Measuring Lithium-ion Polymer Cell Internal Resistance

Share Embed Donate


Short Description

This final year project looks at a simple method for the measurement of single cell LiPo battery internal resistance mea...

Description

Aerospace Engineering.

Aerospace Engineering Individual Investigative Project

Designing, Building and Testing a Lithiumion Polymer Battery Charger With State of Health Monitoring

James E Stott May 2016

Supervisor: Professor David Stone Dissertation submitted to the University of Sheffield in partial fulfilment of the requirements for the degree of Bachelor of Engineering

Abstract The explosive development of lithium-ion polymer (LiPo) batteries has revolutionised the way we interact with technology not previously portable. In particular, remote control vehicle users have unanimously adopted LiPo batteries as their main power source. Despite their widespread use, the largely hobbyist nature of RC users has not proved conducive to the adoption of battery health measurement techniques commonly used in commercial applications that enable informed battery replacement decisions. Consequently, RC vehicle users are incurring extra cost with unnecessary battery replacement due to the inability to decide definitively when to replace them. To minimise waste and cost, a method of quantifying the state of health of RC vehicle batteries enabling informed battery replacement is of interest to users. Internal resistance – the largest variable affecting battery performance – is the strongest indicator of state of health and is therefore the main report consideration. The report aim was to develop a system for RC vehicle users - governed by user requirements for low cost and ease of use - which facilitates informed and unambiguous battery replacement decisions. In order to achieve this, a circuit to measure the internal resistance of a single cell LiPo battery with multiple cell LiPo scalability was produced. With future development the circuit has the potential to yield state of health information and integrate with current hobbyist charging products. The project hypothesis states that with successive cell cycles, the internal resistance will increase. Testing and evaluation of the circuit revealed large variations in cell internal resistance measurements, predominantly due to unstable ambient temperatures during the testing procedure, rendering most data unusable. Despite the test data inconsistencies, trend lines indicate a gradual increase in cell internal resistance with cycle number, permitting the tentative conclusion that the results are congruent with the hypothesis. The circuit will therefore fulfil the project aim with further development of the state of health quantification feature and does meet the user requirements for low cost and ease of use. Overall, the project provides some promising results that indicate the aims and objectives will be fulfilled with further testing and development of the SOH quantification feature to enable better informed battery replacement decisions.

Contents Acknowledgements

i

1.

INTRODUCTION

1

1.1

Background

1

1.2

Motivation

1

1.3

Aims and Objectives

1

1.4

Hypothesis and Solution

2

1.5

Report Overview

2

2.

3.

4.

5.

6.

7.

LITERATURE REVIEW

4

2.1

The Impact of Internal Resistance

4

2.2

Internal Resistance Components

4

2.3

Internal Resistance Measurement Products

6

2.4

Internal Resistance Measurement Techniques

6

2.5

Battery Charging

8

2.6

Circuit Design and Operation

9

2.7

Hypothesis

10

LITERATURE REVIEW AND RESEARCH CONCLUSIONS

11

3.1

Literature Review Conclusions

11

3.2

Research Conclusions

12

DC ‘DUAL PULSE’ TECHNICAL DESCRIPTION

13

4.1

DC ‘Dual Pulse’ Theory

13

4.2

Method of Internal Resistance Measurement

14

CIRCUIT SPECIFICATION

15

5.1

User Specification

15

5.2

Operation Specification

15

CIRCUIT DESIGN

16

6.1

Component Selection

16

6.2

Software Design

17

6.3

Circuit Diagram

19

CIRCUIT TESTING

20

7.1

Testing Considerations

20

7.2

Testing Method

21

8.

TEST RESULTS AND EVALUATION

22

8.1

Test Results

22

8.2

Test Results Evaluation

24

9.

CONCLUSION

26

10.

FUTURE DEVELOPMENT

27

10.1

State of Health Quantification

27

10.2

Charger Integration

27

10.3

Temperature Control

27

PROJECT MANAGEMENT

28

11.1

Gantt Charts

28

11.2

Project Progress

28

11.

12.

13.

