Meaning
Mathematical technique used to calculate the rate of change of one battery variable with respect to another from discrete data points. Numerical differentiation converts raw voltage and capacity measurements into incremental capacity or differential voltage curves. These transformations reveal electrochemical peaks that are invisible in standard charge discharge profiles.
Data Smoothing
Raw experimental results often contain high frequency noise that produces large errors when calculating gradients directly. Successful numerical differentiation requires the application of filters or polynomial fits to the dataset before the derivative is taken. The choice of window size in a Savitzky Golay filter determines the balance between peak resolution and noise suppression.
Phase Identification
Peaks in the resulting dQ/dV or dV/dQ plots correspond to phase transitions or specific redox reactions within the active materials. By employing numerical differentiation, engineers can track the degradation of specific components like the graphite anode or the nickel rich cathode. This method provides a non destructive way to audit the internal health of a sealed cell.
Error Propagation
Small inaccuracies in voltage measurement lead to large artifacts in the derivative output. Errors propagate more aggressively in numerical differentiation when the sampling rate is insufficient to capture the peaks. Sampling precision matters.