In a context of an expected increasing use of Lithium-ion batteries in the transportation sector, Volvo AB is developing its own solutions for large electric vehicles. It is then beneficial to reduce the costs, the energy demand and the raw materials demand by improving the battery systems’ performances. For that purpose, understanding the physical phenomena which come into play in lithium ion cells is necessary. This project’s motivation has been to deepen the existing knowledge on one or a group of these phenomena which include those at the origin of the Open Circuit Voltage (OCV) hysteresis. It is characterized by the difference in charging and discharging voltage when the cell is at a resting state. These voltage differences might result in heat losses in the cells. In this thesis, the behaviour of the Open Circuit Voltage (OCV) under different operating conditions is studied, and a Preisach empirical hysteresis model is developed. The core part of the work consisted in experimental measurements of the Open-Circuit-Voltage of 10 Ah prismatic LiFePO4 (Lithium Iron Phosphate) cells. These measurements were completed using the Galvanostatic Intermittent Titration Technique (GITT) that consists of alternative current pulse and relaxation phases. The tests were performed using relaxation times ranging from 1 hour to 48 hours with the cells being under various cycles (series of charge and discharge). The impacts of the temperature, various current rates from 0.1C to 1C (1 A to 10 A) on the OCV and the voltage relaxation were evaluated. The amplitude of the OCV hysteresis that does not vanish after full relaxation, which is defined by the difference between the OCV charge and the OCV after discharge, was found to vary between 5 mV and 20-25 mV depending on the State-of-Charge of the cells. Two peaks are identified around 20-30% and 65-70% State-of-Charge. The measured OCV hysteresis with 24 hours relaxation is about half of the measured OCV hysteresis with 2-5 hours relaxation. The experiments also measured an apparent smaller OCV hysteresis when the magnitude of the current increases; this trend is to be verified after full relaxation. The temperature has an impact on the OCV which is averaging around ±0.2 mV/K. The analysis of the voltage relaxation behaviour described that at low temperatures and low C-Rates, the cells get closer to equilibrium voltage at a slower pace. In addition, a higher test time, characterized by longer relaxation times after each step and/or a larger number of steps within the same State of Charge (SOC) range, tends to increase the time required for the cell to reach an equilibrium. After completing the OCV measurements, a Preisach hysteresis model is developed based on the experimental results. The model predicts the OCV variations of an Lithium ferrophosphate (LFP) cell at ambient temperature when going through various charge and discharge cycles. Its estimated Root Mean Square Error (RMSE) is 3 mV, but the accuracy of the model could be partially confounded with measurement uncertainty. The main outcomes are a more accurate description of the voltage relaxation behaviour and a new estimation of the amplitude of the OCV hysteresis in LFP cells. / I en värld där det finns en förväntad ökning av användandet av litiumjonbatterier inom transportsektorn, utvecklar AB Volvo sina egna lösningar för stora elfordon som lastbilar. För att reducera kostnaderna, energibehovet och efterfrågan på råvaror, är det nödvändigt att förstå fysiska fenomen inom litiumjoncellerna eftersom det kan hjälpa till att förbättra systemens prestanda. Examensarbetets motivation är att fördjupa kunskapen om fenomenen vid uppkomsten av öppen kretsspänningshysteres inom litiumjärnfosfatceller. Denna hysteres definieras av skillnaden mellan öppen kretsspänning (Open-Circuit-Voltage eller OCV) under laddning och OCV under urladdning. Det orsakar över- och underspänning som ökar värmeförlusterna i litiumjoncellerna. Detta projekt studerar beteendet av både spänningsrelaxation och OCV för ett valt intervall av parametrar. Sedan utvecklas en Preisach empirisk modell. Huvuddelen av arbetet bestod i den experimentella mätningen av OCV av 10 Ah prismatiska LiFePO4 celler (Litiumjärnfosfatceller). Dessa experiment genomfördes medelst en mätprocedur som kallas Galvanostatic Intermittent Titration Technique eller Galvanostatisk intermittent titreringsteknik (GITT). Testerna innehåller växelvis strömpuls- och relaxationsfaser. Spänningsrelaxationsfaserna varade mellan 1 och 48 timmar under olika laddnings- och urladdningscykler. Inverkan av båda temperaturen och strömstyrkan (mellan 0.1C och 1C) på OCV utvärderades. Amplituden för OCV hysteresen som kvarstår efter full relaxation beräknades ligga mellan 5 mV och 20-25 mV beroende på cellersladdningstillstånd. Två hysterestoppar identifierades: en runt 70% och en andra mellan 20% och 30% laddningstillstånd. Hysteresen som mäts med 24 timmar av relaxation är runt hälften av hysteresen som mäts med två till fem timmar av relaxation. Med större strömstyrka är den uppmätta hysteresen lite lägre. Ytterligare tester bör göras för att verifiera att hysteresen fortfarande är lägre efter