The Dynamic Z-Track algorithm enables improved resistance estimation and remaining capacity prediction for dynamic load applications. A comparison algorithm to quantify the improvement in remaining capacity is a compensated end of discharge voltage (CEDV) gauge algorithm. This algorithm uses compensation of the IR drop estimated from a new battery to predict the location when the battery terminal voltage reaches the end of discharge. The Dynamic Z-Track algorithm is able to track the resistance accurately to within a small error tolerance. For an aged battery, the resistance can increase by more than 50% near the end of the battery discharge, and a CEDV gauge underestimates the voltage drop due to load current.

The effect of the gauging algorithms on remaining capacity prediction is shown in Figure 6-1. The Dynamic Z-Track algorithm is able to track the resistance accurately as the battery ages, with a small error tolerance. The CEDV algorithm under-estimates the resistance of the aged battery significantly. For a 1C load shown in Figure 6-1, the Dynamic Z-Track algorithm predicts the SOC when the battery voltage reaches the minimum 3V threshold accurately. The CEDV algorithm does not predict the SOC when the voltage reaches 3V accurately. The algorithm has a 10% over-estimation of the battery remaining capacity versus Dynamic Z-Track. At higher load currents, the IR drop is more significant, so the remaining capacity error in CEDV is larger. For the test case shown, the overestimation error for a 1.75 C average load current is 60%, as seen in Figure 6-2.

The expected improvement in SOC accuracy depends on the battery behavior as the battery ages, and the details of the load current during discharge. For some load currents, there are frequent and long intervals of stable load current, so that an Impedance-Track based gauge algorithm can achieve similar performance to Dynamic Z-Track. A battery with a long lifetime shows a slow increase in resistance. The gap between Dynamic Z-Track and fixed-resistance estimates of remaining capacity is small until the battery ages enough that the resistance shows a significant increase.

Batteries that show rapid increase in resistance can have even larger errors in SOC estimation with dynamic loads than the example in Figure 6-2.