Equivalent circuit models (ECM) are required to provide an on-board model of battery behaviour by battery management systems (BMS). The performance of ECMs is dependent on characterisation test results. The components of the ECM are commonly parameterised using electrochemical impedance spectroscopy (EIS) results, open circuit voltage (OCV) test results, and capacity test results. Therefore, these three tests are important for ECM parameterisation. Although the test procedures for these characterisation tests exist to test Li-ion cells for a range of applications e.g. portable electronic devices, they fail to provide essential information for automotive application due to the different requirements of vehicles (e.g. high power and energy, wide operating environment, long service life). This thesis reports research to improve consistency, accuracy and interpretation of characterisation test techniques for Li-ion battery cells for automotive application. The capacity of the battery pack is a vital parameter required for an ECM to estimate driving range. Existing techniques for predicting the driving range of an electric vehicle use the capacity value in Amp-hours, measured by existing capacity test techniques. In this thesis, experimental evidence that establishes the advantages of using capacity in Watt-hours instead of the capacity in Amp-hours as per the standard test is presented for the first time. Moreover, it is reported that measured battery capacity can vary by up to 5.0 % depending on the length of intermediate rest period. The OCV is another crucial parameter of ECM. The path dependence of OCV is a distinctive characteristic of Li-ion batteries which is known as OCV hysteresis. OCV test procedures used previously do not consider the initial conditions of the cells and capacity variations that show a change in OCV, leading to an apparent increase in, or erroneous, hysteresis. Using a new methodology which addresses issues mentioned above, OCV and OCV hysteresis has been quantified for different Li-ion cells for the first time. The test results show that a battery’s OCV is directly related to the discharge capacity, not the more commonly used SoC. The maximum hysteresis was found in a LiFePO4 (LFP) cell and lowest in a LTO cell, although still measurable. A dynamic hysteresis model is used to show how better OCV prediction accuracy can be achieved by a BMS when hysteresis voltage is a function of SoC instead of assuming it to be a constant, as traditionally done. EIS is commonly used to parameterise an ECM. For the first time this thesis reports that the time period between the removal of an electrical load and an EIS measurement affects the results. The study of five commercially available cells of varying capacities and electrode chemistries show that, regardless of the cell type, the maximum impedance change takes place within the first 4 hours of the relaxation period. Therefore a standardised relaxation period of minimum 4 hours should be allowed before performing EIS test. In addition to ECM parameterisation, EIS has been considered for online measurement, integrated with a BMS. This thesis concluded that the use of EIS as a fast measurement tool will be unreliable because of the relaxation effect. The flaws with capacity, EIS and OCV tests for automotive applications have been discussed. Through experimental evidence and electrochemical explanation it has been demonstrated that these tests can be made more consistent (e.g. by allowing fixed relaxation period in EIS test), have improved accuracy (e.g. incorporating hysteresis as a function of SoC) and better interpretation of test results (e.g. Watt-hours instead of Amp-hours in capacity test) are possible. Therefore, the overall contributions of this thesis to the scientific community are better consistency, accuracy and interpretation of these three tests. With the use of a case study, it has been shown that this new knowledge will improve performance of ECM, and thus BMS. This is not only for automotive but also more general applications through adopting the proposed new methodologies.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:682935 |
| Date | January 2015 |
| Creators | Barai, Anup |
| Publisher | University of Warwick |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://wrap.warwick.ac.uk/77676/ |
Page generated in 0.0016 seconds