With the rapid adoption of electric vehicles globally, particularly in China’s electric vehicle market, the demand for reliable and efficient charging infrastructure has become paramount. As a key component in the energy transfer process, electric vehicle charging piles must exhibit high measurement accuracy and energy efficiency to ensure user trust and optimal resource utilization. However, existing metering technologies often suffer from limitations such as temperature drift, external interference, and inadequate calibration, leading to inaccuracies in energy measurement. Similarly, energy efficiency evaluation methods lack standardization, hindering comprehensive performance assessments. In this study, we address these challenges by proposing integrated solutions for improving measurement precision and developing a robust energy efficiency evaluation framework. Our approach combines hardware optimizations, advanced algorithms, and digital calibration strategies, alongside a dynamic evaluation model, to enhance the reliability and sustainability of electric vehicle charging systems. The proliferation of electric vehicle charging stations in China EV ecosystems underscores the urgency of these advancements, as they directly impact operational costs and environmental outcomes.
To begin, we focus on enhancing measurement accuracy through hardware-level improvements. The selection of sensors plays a critical role; we employ dual-Hall differential structures that counteract common-mode interference and temperature drift by arranging two Hall elements in opposite configurations. Additionally, magnetic shielding layers are incorporated to minimize external magnetic disturbances, thereby improving sensor stability. For signal conditioning circuits, we utilize low-temperature-drift operational amplifiers, such as the AD8554 with a drift coefficient below 1μV/℃, and design temperature-compensated reference voltage sources. To address precision in low-current ranges, a dual-range circuit with switchable gains is implemented, automatically shifting to high-gain channels when load currents fall below 20% of the rated value. In the analog-to-digital conversion stage, a 16-bit SAR-type ADC chip, ADS8885, is adopted for its high resolution, differential input capability, and strong anti-interference performance. The optimized hardware achieves a signal-to-noise ratio greater than 85dB, gain error below 0.02%, linearity under 0.01%, temperature drift coefficient less than 100ppm/℃, and common-mode rejection ratio exceeding 100dB, providing a solid foundation for accurate measurements in electric vehicle charging applications.
At the algorithm level, we develop three core modules: digital filtering, harmonic compensation, and adaptive calibration. Digital filtering employs an enhanced Kalman filter algorithm, which models the current signal’s state space to separate it from noise. By dynamically adjusting filter parameters based on current characteristics, this approach avoids the attenuation of dynamic signals typical of conventional filters, thereby improving filtering efficiency and response speed. Harmonic compensation relies on Fast Fourier Transform (FFT) analysis to extract the amplitude and phase of dominant harmonics, using a mathematical model to mitigate their impact and restore the integrity of the fundamental signal. The compensation is calculated using the following formula: $$I(t) = I_0 \sin(\omega t) + \sum_{n=2}^{\infty} A_n \sin(n\omega t + \phi_n)$$ where \(I_0\) represents the fundamental component, and \(A_n\) and \(\phi_n\) denote the amplitude and phase of the n-th harmonic, respectively. The adaptive calibration module continuously monitors environmental parameters, such as temperature, and dynamically adjusts the measurement system’s gain coefficients and zero offsets, significantly reducing errors induced by varying conditions. This holistic algorithm-based approach ensures that measurement precision reaches ±0.2%, as validated through extensive testing in electric vehicle charging scenarios.
For digital calibration, we implement a multi-point, multi-condition strategy that establishes a comprehensive calibration database covering full temperature and load ranges. This process includes two phases: factory calibration and online calibration. During factory calibration, conducted in a controlled temperature environment, high-precision power analyzers collect data at load points of 5%, 20%, 50%, 80%, and 100% of the rated capacity. A least-squares fitting method is then applied to derive a reference calibration curve. Online calibration leverages built-in temperature sensors and current detection circuits within the charging pile to acquire real-time operational parameters, dynamically adjusting measurement error compensation based on pre-stored calibration data. The calibration model is expressed as: $$V_{\text{corrected}} = a \cdot V_{\text{raw}} + b \cdot T + c$$ where \(V_{\text{corrected}}\) is the compensated value, \(V_{\text{raw}}\) is the raw measurement, \(T\) is temperature, and \(a\), \(b\), and \(c\) are coefficients determined through experimental fitting. Calibration data are stored in non-volatile FLASH memory to ensure integrity after power loss and facilitate ongoing optimization, contributing to the reliability of electric vehicle charging infrastructure in diverse environments.
Transitioning to energy efficiency evaluation, we construct a hierarchical indicator system based on an “input-output-loss” analysis framework. This system assesses efficiency across three dimensions: charging efficiency, conversion efficiency, and overall system efficiency. Charging efficiency, denoted as \(\eta_1\), reflects the energy transfer performance during the charging process, while conversion efficiency, \(\eta_2\), evaluates the effectiveness of power conversion units. The comprehensive指标体系 is summarized in the table below, which includes key metrics such as no-load loss rate, load loss rate, power factor, total harmonic distortion (THD), peak efficiency, and weighted efficiency. These indicators provide a multi-faceted view of energy performance, essential for optimizing electric vehicle charging piles in China EV applications.