SELF REVIEW

30

12.1

Project Progress

30

12.2

Personal Development

30

REFERENCES

31

Acknowledgements I would like to offer my sincere thanks to my project supervisor, Professor David Stone, for his continued and boundless support throughout this project, including his provision of sound careers and further study advice. His reassuring manner has proved invaluable for re-focussing on objectives and re-directing efforts in times of confusion. Thanks are also due to my second project supervisor, Dr Daniel Gladwin, for kindly providing his signature for part orders and advice in David’s absence.

i

1

INTRODUCTION

1.1 Background The invention of the lithium-ion battery in the 1970s (Whittingham, M.S., 1976) never looked set to transform the portable electronics field. It wasn’t until research into stable electrode materials at The University of Oxford in 1979 (Mizushima, K. et al., 1980), that their use in consumer electronics became a key focus. In 1991, Sony and chemical company Asahi Kasei released the first commercial lithium-ion battery (Sony Energy Devices Corporation, 2016), laying the foundation for a revolution in portable technology.

1.2 Motivation A major use of lithium battery technology has emerged among amateur RC (remote control) vehicle users. LiPo (Lithium-ion Polymer) batteries are the predominant power source for RC planes, cars and UAVs (Unmanned Aerial Vehicles) due to their high energy densities, large discharge currents and low weight. The phenomenon of decreasing battery performance with successive charge cycles is recognised by RC vehicle users, but is rarely quantified, making battery replacement decisions largely ambiguous resulting in unnecessary expense and waste. Clearly, this presents a problem to the user. Battery SOH (State Of Health) is a health metric encompassing parameters that contribute to performance reductions and represents the condition of an ageing battery compared to a new one. A key component of battery SOH is internal resistance, which directly affects performance criteria such as discharge current and power transfer efficiency. A method of quantifying battery SOH through internal resistance measurement is of interest to RC users to enable unambiguous replacement decisions, minimising cost and waste.

1.3 Aims and Objectives The aim of the project was to develop a battery charging system for RC vehicle users that facilitates informed and unambiguous battery replacement decisions. The objectives of the project were to design and build a LiPo charging circuit to charge and measure the internal resistance of single cell LiPo batteries, and to test and evaluate the circuit performance. A final objective was to demonstrate that the system is scalable to function with multiple cell batteries and that with further development it is capable of producing SOH readings. The

1

project aims and objectives are defined by user requirements for low cost and ease of use as well as technical and time constraints limiting the breadth of research and development.

1.4 Hypothesis and Solution The project hypothesis states that with successive cell cycles, the internal resistance will increase to a point greater than the manufacturer defined limit, rendering the battery unable to deliver the expected performance. The solution is based upon a widely studied battery internal resistance measurement method called the DC ‘dual pulse’ method that utilises Ohm’s law, and was recommended for its mathematical simplicity and ease of implementation. The circuit comprises a microcontroller and power electronics to measure cell internal resistance using the DC ‘dual pulse’ method, subjecting a battery to varying loads, measuring the voltage drop and calculating the result using Ohm’s law.

1.5 Report Overview The report opens with a literature review that compares and contrasts various literary sources relevant to the project in order to identify past developments that can be utilised in achieving the aims and defining the scope of the project. Following on, chapter 3 evaluates the findings of the literature review and clearly identifies elements of previous works that will be used in the project. Chapter 4 describes the theory and operation of the DC ‘dual pulse’ method and highlights some considerations for the circuit design, build and test stages. Chapter 5 defines a circuit specification to provide an easy means of evaluating the circuit test results through comparison with the circuit specification. Chapter 6 guides the reader through the circuit design process, introduces the circuit diagram and shows the program code from which the reader can gain a technical understanding of the circuit operation. In chapter 7, considerations during the test procedure are outlined and followed by a description of the circuit testing method. Chapter 8 lays out the circuit test results in graphical form and analyses them for reliability, accuracy and congruency with the circuit specification, aims and hypothesis.

2

Spurious test data results are identified and their respective causes highlighted with analysis into the sources of uncertainty and how they manifest themselves in the test data. Chapter 9 provides a project conclusion before chapter 10 suggests areas for future development to fully achieve the project aims and objectives, as well as suggesting methods of reducing the uncertainty in internal resistance measurements. Chapter 11 delivers an analysis of the project management and contrasts expected with actual progression, allowing conclusions to be drawn as to the effectiveness of time management and planning techniques. Chapter 12 gives a self review of the performance and personal developments made during the project. Finally, chapter 13 sets out the references used throughout the report.

