Optimal Supervision And Use of Battery Systems
This project is about developing mathematical models and software for automatic battery testing.
The dynamic behavior of batteries can be predicted using theoretical cell model for basic processes. The model is calibrated for two types of valve-regulated lead-acid batteries (gel-cell and AGM batteries), and is applied for viewing the unobservable processes in a battery by observable processes. The unobservable parameters like overpotential, reaction rate, porosity, acid concentration, and other parameters of electrode can be evaluated by total current, terminal voltage and temperature of surrounding atmosphere of battery. The calibrated model is applied to distinguish between outwardly equal batteries with different backup time and cut-off time. It is shown that
- difference in morphology of electrodes,
- thickness of electrodes and
- quantity of electrolyte in separator
are the main distinguishing parameters between batteries. These parameters are tested online by current-voltage measurements using fast calculation method proposed in this project.
A theoretical cell model is applied for evaluation of the valve-regulated lead-acid batteries under charge, discharge and overcharge conditions; it is modified, calibrated and tested against experimental data. The model is modified in several ways. A new formula (more non-linear) for electrode morphology is used. The charging factor is applied in the state-of-charge model, electrode porosity and acid concentration models. The recombination of oxygen is considered as a mass-transport limited evolution process. A new model is proposed for resistance of battery. The modified cell model is calibrated on test batteries. The model prediction accuracy can be improved for a full range of charge-discharge processes, including deep discharge and overcharge of battery. The behavior of single batteries in string is different during overcharge, their individual behavior can be predicted. The modified cell model is applied for evaluating the unobservable resistance and gas formation processes by observable current-voltage and temperature measurements. Battery monitoring by resistance and gas formation processes during overcharge is analyzed.
The battery model is implemented in software designed to be integrated in a telecommunications UPS system as a prototype for battery monitoring software. The software is for failure detection and lifetime forecasting of batteries.
- Keywords: battery monitoring, simulation, state estimation, state-of-health, VRLA battery
- Duration: 2000-2001
- Research area: Power Electronics & Electric Machines, Operations & Maintenance
- Publications (total: 11)
- Control Engineering Laboratory (TKK)
- Electronics Laboratory (University of Oulu)
- Ascom Energy Systems
- Nokia Networks
- Teuvo Suntio
- Ander Tenno
- Robert Tenno
- Ander Tenno: Modelling and evaluation of valve-regulated lead-acid batteries (2004)