| Evaluation Dimension | Specific Indicator | Calculation Method |
|---|---|---|
| Charging Efficiency | No-load Loss Rate | \(P_{\text{no-load}} / P_{\text{rated}}\) |
| Charging Efficiency | Load Loss Rate | \(P_{\text{loss}} / P_{\text{output}}\) |
| Charging Efficiency | Charging Efficiency \(\eta_1\) | \(\eta_1 = P_{\text{output}} / P_{\text{input}}\) |
| Conversion Efficiency | Power Factor (PF) | PF = Real Power / Apparent Power |
| Conversion Efficiency | Total Harmonic Distortion (THD) | THD = \(\sqrt{\sum_{n=2}^{\infty} (A_n / A_1)^2}\) |
| Conversion Efficiency | Conversion Efficiency \(\eta_2\) | \(\eta_2 = P_{\text{output}} / P_{\text{input-conv}}\) |
| System Efficiency | Peak Efficiency \(\eta_{\text{peak}}\) | Maximum \(\eta_{\text{system}}\) under optimal conditions |
| System Efficiency | Weighted Efficiency \(\eta_{\text{weighted}}\) | \(\eta_{\text{weighted}} = \sum (w_i \cdot \eta_i)\) for various load points |
In terms of evaluation methods and models, we adopt a layered modeling approach that encompasses power electronic conversion units, control units, and auxiliary systems. For the conversion unit, we analyze switching and conduction losses in IGBT/MOSFET power devices using the following models: $$P_{\text{switch}} = (E_{\text{on}} + E_{\text{off}}) \times f$$ $$P_{\text{conduction}} = V_{\text{CE}} \times I_{\text{C}}$$ where \(E_{\text{on}}\) and \(E_{\text{off}}\) represent the turn-on and turn-off energies, respectively, \(f\) is the switching frequency, \(V_{\text{CE}}\) is the collector-emitter voltage, and \(I_{\text{C}}\) is the collector current. The energy consumption of control units, including DSP/MCU, drive circuits, and sampling circuits, is quantified through lookup tables, while auxiliary systems account for devices like fans and displays. The total system loss is then expressed as: $$P_{\text{total loss}} = P_{\text{conversion loss}} + P_{\text{control loss}} + P_{\text{auxiliary loss}}$$ To enhance evaluation accuracy, we develop a dynamic model that incorporates factors such as temperature and load rate: $$\eta_{\text{system}} = \eta_0 + \alpha (T – T_0) + \beta (L – L_0)$$ where \(\eta_0\) is the reference efficiency under standard conditions, \(T_0\) is the reference temperature, \(L\) is the load rate, and \(\alpha\) and \(\beta\) are temperature and load coefficients, respectively. This model allows for real-time adjustments, ensuring precise energy efficiency assessments across varying operational scenarios in electric vehicle charging piles.
To validate our proposed methods, we designed an experimental test platform using a charging pile prototype with dual-Hall sensor improvements, rated at 60kW, output voltage range of 200–750V, and maximum output current of 200A. The experiment was divided into two parts: measurement accuracy testing and energy efficiency evaluation. For accuracy tests, data were collected at temperatures of -20℃, 25℃, and 50℃, with load levels of 10%, 20%, 40%, 60%, 80%, and 100% of the rated capacity. Each condition lasted 30 minutes, with a sampling interval of 1 second. Energy efficiency evaluation was conducted at 25℃, with load points increasing from 10% to 100% of rated power, and data recorded after 15 minutes of stabilization at each point. The results demonstrated that our optimizations achieved a measurement accuracy of ±0.2% and an energy efficiency evaluation precision of ±2%, confirming the effectiveness of our approach in real-world electric vehicle charging environments.
The energy efficiency evaluation results, as shown in the table below, reveal a nonlinear relationship between efficiency and load rate, with temperature exerting a significant influence on performance. Specifically, a temperature increase of 10℃ led to an average efficiency reduction of 0.5%, primarily due to changes in the switching characteristics of power devices. These findings highlight the importance of dynamic modeling for accurate assessments in the context of China’s expanding electric vehicle infrastructure.
| Load Rate (%) | Conversion Efficiency (%) | Charging Efficiency (%) | System Efficiency (%) | Power Factor | THD (%) |
|---|---|---|---|---|---|
| 10 | 88.5 | 90.2 | 87.8 | 0.92 | 3.5 |
| 30 | 92.3 | 93.5 | 91.8 | 0.95 | 2.8 |
| 50 | 95.1 | 95.8 | 94.5 | 0.98 | 2.2 |
| 80 | 96.0 | 96.5 | 95.2 | 0.99 | 1.8 |
| 100 | 95.5 | 96.0 | 94.8 | 0.99 | 1.5 |
In conclusion, enhancing measurement accuracy and energy efficiency evaluation for electric vehicle charging piles is crucial for advancing sustainable transportation, particularly in the rapidly growing China EV sector. Our integrated approach, combining hardware refinements, algorithmic innovations, and digital calibration, has proven effective in achieving high precision and reliability. The proposed energy efficiency指标体系 and dynamic model provide a standardized framework for performance assessment, facilitating better resource management and operational optimization. Future work should focus on refining calibration techniques for dynamic environments and expanding the applicability of evaluation models to accommodate complex scenarios, thereby supporting the continued evolution of electric vehicle charging infrastructure. As the adoption of electric vehicles accelerates, these advancements will play a pivotal role in ensuring energy efficiency and user satisfaction across global markets.