3

2

Literature Review

The following chapter analyses existing internal resistance research and technical articles, and discusses their relationship to the investigation in order to justify the report aim and define the report scope. Throughout this chapter, the review is grouped into themes with appropriate headings to effectively identify both similar and conflicting research connections.

2.1 The Impact of Internal Resistance Typically, designers look to maximise the power transfer efficiency between source and load to minimise the required battery capacity and hence weight. Internal resistance reduces power transfer efficiency and is therefore an important consideration in RC vehicle design. Linden (2010) describes that, “The power capability of a battery is determined by its cell voltage and its impedance” (p. 24.68). At this point, it is important to make clear the difference between internal impedance and resistance. Internal impedance is a measure of the opposition to current flow when a battery is subjected to a complex AC load, whereas internal resistance is the equivalent when subjected to a simple DC load. As RC vehicles present simple DC loads, internal resistance constitutes the main opposition to current flow in the RC application of LiPo batteries and will remain the measurement of interest throughout the report. Physically, internal resistance limits the flow of energy from the battery to the load due to Ohm’s Law, which in turn curtails the deliverable power and reduces the power transfer efficiency. Buchmann (2001) highlights the impact of internal resistance on high current delivery, reinforcing the importance of minimising internal resistance; “A battery with low internal resistance can deliver high current on demand; a battery with high internal resistance cannot deliver high current” (p. 109). An article discussing the prediction of battery performance using internal resistance states that, “The resistance of the internal circuit path is what influences the performance of a cell and is therefore the important parameter that needs to be measured” (Albér, n.d., p. 2).

2.2 Internal Resistance Components Resistance or impedance – a measure of the opposition to current flow in a circuit –seems a simple concept. However, in the context of a battery, it becomes a topic requiring greater consideration due to multiple resistance and impedance components. Using the assumption that battery total resistance obeys Ohm’s law and is independent of frequency, internal

4

resistance can be used in battery SOH measurement. Although this is an appropriate assumption for the DC applications in question, later chapters discuss measurements made using complex AC loads, yielding impedances as opposed to resistances. According to Buchmann (2001), “A battery as a power source combines ohmic, inductive and capacitive resistances” (p. 150), which when subjected to a complex load will exhibit an internal impedance. The various resistive components of a battery can be visualised using an equivalent electrical circuit called the Randles model shown in figure 1. For complex AC loads, each resistive component must be considered, as opposed to only the ohmic resistance RO for DC loads. Internal

impedance

aside,

the

internal resistance of a battery is comprised of both ohmic and electrochemical

resistances,

combining to yield the total TER (Total

Effective

Resistance).

Figure 1 The Randles model showing the various According to Albér (n.d.), the total resistances, impedances and capacitances of a typical conductance path through a battery battery. Buchmann (2001). (P. 150). includes, “the metallic or ohmic RO = ohmic resistance ZW = Warburg Impedance QC = constant phase loop Rt = transfer resistance path, as well as the path that is L = inductor involved electrochemically” (p. 2), where each has an associated resistance. A technical bulletin produced by Energizer (2005) and Linden (2010) define the resistances as the electronic and ionic resistances respectively, which combine to yield TER. In a technical article on internal resistance measurement by Schweiger et al. (2010), ohmic resistance is defined as the resistivity of battery components including anode and cathode materials, current collectors and electrolyte. Energizer (2005) attributes ionic (electrochemical) resistance to electrochemical factors including electrolyte conductivity, ion mobility and electrode surface area. This is in line with the definitions found in Linden (2010), Albér (n.d.) and Buchmann (2001). The Energizer (2005) technical article makes reference to the different development times of electronic (ohmic) and ionic (electrochemical) resistances during discharge, which will become an important consideration when analysing internal resistance measurement methods later in the report.

5

2.3 Internal Resistance Measurement Products Despite the requirement for low internal resistance to maximise power delivery in RC applications, a poor range of consumer measurement products is available. A hobby website by Battery University (2011) presents several methods of internal resistance measurement with accompanying theory, but fails to address the accuracy and reliability of the methods so important to making good estimates of battery SOH. This is in keeping with many other hobbyist sources located on the Internet. Turnigy – a major manufacturer of hobbyist batteries and chargers – produces an ‘all in one’ LiPo battery checker that includes an internal resistance measurement capability. This represents the type of unit a typical RC hobby user might purchase to measure battery SOH. The accompanying Turnigy (n.d.) data sheet stresses, “there is no perfect test for cell internal resistance” (p. 9). It states “internal resistance can vary considerably even within the same cell, this makes the task of determining the internal resistance a lot harder” (p. 9). The Turnigy (n.d.) data sheet identifies that internal resistance is a function of variables including cell temperature, charge state and age, highlighting the difficulty of accurately and reliably measuring battery internal resistance. As the cell temperature decreases, the internal resistance increases, becoming particularly poor around zero degrees Celsius. The lower the cell charge, the higher the internal resistance, which also increases with cell age. These variables and their relevance to the project will be considered in later sections.