full relaxation. Temperaturen har en begränsad effekt på den totala hysteresen, men entropikoefficientensvärdet är i genomsnitt runt ± 0.2 mV/K. Analysen av relaxations beteende beskriver att en högre temperatur och strömstyrka ökar hastigheten med vilken jämviktspotentialen nås efter strömpulsen. Dessutom orsakaren ökning av den totala testlängden en långsammare relaxering. En längre testtid karaktäriseras av en längre relaxationstid efter varje strömpuls och/eller flera steg för laddningstillstånd. Efter OCV-mätningarna, byggdes en Preisach hysteresmodel med hjälp av de experimentella resultaten. OCV-variationer under olika laddnings och urladdnings cykler modellerades vid rumstemperatur med ett uppskattat minsta kvadratfel på cirka 3 mV. Modellen testades inte med ett begränsat antal cykler så den exakta noggrannheten behöver ytterligare verifieras för att få ner mätosäkerheten. Det huvudsakliga bidraget från detta examensarbete är uppskattning av amplituden för den hysteresen och beskrivningen av spänningsrelaxering efter olika strömpulser, såväl i längd som i amplitud.

Batterimodeller används för att representera batterier. För ändamål som batterihanteringssystem används idag främst empiriska modeller som representerar ett batteri med en motsvarande kretsmodell. Några nackdelar för dessa modeller ligger i dess oförmåga att simulera interna tillstånd och en tidskrävande parametriseringsprocess. Dessa nackdelar motiverar ingenjörer att vända sig till modeller som är baserade på fysiska lagar som ett alternativ eftersom de kan ge insikt i vad som händer inuti batteriet. Batterimodellerna som är baserade på de fysiska lagarna har alltför krävande beräkningar för att kunna användas för vissa applikationer, som batterihanteringssystem. Singel-partikelmodellen (SPM) är en fysikbaserad modell som används i detta avhandlingsprojekt. Syftet med projektet var att hitta en metod för att parametrisera SPM för nya kommersiella cylindriska HTPFR18650 1100mAh 3.2V litiumjärnfosfatceller. En litteraturundersökning och experiment användes för att extrahera parametervärdena. 17 parametrar valdes från litteraturundersökningen eftersom de kunde användas för att parametrisera modellen. Geometriska parametrar hittades genom en cellöppning. Tre typer av icke-destruktiva experiment som var inspirerade av litteraturen utfördes för att extrahera värden för de andra icke-geometriska parametrarna. Ett cykeltest med låg strömhastighet utfördes för att få en pseudo-OCV-kurva och för att extrahera kapacitetsrelaterade parametrarna. En känslighetsanalys genomfördes för galvanostatisk intermittent titreringsteknik testet (GITT) och pulstestet för de parametrar som var kopplade till transportoch kinetiska fenomen. Python matematisk batterimodellering (PyBaMM) användes för att simulera experimenten. Parametersamlingen Prada 2013 användes som standardvärden. Standardvärdena för de valda parametrarna ersattes av de värden som hittades genom experiment. Känslighetsanalysen visade att några av de valda parametrarna var känsliga för experimenten medan andra inte var det. Parametrarna extraherades genom fysiska relationer och genom att anpassa parametervärde för simuleringen så att den passar den experimentella datan under urladdningsförloppet. Värden för 14 av de 17 parametrarna extraherades i metoden. Den parametriserade modellen validerades mot två potentiella applikationer, en för ett batterielfordon och den andra för ett mild-hybridfordon. Den parametriserade modellen visade att den negativa partikelradien inte kan hittas med den föreslagna parametriseringsmetoden. Simuleringen visade sig också matchade den experimentella datan bättre under urladdning av cellerna jämfört till uppladdning. Flera förbättringar för framtida arbete har föreslagits, såsom att utvidgning av känslighetsanalysen, att erhålla OCV-kurvan från GITT istället för att använda pseudo-OCVkurvan, att använda strängare gränser vid kurvanpassningarna samt att skapa mer optimala tester för att extrahera parametervärdena. / Battery models are used to represent batteries. For purposes like battery management systems, empirical based models like the equivalent circuit models are widely used. These models have downsides regarding for example inability to simulate internal states and parametrization time that make engineers look at physics-based models as an alternative. The physics-based models are made up of physical relationships that offer insights into what is happening inside the battery. These are too computationally demanding to be used for certain applications, like battery managements systems. The Single Particle Model (SPM) is a physics-based model that is utilized in this thesis project. The aim of the project is to find a method to parametrize the SPM for fresh commercial cylindrical HTPFR18650 1100mAh 3.2V lithium iron phosphate cells. Literature survey and experiments were used to extract the parameter values. 