有色金属(冶炼部分)

在全球能源转型与“双碳”战略背景下，锂离子电池的跨国贸易流向已成为决定未来全球电子废弃物地理分布与资源循环格局的关键变量。为突破传统基于产量统计或单一国家数据的预测局限，本研究构建了基于海关贸易大数据的混合驱动全生命周期预测模型。研究采集并清洗了2023–2024年约39万条全球关单数据，结合真实锚点与动态推算法构建覆盖2020–2030年的投入矩阵，并采用修正韦伯寿命分布进行退役演变模拟。结果表明：全球锂电池退役规模将从2023年约69万t增长至2035年约646万t；消费电子退役量于2031年达峰值约128万t，动力电池退役量持续攀升至2035年约611万t。2028年为全球电池回收市场结构性拐点，届时动力电池退役量将首次超越消费电子；2032年动力电池再生钴供给将首次超越消费电子，标志着全球再生钴供应格局的根本性转变。本研究揭示了不同应用场景电池退役的差异化演变规律，为有色金属回收企业的全球化战略布局提供了数据支撑。

Lithium-ion batteries serve as core energy storage components for electric vehicles and consumer electronics, and their cross-border trade flows have become critical variables shaping the future geographical distribution of global e-waste and resource recycling patterns. Traditional forecasting methods relied on macroeconomic production statistics or single-country vehicle registration data, which failed to capture the crossborder movement trajectories of battery products throughout their life cycles and systematically underestimated retirement potential in emerging markets. A hybrid-driven full life cycle prediction model based on customs trade data was developed to address these limitations. Approximately 390, 000 global customs declaration records from 2023 to 2024 were collected and processed through a multi-layer semantic filtering mechanism that enabled precise classification of consumer electronics and power batteries. A blacklist database containing chemical formulas such as perchlorate, carbonate, hydroxide, and acetate was established to eliminate lithium salt raw materials. Keywords including notebook, mobile phone, tablet, and handheld device were used to identify consumer electronics batteries, while terms such as power battery, battery pack assembly, electric vehicle, and hybrid were applied to classify power batteries. A weight median imputation algorithm was implemented for records with missing weight data. After processing, 220, 000 valid analytical samples were obtained from the original dataset, comprising approximately 87, 000 consumer electronics samples and 148, 000 power battery samples. A hybrid input matrix covering 2020–2030 was constructed by combining real anchor points from 2023–2024 customs data with dynamic extrapolation using compound annual growth rates. Consumer electronics batteries were estimated at 3% annual growth based on global smartphone and laptop shipment trends, while power batteries were back-calculated at 30% CAGR for historical years and projected at 10% for future years considering tariff barriers and localized production trends. Modified Weibull lifetime distribution parameters were applied for retirement evolution simulation, with shape parameter β=2.5 and characteristic lifetime η=4 years for consumer electronics batteries, and β=3.5 and η=10 years for power batteries. Weight correction coefficients of 0.75 and 0.85 were applied to consumer electronics and power batteries respectively to exclude packaging materials and non-active components. Model results indicated that global lithium battery retirement would grow from approximately 0.69 million tons in 2023 to 6.46 million tons by 2035, representing an 8.4-fold increase. Consumer electronics battery retirement was projected to peak at approximately 1.28 million tons in 2031 before declining, while power battery retirement continued climbing to approximately 6.11 million tons by 2035. The contribution of power batteries to total retirement increased from 9.3% in 2023 to 94.6% in 2035. The year 2028 was identified as a structural inflection point when power battery retirement(approximately 1.59 million tons) would exceed consumer electronics retirement(approximately 1.22 million tons), marking a transition from lightweight e-waste processing to heavy-duty power battery handling. In 2032, recycled cobalt supply from power batteries was projected to surpass that from consumer electronics, indicating a fundamental shift in global secondary cobalt supply structure. Theoretical recovery potential for cobalt was estimated to peak at approximately 409, 000 tons in 2033, while nickel recovery potential would reach approximately 751, 000 tons and lithium approximately 99, 000 tons by 2035. The differentiated evolution patterns of battery retirement across application scenarios and the temporal distribution characteristics of critical metal recovery potential were revealed through the analysis.