2.4 Internal Resistance Measurement Techniques According to Buchmann (2001), “several methods of measurement are available of which the most common are applying DC loads and AC signals. Depending on the level of capacity loss, each technique provides slightly different readings” (p. 111). Albér (n.d.) also highlights the range of internal resistance measurement techniques available – “Instruments presently available use either an AC current injection method or a momentary load test” (p. 3). According to Buchmann (2001) and Schweiger et al. (2010), applying an AC current to a battery reveals the voltage and current phase shift, allowing internal impedance to be calculated and battery SOH to be quantified. EIS (Electrochemical Impedance Spectroscopy) is an application of the AC current injection method. EIS “is an effective method of analysing the mechanisms of interfacial structure and to observe the change in formation when cycling the battery as part of everyday use” (p. 202) according to Buchmann (2001). Schweiger et al. (2010) highlights the wide range of battery

6

parameters measurable with EIS - “EIS can provide detailed information of the cell under examination; parameters such as corrosion rate, electrochemical mechanisms, battery life and of course internal resistance” (p. 5609). The Energizer (2005) technical bulletin describes EIS as “an impedance test across a range of frequencies to portray internal resistance accurately” (p. 1). Typical RC vehicle users rarely cycle batteries daily, diminishing the usefulness of EIS to measuring the SOH of amateur RC vehicle batteries. Internal impedance measurements obtained through EIS are highly dependent upon battery characteristics and are not universal according to Buchmann (2001) – “Each battery type generates its own set of signatures, and without a library of well defined reference readings with which to compare the measurements, EIS has little meaning” (p. 203). For EIS to be an effective method of RC vehicle battery SOH measurement, a large library of battery reference readings must exist. Despite the in depth battery analysis EIS offers, it requires extensive measurement equipment and is generally time consuming according to Schweiger et al. (2010). Finally Albér (n.d.) highlights several problems with AC measurement methods – “The problem with AC measurements is that they are susceptible to charger ripple currents and other noise sources. Some instruments cannot be used while the battery is on-line” (p. 3). In order to mitigate the effects of noise on internal resistance measurement, Albér (n.d.) offers the DC load method as an alternative to AC methods due to its use of A/D convertors capable of ignoring AC signals flowing through the battery. “A common method of measuring internal resistance is the DC load test which applies a discharge current to the battery whilst measuring the voltage drop” (p. 145) according to Buchmann (2001). The voltage drop method “is a fast and convenient method for the measurement of internal resistance” (p. 5623) according to Schweiger et al. (2010). One DC method effectively shorts the battery to deliver maximum current for a very short period of time, allowing an approximation of internal resistance to be made according to Energizer (2005) and Linden (2010). However, Linden (2010) states that the ammeter resistance must be extremely low, no more than 10% of the battery internal resistance. Alternatively, a DC load test called the ‘dual pulse’ or ‘voltage drop’ method consecutively applies two smaller loads to a battery and calculates the internal resistance from the known load resistance and voltage drop. The resulting smaller discharge currents permit the use of larger resistors less susceptible to thermal effects, improving the internal resistance measurement accuracy. Linden (2010) writes, “a more accurate method of calculation (as opposed to flash amps) is the voltage drop method. In this method, a small initial load is

7

applied to stabilise the battery. A load approximating the application load is then applied” (p. 9.27). A variety of other internal resistance measurement methods are available including energy loss and calorimetric methods according to Schweiger et al. (2010). However, these methods do not present reliable, easily obtainable internal resistance measurements according to the article findings.

2.5 Battery Charging Typically, a CC/CV (constant current/constant voltage) method is employed to charge LiPo batteries according to Buchmann (2001). Buchmann (2001) defines a complete LiPo charge as when the upper battery voltage threshold has been reached – typically 4.1V – and the charge current has reached 3% of the nominal charge current. Various charging circuit assemblies can be implemented using MOSFETs, digital electronics or ‘off the shelf’ ICs (Integrated Circuits) to achieve the CC/CV charging scheme shown in figure 2.