17 parameters were selected from the literature survey since they could be used to parametrize the model. Geometrical parameters were found through a cell opening. Three types of nondestructive experiments inspired by literature were performed to extract values for the other non-geometric parameters. A low-rate cycling test was performed to get pseudo-OCV curve and to extract capacity related parameters. A sensitivity analysis is done for the GITT and the Pulse test for the parameters that were connected to the transport and kinetic phenomena. Python mathematical battery modelling (PyBaMM) was used to simulate the experiments. The Prada 2013 parameter set was be used as default values. The default values for the selected parameters were replaced by the values found through experiments. The sensitivity analysis showed that some of the selected parameters were sensitive while others were not. The parameters were extracted through physical relations and through curve fitting procedures during discharge. Values for 14 out of the 17 parameters were extracted in the method. The parametrized model was validated against two potential applications, one for a battery electric vehicle and the other for a mild hybrid. The parametrized model showed that the negative particle radius cannot be found through the proposed parametrization procedure. The simulation matched the experimental data better for discharging cells than charging cells. Several improvements for future work have been suggested such as extending the sensitivity analysis, obtaining the OCV-curve from GITT instead of low-rate cycling, having stricter bounds for the curve fitting as well as creating more optimal tests to extract the parameter values.

Direct ethanol fuel cells are of interest due to the high energy density, ease of distribution and handling, and low toxicity of ethanol. Difficulties lie in finding a catalyst that can completely oxidize ethanol and resist poisoning by intermediate reaction species. Degradation of the catalyst layer over time is also an issue that needs to be addressed. In this work, niobium doped-titanium dioxide (Nb-TiO2) is investigated as a platinum (Pt) support due to its increased resistance to corrosion compared to the common catalyst support, carbon. It has also been seen in the literature that TiO2 is able to adsorb OH and assist in freeing Pt sites by further oxidizing COad to CO2 and thereby increasing the catalytic activity of catalysts toward ethanol oxidation. The TiO2 support is mixed with carbon, forming Nb-TiO2-C, in order to increase the conductivity throughout the support. The electrochemical activity and direct ethanol fuel cell (DEFC) performance of this novel catalyst is investigated and compared to that of two common catalysts, carbon supported Pt (Pt/C) and carbon supported platinum-tin (PtSn/C). While the conductivity of the Pt/Nb-TiO2-C electrodes was low compared to that of the carbon supported electrodes, the overall catalytic activity and performance of the TiO2 supported catalyst was comparable to that of the Pt/C catalyst based on the electrochemically active surface area. / Master of Science

Battery Management System (BMS) is an essential part in battery powered applications where large battery packs are in use. BMS ensures protection, controlling, supervision and accurate state estimation of battery pack to provide efficient energy management. However the particular application determines the accuracy and requirements of BMS where it has to implement; in electric vehicles (EVs) accuracy cannot be compromised. The software part of BMS estimates the states of the battery pack and takes the best possible decision. In EVs one of the key tasks of BMS’s software part is to provide the actual state of charge (SOC), which represents a crucial parameter to be determined, especially in lithium iron phosphate (LiFePO4) batteries, due to the presence of the high hysteresis behavior in the open circuit voltage than other kind of lithium batteries. This hysteresis phenomena appears with two different voltage curves during the charging and discharging process. The value of the voltage that the battery is going to assume during the off-loading operation depends on several factors, such as temperature, loop direction and ageing. In this research work, hybrid method is implemented in which advantages of several methods are achieved by implementing one technique combined with another. In this work SOC is calculated from coulomb counting method and in order to correct the error of SOC, an hysteresis model is developed and used due to presence of hysteresis effect in LiFePO4 batteries. An hysteresis model of the open circuit voltage (OCV) for a LiFePO4 cell is developed and implemented in MATLAB/Simulink© in order to reproduce the voltage response of the battery when no current from the cell is required (no load condition). Then the difference of estimated voltage and measured voltage is taken in order to correct the error of SOC calculated from coulomb counting or current integration method. To develop the hysteresis model which can reproduce the same voltage behavior, lot of experiments have been carried out practically in order to see the hysteresis voltage response and to see that how voltage curve change with the variation of temperature, ageing and loop direction. At the end model is validated with different driving profiles at different ambient temperatures.

