請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/88578
標題: | 快速電池健康度估測應用於18650鋰離子電池 Quick State of Health Estimation for 18650 Li-Ion Batteries |
作者: | 張育國 Yu-Kuo Chang |
指導教授: | 黃振康 Chen-Kang Huang |
關鍵字: | 電動農機,內電阻,電池健康狀態 (State of Health, SOH),電池電量狀態 (State of Charge, SOC),支持向量迴歸 (Support Vector Regression, SVR), Electric agricultural machinery,Internal resistance,State of Health (SOH),State of Charge (SOC),support vector regression (SVR), |
出版年 : | 2023 |
學位: | 碩士 |
摘要: | 隨著各方重視淨零碳排,農機電動化為重要的發展方向。其中電池與能源的管理相當重要,必須了解電池之壽命以及使用狀況,延長使用壽命。本研究旨在開發電池管理系統之評估電池健康狀態 (State of Health, SOH) 快速估算方法以及可用電量估測。在獲取電池充放電實驗之完整電壓、電流及溫度,使用Arduino Nano控制板,搭配電壓、電流以及溫度感測器製作監控裝置,同時將數據記錄於記憶卡中,完成資料收集。快速估算法藉由大電流放電,觀察電壓下降幅度,以快速估算法與等效電路模型計算電池之內電阻,不需要額外量測儀器,可在短時間得知電池之SOH。以快速估算法得知,大部分廠牌電池SOH為0% 時,其電池容量約為原始容量之60 %。使用快速估算法與儀器量測之內電阻計算SOH,其均方根誤差 (Root Mean Square Error, RMSE) 為3.0 %;使用等效電路模型計算出內阻與儀器量測之內電阻計算SOH,其RMSE為3.8 %。利用完整公開之電池資料集,經由訓練模型預測電池容量,預測結果之RMSE為178 mAh。在老舊電池之容量浮動約200 mAh的狀況,預測結果會較接近真實容量。真實容量和預測容量計算SOH,RMSE為5.7 %。因有完整的循環狀況,此方法適合用於估測老化情形較嚴重之電池的容量,進而估算電池之SOH。本研究對於開發電池管理系統之電池內電阻以及SOH之估算,希望對於未來電動農機或者田間充電站之充放電管理系統的發展,可以提供一些實驗經驗以及數據,並產生更多討論和研究想法。 Due to the increasing focus on achieving net-zero carbon emissions, traditional agricultural machinery is gradually transitioning to electric systems for operation. Effective battery and energy management is very important. It is essential to understand battery lifespan and usage conditions in order to prolong operational time and achieve sustainability goals. This research aims to develop a fast estimation method for evaluating State of Health (SOH) and estimating available energy in a battery management system. To obtain complete voltage, current and temperature data during battery charging and discharging experiments, Arduino Nano control board was utilized along with voltage, current and temperature sensors to create a monitoring device. The collected data was then recorded on a memory card for data collection purposes. In response to the agricultural environment, a fast estimation algorithm was proposed, which utilizes high-current discharging to observe voltage drop and calculate battery internal resistance without the need for additional measurement instruments. This method allows for the quick assessment of the battery's SOH within a short timeframe. According to this method, when the SOH of most battery brands is 0%, their capacity is approximately 60% of the original capacity. The root mean square error (RMSE) for estimating SOH using the fast estimation method and instrument measurement for internal resistance calculation is 3.0%. On the other hand, the RMSE for estimating SOH using the Internal resistance obtained through equivalent circuit model and instrument measurement is 3.8%. By utilizing publicly available battery datasets, a trained model was employed to predict battery capacity, resulting in a RMSE of 178. In cases where the capacity of older batteries fluctuates by approximately 200mAh, the predicted results are closer to the actual capacity. The RMSE for calculating SOH by comparing real capacity and predicted capacity is 5.7%. This method is particularly suitable for estimating the capacity of heavily aged batteries by considering complete cycling conditions, thereby providing an estimation of the battery's SOH. The objective of this research, which focuses on estimating battery internal resistance and SOH for the development of a battery management system, is to provide experimental experience, data, and generate more discussions and research concepts for the development of charging and discharging management systems in electric agricultural machinery or on-field charging stations. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/88578 |
DOI: | 10.6342/NTU202302466 |
全文授權: | 同意授權(全球公開) |
顯示於系所單位: | 生物機電工程學系 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-111-2.pdf | 2.84 MB | Adobe PDF | 檢視/開啟 |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。