Figure 2 A graph showing the CC/CV LiPo charging scheme implemented in battery chargers. It is clear from the graph that at the upper battery threshold voltage 4.1V, the charging scheme switches from constant current to constant voltage. Buchmann (2001).

8

2.6 Circuit Design and Operation According to Linden (2010), Energizer (2005) and Buchmann (2001), various battery conditions including temperature, state of charge and age affect internal resistance. The hobby website produced by Battery University (2011) states that “resistance levels are highest at low state of charge and immediately after charging.” As previously mentioned, the internal resistance of LiPo batteries varies inversely proportional to temperature, exhibiting high internal resistance around electrolyte freezing level. Figure 3 shows the effect of temperature on the internal resistance of a fresh AA battery; the rapid internal resistance increase is clearly visible around the water-based electrolyte freezing temperature negative twenty degrees Celsius.

Therefore, an

important consideration in the circuit design and operation is to ensure a constant

temperature

measurement consistent

during

process readings.

to

the yield

Buchmann

Figure 3 A graph showing the change in internal internal resistance of an AA battery with resistance measurements taken from temperature. Energizer (2005). (2001)

suggests

that

a fully charged battery immediately after charging are higher than those taken several hours after removal from the charger. Buchmann (2001) also defines a full LiPo cell cycle as a discharge to 3V and warns that many modern batteries contain protection circuits that can distort internal resistance measurements. Finally, Buchmann (2001) gives insight into the method by which some battery quick testers calculate SOH readings - “Some quick testers simulate the equipment load and observe the voltage signature of the battery under these conditions. The readings are compared to reference settings in the tester. The resulting discrepancies are calculated against the anticipated or ideal settings and displayed as SOH readings.” (p. 192). In order to quantify battery SOH it is clear that a form of trend monitoring must be implemented, consisting of historical internal resistance measurements and manufacturer defined limits to compare against readings.

9

2.7 Hypothesis According to Buchmann (2001), the usage of LiPo batteries does not contribute as much to the increase in internal resistance as ageing. However, “the wear down effects caused by usage and aging are more pronounced in LiPo batteries” (p. 164), confirming the hypothesis that with cycling, the internal resistance of LiPo batteries will increase. It describes that cell oxidisation increases internal resistance and is “the ultimate cause of failure” (p. 110). Albér (n.d.) attributes increasing battery internal resistance to cycling and aging as corrosion, sulfation and grid growth occur within the cell, which is congruent with the hypothesis.

10

3 Literature Review and Research Conclusions From the literature review and an analysis of the state of the art, the following chapter draws conclusions regarding the market requirement, circuit design and hypothesis. This chapter also highlights some issues encountered during research and makes clear the reasons for narrowing the project aims and objectives scope.

3.1 Literature Review Conclusions The requirement from RC vehicle users for LiPo batteries with low internal resistance is clear, confirming that there is an opportunity to market an internal resistance measurement circuit capable of quantifying battery SOH. A distinct lack of currently available internal resistance measurement products offers a unique chance to introduce a useful new technology into the RC vehicle market. On analysis of the sources, it’s clear there is unanimous agreement that internal resistance has a pronounced effect on the ability of a battery to deliver the required power. The definition of SOH - a metric encompassing parameters that contribute to performance reductions - can now be expanded to identify ‘performance reductions’ as a reduction in the ability to deliver the required power, due to increased battery internal resistance. The DC ‘dual pulse’ or ‘voltage drop’ method is best suited to satisfying the report aim. Due to the non-complex nature of RC vehicle loads, internal resistance as opposed to impedance is deemed to be an appropriate measure to quantify battery SOH. Although used extensively in commercial applications where high accuracy is required, AC current injection methods do not fulfil the ease of use and low cost requirements of the project. The simplicity of the DC ‘dual pulse’ method allows readily available power electronics to be assembled into a basic circuit, minimising cost and simplifying the measurement process. Despite the improved accuracy of AC over other measurement methods, the extensive equipment required and large measurement times of EIS do not lend themselves to fulfilling the project aim and objectives. The usability of readings obtained via AC methods, EIS in particular, are dependent upon a well-defined library of internal resistance readings for various battery types. Suppliers of LiPo batteries used by RC vehicle users do not currently provide the required data libraries for validating internal measurement readings, rendering AC methods unsuitable for this project.

