在當今的現代社會中，電池技術扮演了關鍵角色，應用範圍廣泛。包括電動車、可再生能源儲能系統以及移動設備。這些應用領域的共同特點是它們都依賴於高效且可靠的能量供應，而電池的電量狀態(State of Charge, SOC)是實現這一目標的關鍵因素之一。 以電動車為例，了解電池的SOC對於確保車輛的正常運行和提供預測的駕駛範圍至關重要，以避免電力耗盡。此外，電池的SOC也影響充電的頻率和時間，因此它關乎用戶的便利性和充電基礎設施的需求。 電池管理系統(Battery Management System, BMS)是一項專門用於監督電池的技術。BMS包含了各種感應器，在確保安全且高效利用電池內儲存的能源方面發揮著至關重要的作用。BMS的主要功能之一是估算電池的SOC。然而，由於鋰離子電池的電化學特性，使得這項功能具有相當大的挑戰性。鋰離子電池的SOC無法通過傳統感測器直接測量，但可以從電流和電壓等參數間接估算出來。本文提出了一個用於預測電動車上鋰離子電池SOC的深度神經網絡(Deep Neuron Network, DNN)模型。其目標是提高電池SOC估算的準確性以及穩定性。 本文的研究過程以Python進行演算法驗證，接著使用Verilog HDL進行模擬。隨後通過FPGA、Design Compiler和IC Compiler進行電路驗證。 ;Due to environmental pollution, electric vehicles (EVs) are trends in transportation. EVs produce zero tailpipe pollution and use less fuel than similar conventional vehicles. However, EVs still have some problems that should be addressed, such as lack of charging station, high cost of infrastructure, limited range or range anxiety. Consequently, the estimation of parameters related to electric vehicle batteries is a critical issue. Battery Management System (BMS) is a technology designed for the supervision of batteries. BMS plays a pivotal role in ensuring the safe and efficient utilization of energy stored within batteries, containing various sensors. One of the primary tasks of the Battery Management System (BMS) involves the estimation of the battery state of charge (SOC). Nevertheless, this task poses a considerable challenge due to the electrochemical nature of lithium-ion batteries. The SOC of lithium-ion batteries cannot be directly measured through conventional sensors, but it can be indirectly estimated from parameters such as current and voltage. This paper proposes a deep neural network model (DNN) for predicting the Li-ion SOC of EVs. The objective is to enhance accuracy and stability of battery SOC estimation. The research process of this paper begins with algorithm validation with Python, followed by simulation using Verilog HDL. Subsequently, circuit verification is conducted through FPGA, Design Compiler, and IC Compiler.