Rotating blades are considered to be the one of the most common cause of failures in rotating machinery. Blade failure modes normally occur as a result of cracks due to unexpected operating conditions, which are normally caused by accidents of foreign objects damage, high cycle fatigue, blade rubbing, blade root looseness, and degradation from erosion and corrosion. Thus, detection of blade faults has an important role in reducing blade related failures and allowing repairs to be scheduled for the machinery. This in turn will lead to reduction in maintenance costs and thus raise productivity and safety aspects of operation. To maintain vital components of rotating machines, such as blades, shafts, bearings and gear boxes, at optimal levels, detection of failures in such components is important, because this will prevent any serious damage that could affect performance. This research study involves laboratory tests on a small rig with a bladed disc rotor that applied vibration measurements and analysis for blade fault detection. Three measurements: shaft torsional vibration, on-bearing vibration (OBV) and on-casing vibration (OCV), are used. A small test rig of a single stage bladed disc holding 8-blades was designed and manufactured, to carry out this research study to assess the usefulness and capability of each vibration technique in detection of incipient defects within machine blades. A series of tests was conducted on a test rig for three different cases of blade health conditions: (a) healthy blade(s) with mistuned effects, (b) blade root looseness and (c) cracks in a blade on two different blade sizes (long and short blades) in order to discover changes in blades' dynamic behaviour during the machine running-up operation. The data were collected using the three measurements during machine run-up and then recorded. The measured vibration data were analysed by computing the blades' resonance at different engine orders (EOs) related to the blade(s) resonance frequencies and their higher harmonics, to understand the blade(s) dynamics behaviour for the cases of healthy and faulty blade(s). Data have been further processed using a polar plot presentation method which provides clear results that can be used for monitoring blade integrity. To validate the obtained experimental results, a simplified mathematical model was also developed. Finally, a comparative study between three methods was undertaken to understand the relative advantages and limitations in the blade heath monitoring.

Fuel cells are known for their high efficiency and have the potential to become a major technology for producing clean energy, especially when the fuel, e.g. hydrogen, is produced from renewable energy sources such as wind or solar. Currently, the two main obstacles to wide-spread commercialization are their high cost and the short operational lifetime of certain components. Polymer electrolyte membrane (PEM) fuel cells have been a focus of attention in recent years, due to their use of hydrogen as a fuel, their comparatively low operating temperature and flexibility for use in both stationary and portable (automotive) applications. Perfluorosulfonic acid membranes are the leading ionomers for use in PEM hydrogen fuel cells. They combine essential qualities, such as high mechanical and thermal stability, with high proton conductivity. However, they are expensive and currently show insufficient chemical stability towards radicals formed during fuel cell operation, resulting in degradation that leads to premature failure. The incorporation of durability improving additives into perfluorosulfonic acid membranes is discussed in this work. iv Cerium oxide (ceria) is a well-known radical scavenger that has been used in the biological and medical field. It is able to quench radicals by facilely switching between its Ce(III) and Ce(IV) oxidation states. In this work, cerium oxide nanoparticles were added to perfluorosulfonic acid membranes and subjected to ex-situ and in-situ accelerated durability tests. The two ceria formulations, an in-house synthesized and commercially available material, were found to consist of crystalline particles of 2 – 5 nm and 20 – 150 nm size, respectively, that did not change size or shape when incorporated into the membranes. At higher temperature and relative humidity in gas flowing conditions, ceria in membranes is found to be reduced to its ionic form by virtue of the acidic environment. In ex-situ Fenton testing, the inclusion of ceria into membranes reduced the emission of fluoride, a strong indicator of degradation, by an order of magnitude with both liquid and gaseous hydrogen peroxide. In open-circuit voltage (OCV) hold fuel cell testing, ceria improved durability, as measured by several parameters such as OCV decay rate, fluoride emission and cell performance, over several hundred hours and influenced the formation of the platinum band typically found after durability testing.