دانلود مقاله تحلیل و مدیریت داده ها برای استفاده بهینه از الگوریتم پیشرفته خطا یابی در شبکه
چکیده
مقدمه
روش مکان یابی خطا
مطالعه موردی
تجزیه و تحلیل میزان حساسیت
نتیجه گیری
منابع
Abstract
Introduction
Fault location method
Case study
Sensitivity analysis
Conclusions
References
چکیده
از آنجایی که نیاز به خودکارسازی طرح های تشخیص عیب شبکه برق در حال افزایش است، استفاده از فناوری هایی مانند هوش مصنوعی (AI) می تواند راه حل های عملی برای مشکل ارائه دهد. هوش مصنوعی می‌تواند بر چالش‌هایی که توپولوژی‌های پیچیده مانند شبکه‌های هوشمند ولتاژ پایین (LV) ایجاد می‌کنند غلبه کند و ثابت کند که ابزار قدرتمندی در توسعه روش‌های پیشرفته تشخیص عیب است. یک پارامتر مهم برای موفقیت هر روش مبتنی بر هوش مصنوعی، کیفیت داده است. بنابراین، در این مقاله تجزیه و تحلیل داده‌ها به منظور ارزیابی نوع داده‌های تولید شده توسط یک شبکه LV کوچک و پاسخ الگوریتم هوش مصنوعی به آن‌ها انجام می‌شود. در چارچوب این تحلیل، مهمترین ویژگی ها و مترها شناسایی شدند. علاوه بر این، به عنوان پاسخ به حجم زیاد داده های موجود، یک استراتژی مدیریت داده پیشنهاد شده است. استراتژی ترکیبی از ویژگی های اصلی و تغییر شکل یافته است. برای این منظور، پنج روش کاهش بعد آزمایش و مقایسه شده است. Truncated-SVD مناسب‌ترین تلقی می‌شود و متعاقباً برای شکل‌دهی مجدد مجموعه داده‌ای که به مدل مکان خطای XGBoost معرفی می‌شود، استفاده می‌شود. ادغام تکنیک کاهش ابعاد در الگوریتم منجر به کاهش زمان محاسباتی و اندازه مجموعه داده و تعمیم پذیری بیشتر الگوریتم می شود. بنابراین، کاربرد روش پیشنهادی محدود به توپولوژی شبکه نیست. استحکام روش در برابر پارامترهای تأثیرگذار مختلف مانند مقاومت خطا، اندازه مجموعه داده، از دست دادن داده ها و سطح نفوذ فتوولتائیک تأیید شد. الگوریتم کلی در هنگام آزمایش بر روی شبکه معیار CIGRE LV به میانگین مربع خطای 13.26 و دقت آموزش و آزمون بیش از 99 درصد دست یافت.
توجه! این متن ترجمه ماشینی بوده و توسط مترجمین ای ترجمه، ترجمه نشده است.
Abstract
As the need for automatization of the electricity grid’s fault diagnosis schemes is rising, the application of technologies such as the artificial intelligence (AI) can provide practical solutions to the problem. AI can overcome the challenges that complex topologies like those of the low voltage (LV) smart grids pose and prove to be a powerful tool in the development of advanced fault diagnosis methods. An important parameter for the success of any AI-based method is the quality of data. Therefore, in this paper a data analysis is performed in order to evaluate the type of data produced by a small LV grid and an representative AI algorithm’s response to those. In the context of this analysis, the most important features and meters were identified. Furthermore, as a response to the large volume of available data, a data management strategy is proposed. The strategy combines original and reshaped features. For this purpose, five dimensionality reduction methods are tested and compared. Truncated-SVD is deemed the most appropriate and is subsequently utilized for the reshaping of the dataset that is introduced to the XGBoost fault location model. The integration of the dimensionality reduction technique in the algorithm results in the decrease of the computational time and the dataset’s size and in a higher generalizability of the algorithm. Thus, the application of the proposed method is not limited by the grid’s topology. The method’s robustness was verified against various influencing parameters such as the fault resistance, the size of the dataset, the loss of data and the photovoltaics’ penetration level. The overall algorithm achieved a mean squared error of 13.26 and a training and test accuracy of more than 99% when tested on the CIGRE LV benchmark grid.
Introduction
The transformation of the traditional electricity grids into smart grids is well underway, mandating the redefinition of the grid operation principles. One of the vital operating parts of the grid requiring redesign is the protection system and more specifically the fault diagnosis schemes, since the bidirectionality of power flows and the intermittency of generation sources pose additional challenges. Fault diagnosis refers to the detection, classification and location of a fault. Rapid and automatized fault diagnosis leads to increased reliability of the electricity grid, aligned with the needs of the modern society. With the vast changes in the grid’s topology, the traditional fault diagnosis methods have become outdated and inefficient. Therefore, the necessity for novel accurate and fast fault diagnosis methods has soared.
Conclusions
In this study a novel artificial intelligence (AI) – based fault location method for low voltage grids is presented. The transformation of traditional electricity grids to smart grids has rendered most of the conventional fault location methods obsolete, however, it has also offered opportunities for innovation due to the increased observability over the The proposed algorithm aims at solving these problems by optimizing the application of AI in a fault location method. This is achieved by evaluating the collected data and developing a data management strategy. More specifically, first, the data analysis results point to the form with which the recorded variables should be included in the dataset. The dataset is then processed following the data management strategy proposed. This part of the algorithm reduces the data volume, and thus the CT of the method, by transforming the least informative features with the use of the Truncated-SVD technique while keeping the 10 most informative features in their original form. This step results in the efficient exploitation of all the available data and, at the same time, the generalization of the algorithm. The reduced dataset is used as an input for the training of an XGBoost model, which combines low overfitting with high computational speed and accuracy. The final algorithm is characterized by superior performance, with a mean squared error of 13.26 and a training and testing accuracy above 99% when evaluated with data generated from the simulation of the CIGRE European LV benchmark.