This technology uses automatic speech recognition (ASR) and ultrasonic non-destructive testing (NDT) to accurately assess the state of health (SoH) and internal condition (electrolyte and by-product content) of the lithium-ion batteries of electric vehicles. The accuracy of SoH prediction exceeds 93%. The reliability and functionality of this technology surpass international commercial technologies. With the novel AI model and NDT method, this technology will be patented in various countries.

1. Novel ultrasonic transducers for lithium-ion batteries: The signal remains a high resolution after penetrating the battery. 2. Multiple detection and data synthesis: Presenting the detailed physicochemical properties of aged lithium-ion batteries. 3. Combination of AI speech recognition and ultrasonic non-destructive testing: Enhancing detection accuracy. 4. Integration of generative adversarial networks in the AI model: Reducing the time and number of samples required for AI model training.

1. Quality control in lithium-ion battery manufacturing: Identifying defective batteries, such as those with insufficient electrolyte or structural defects. 2. Optimizing the performance of the battery management system (BMS): Real-time monitoring of the SoH and internal condition of the lithium-ion batteries can enhance the safety and power management efficiency of electric vehicles. 3. Recycling of retired lithium-ion batteries: Identifying non-failed retired lithium-ion batteries for reuse.