11

As internal resistance is a function of variables including cell temperature and state of charge, the requirement to control these variables during internal resistance measurement has been clearly demonstrated. Therefore a major consideration during circuit operation will be to ensure constant cell temperature and charge level to produce consistent internal resistance measurements. Finally, the hypothesis is backed up by several literary sources, validating the motivation of this report to carry out the design, build and test stages.

3.2 Research Conclusions Research into charging methods revealed that purchasing an ‘off the shelf’ LiPo battery charging IC offered the best solution to fulfil the project aim. Although the construction of a CC/CV charger from scratch was technically possible, the required theory and development time proved too big a hurdle to overcome, leaving the ‘off the shelf’ solution as the preferred option for implementing charging into the final circuit. Despite multiple attempts to assemble an ‘off the shelf’ battery charging solution, each attempt ended with the destruction of the charging IC. Various factors contributed to the difficulty in producing a working LiPo battery charging circuit including limited personal technical ability and time constraints. Therefore, it was decided to separate the battery charging and internal resistance measurement circuits, assigning the charging circuit to potential future system development. Consequently the scope of the project narrowed, leaving extra time for refined and persistent work on an internal resistance measurement circuit where the best potential for development lies. As a result of the narrowed scope, the aim of the project became to develop a system for RC vehicle users that facilitates informed and unambiguous battery replacement decisions, whilst objectives related to battery charging were withdrawn.

12

4 DC ‘Dual Pulse’ Technical Description The following chapter describes the theory of the DC ‘dual pulse’ internal resistance measurement method and the steps required to measure battery internal resistance.

4.1 DC ‘Dual Pulse’ Theory The ‘dual pulse’ method allows the simple calculation of battery internal resistance using Ohm’s law. When placed under a load and discharged, the battery terminal voltage will drop due to the voltage drop across the internal ohmic resistance as governed by Ohm’s law. Using equation 1, the battery internal resistance can be calculated from the terminal voltage change and the known load resistance. Both Energizer (2005) and Albér (n.d.) make a recommendation for short discharge times to minimise the effect of polarization on internal resistance measurements. As previously mentioned, the TER is combined from internal ohmic and ionic resistances. The effect of ohmic resistance is seen immediately after load application, whilst the effect of ionic resistance occurs later during the discharge period. Studying figure 4, these phenomena can be visualised; the voltage drop dv1 occurs instantaneously and is due to the ohmic resistance of the battery. The non-linear voltage drop dv2 occurs over a period of milliseconds and is due to ionic resistance resulting from polarisation effects. In order to minimise the amount of ionic resistance measured and accurately measure the ohmic (internal) resistance, Energizer (2005) and Albér (n.d.) recommend keeping the discharge duration between 50 and 100ms.

Figure 4

A LiPo discharge curve showing terminal voltage against time, highlighting

the effects of ohmic and ionic resistances on the terminal voltage. Linden (2010).

13

Energizer (2005) and Linden (2010) recommend applying a light background load prior to the full load application to stabilise the battery discharge and provide equilibration for consistent measurements. Equation 1 can be used to calculate the internal resistance of a LiPo battery using the DC ‘dual pulse’ method.

Rin = Where

(V1

V2 )RL V2

(1)

Linden (2010)

Rin = internal resistance, ohms V1 = stabilised closed circuit voltage, V V2 = closed circuit voltage at load application, V RL = application load, ohms

4.2 Method of Internal Resistance Measurement 1.

Connect the cell to a light stabilisation load for a short period of time to stabilise the discharge and provide equilibration for consistent internal resistance measurements.

2.

Measure the cell terminal voltage under light load application.

3.

Disconnect the light stabilisation load and connect the cell to a heavy discharge load for a short period of time.

4.

Measure the cell terminal voltage under heavy load application.

5.

Disconnect the heavy discharge load and use equation 1 to calculate the cell internal resistance from the voltage measurements and the known load resistance.

14

5 Circuit Specification The following chapter defines what the internal resistance measurement circuit should do and how it should operate. In later chapters, it will be used as a benchmark for evaluation of the actual test results against expected results.

5.1 User Specification •

The circuit must accurately calculate the internal resistance of a single LiPo cell and present the internal resistance value to the user.



