Session
Technical Session XII: Advanced Subsystem or Component Developments II
Abstract
The most common battery charging approaches for nickel-hydrogen batteries suffer from two key limitations. They either consistently under- or overcharge the battery or they require accurate knowledge of the battery capacity. For the ORBCOMM constellation of 36 small communications satellites, regular overcharging is unacceptable because it decreases battery life. Worse, the lost life occurs when the satellite constellation is facing its greatest usage demand. On the other hand, obtaining an accurate knowledge of the battery capacity requires an upfront effort to develop the necessary engineering tools and characterize battery performance. In addition, the satellite operator must bear the recurring expense of continuously estimating the capacity over the 5-year design life of each spacecraft. To avoid these limitations, the ORBCOMM system employs a patented battery charging approach that estimates the battery charging efficiency by computing the amount of charge received by the battery as a function of the input charge. The full point of the battery is easily determined since the charge efficiency drops as the battery charge approaches its capacity. The battery charge is estimated by computing the moles of H2 gas in the cell; the input charge is simply the integral of the charge current. So, the charge efficiency is proportional to the rate of change of the moles of H2 gas with respect to the input charge, or dM/dC. This maintenance-free approach enables the charging routine to perfectly fill the batteries-without overcharging-independent of the battery capacity. Any algorithm based upon a derivative estimate poses a noise rejection challenge. The dMldC approach is no different. The battery charge control system must be adept enough to respond to decreases in dM/dC yet stable enough to avoid the inherent signal to noise problem at low charge rates. The effectiveness and pitfalls of the dM/dC battery charge algorithm are highlighted using data from the 26 ORBCOMM constellation spacecraft launched between December 1997 and September 1998.
A Maintenance-Free Battery Charge Control Approach for Nickel-Hydrogen Batteries
The most common battery charging approaches for nickel-hydrogen batteries suffer from two key limitations. They either consistently under- or overcharge the battery or they require accurate knowledge of the battery capacity. For the ORBCOMM constellation of 36 small communications satellites, regular overcharging is unacceptable because it decreases battery life. Worse, the lost life occurs when the satellite constellation is facing its greatest usage demand. On the other hand, obtaining an accurate knowledge of the battery capacity requires an upfront effort to develop the necessary engineering tools and characterize battery performance. In addition, the satellite operator must bear the recurring expense of continuously estimating the capacity over the 5-year design life of each spacecraft. To avoid these limitations, the ORBCOMM system employs a patented battery charging approach that estimates the battery charging efficiency by computing the amount of charge received by the battery as a function of the input charge. The full point of the battery is easily determined since the charge efficiency drops as the battery charge approaches its capacity. The battery charge is estimated by computing the moles of H2 gas in the cell; the input charge is simply the integral of the charge current. So, the charge efficiency is proportional to the rate of change of the moles of H2 gas with respect to the input charge, or dM/dC. This maintenance-free approach enables the charging routine to perfectly fill the batteries-without overcharging-independent of the battery capacity. Any algorithm based upon a derivative estimate poses a noise rejection challenge. The dMldC approach is no different. The battery charge control system must be adept enough to respond to decreases in dM/dC yet stable enough to avoid the inherent signal to noise problem at low charge rates. The effectiveness and pitfalls of the dM/dC battery charge algorithm are highlighted using data from the 26 ORBCOMM constellation spacecraft launched between December 1997 and September 1998.
