary Substation. Through on-line monitoring the concentrations of the dissolved gases, the proposed diagnostic system offers a way to interpret the incipient fault causes. The artificial neural network (ANN) algorithm is employed to analyze the types of possible faults.