以分類及迴歸法進行低壓需量契約用戶估測

Capactiy Estimation of Low-Voltage Contract-consumers Using Clustering and Regression

謝家偉(Chia-Wei Hsieh)

指導教授 ： 洪穎怡

中原大學/工學院/電機工程研究所/碩士(2004年)

https://doi.org/10.6840/cycu200400401

摘要

自教育普及以來，人們生活水準提高後，開始注重本身的權利問題及生活週遭的環境品質，尤其對電力公司置放在住家附近的電力設備排斥性極高，然而為了供應民生用電，這些設備又不得不存在的情形下，建築商及水電承裝業為因應市場的要求，會在電力公司營業規則上作技術性迴避電力設備的放置。
契約用戶對電力有特定的需求量，因此電力公司配電設計員會考量區域負載的發展，通常會新設置電力設備，尤其在都市商業區及學校附近商街，更為注重，但因用戶以低估契約量的技術性迴避，造成設計員的困惱，使得類似區域的電力設備常處於滿載或超載的運轉下，徒增維護、供電品質、設備壽命的問題，所以新設低壓契約用戶的負載估測，是本研究提出的解決方法。
本研究以台南市中區200戶商業低壓契約用戶為例，將用電行為應用聚類演算法（硬式C均值、相對硬式C均值、模糊C均值、相對模糊C均值）加以分群，觀察分群後的散佈圖，並以線性迴歸法求出各聚類演算法分群後的迴歸方程式，比較彼此的誤差平方和與迴歸平方和，另以53筆測試戶，進行迴歸估測，觀察估測值與實際值的誤差最大值、誤差最小值和變異數研判較佳的分類法迴歸模式，以做為日後對低壓契約用戶用電參考指標。

關鍵字

迴歸分析 ； 相對硬式C均值 ； 相對模糊C均值 ； 硬式C均值 ； 負載預測 ； 模糊C均值

並列摘要

Due to education prevailing, people improve living standard and seek for self-authority and life quality. People reject power-equipments surrounding their house, but these equipments supply power for their livelihood. Usually, building constructor and electrician circumvent Taipower’s rules to reject power equipment near consumers’ houses.
Engineers design and lay equipments, while contract-consumers need specific power in areas, e.g., shopping and culture/schooling districts. In order to circumvent this circumstance, some consumers usually underestimate measure of power-depletion. This will result in some problems, e.g., full load or overload operation, difficulty in equipment maintenance, power quality degradation and equipment life lessening. Therefore, the purpose of this thesis is solving the load estimation for both Taipower engineers and contract-consumers.
In this thesis, two hundred contract-consumers, who locate in the middle-district of Tainan city, were studied. HCM（Hard c-means）, AHCM（Alternative hard c-means）, FCM（Fuzzy c-means） and AFCM（Alternative fuzzy c-means）were used to cluster different contract-consumers. SSE（Sum of Squares due to Error）and SSR（Sum of Squares due to Regression）were examined after performing the line regression. Fifty-three consumer data were used for evaluating the performance of the proposed method by comparing the estimated and real values. Maximum error, minimum error and MS (Mean Square) were considered as measures. The best model can be used a tool for engineers for future design.

並列關鍵字

Regression analysis ； Alternative hard c-means ； Hard c-means ； Fuzzy c-means ； load forecasting ； Alternative fuzzy c-means

參考文獻

[3]K.l.Wu and M.S.Yang, “Alternative c-means clustering algorithms,”2001.

[8]胡裕仁, “模糊分類在醫學影像分割上的應用,” 中原大學數學系碩士論文,PP.10-12,中華民國九十年六月。

The Mcgraw-Hill Companies, Inc., pp.401-405, 2001.

[9]張榮芳, “電力用戶負載歸類及整合,” 中山大學電機工程學系研究

[10]I.Mogharm and S.Rahman, “Analysis and evalution of five short-term load forecasting techniques,” IEEE Trans. on Power Apparatus and Systems, Vol.4, No.4, pp.1484-1491, 1989.