The circuit must be scalable to measure the internal resistance of multiple cell LiPo batteries.



The circuit must integrate easily with common LiPo battery charging technologies.



The circuit must be easy to operate and inexpensive.



The circuit must be able to facilitate internal resistance trend monitoring.

5.2 Operation Specification •

The circuit must apply stabilisation and application loads to the cell consecutively.



The circuit must not discharge the cell lower than 3V.



The circuit must apply the stabilisation load for 20ms and the application load for 100ms.

15

6 Circuit Design The following chapter describes the circuit design process and makes the justifications for technical and operational decisions clear. After defining the component selection, the software design is outlined before the final circuit diagram is presented.

6.1 Component Selection In order to fulfil the user requirements for low cost and ease of use, as well as providing good scope for future development, the implementation of the DC ‘dual pulse’ method in software with a microcontroller and accompanying power electronics has been selected. The highly popular Arduino Uno microcontroller offers a wide array of input and output ports to facilitate the scalability requirement whilst the strong program library and supporting community offers quick and easy troubleshooting solutions. The large on board memory also offers space to implement internal resistance trend monitoring as detailed in the specification. The Arduino Uno receives power via the on-board USB power connection from a PC host and is programmed in the C programming language using the standard Arduino IDE software. The DC ‘dual pulse’ method requires the consecutive application – switching - of two loads in order to accurately measure cell internal resistance. Using power MOSFETs in a switch configuration best fulfils the low cost requirement due to their widespread use and economical purchase prices, whilst also offering good power handling capabilities able to withstand the large discharge currents. The stabilisation and discharge currents have been set at 0.5A and 2A respectively, requiring MOSFETs capable of handling at least 8.4W, assuming a maximum cell voltage of 4.2V. To further reduce cost, the MOSFETs will be logic compatible to negate the need for gate driver chips, enabling switch control directly from the Arduino Uno. Due to the large currents encountered during internal resistance measurement, the discharge resistors must also possess appropriate power handling capabilities and low temperature coefficients to maintain accurate internal resistance readings. Assuming the stabilisation and discharge currents previously mentioned, power resistors capable of handling at least 8.4W have been selected. The large current through the power resistors will lead to ohmic heating, reducing the resistances. As equation 1 requires a constant load resistance to produce accurate and consistent results, load resistors with low temperature

16

coefficients have been selected, minimising the variation in load resistance over the discharge period.

6.2 Software Design The internal resistance measurement method described in chapter 4 has been implemented in software in the C programming language and is shown below. Using the I/O ports and A/D convertors of the Arduino Uno, the required load switching scheme and voltage measurements for cell internal resistance calculations as shown in chapter 4 have been realised in software. int dischargeOnPin = 10; int stabiliseOnPin = 11; int depleteOnPin = 12;

// Assign discharge switch variable to I/O port 10 // Assign stabilise switch variable to I/O port 11 // Assign deplete switch variable to I/O port 12

void setup() { pinMode(dischargeOnPin, OUTPUT); pinMode(stabiliseOnPin, OUTPUT); pinMode(depleteOnPin, OUTPUT); Serial.begin(9600); }

// Set I/O port 10 to output // Set I/O port 11 to ouput // Set I/O port 12 to output // Initialise serial comms with PC

void loop() { int openV = analogRead(A0); float openVoltage = openV * (5.0 / 1023.0); if (openVoltage > 3.0) { Serial.print(openVoltage,3);

// Read the open cell voltage // Convert to floating point // If the cell has not fully discharged // Print the open cell voltage to the // serial monitor Serial.print("\t"); // Tabulate the serial monitor digitalWrite(stabiliseOnPin, HIGH); // Start the stabilisation discharge delay(20); // Discharge at 0.5A for 20ms int stabiliseV = analogRead(A0); // Read the stabilisation discharge cell // voltage float stabiliseVoltage = stabiliseV * (5.0 / 1023.0); // Convert to floating point digitalWrite(stabiliseOnPin, LOW); // Stop the stabilisation discharge digitalWrite(dischargeOnPin, HIGH); // Start the application load // discharge delay(100); // Discharge at 2A for 100ms int dischargeV = analogRead(A0); // Read the application load // discharge cell voltage float dischargeVoltage = dischargeV * (5.0 / 1023.0); // Convert to floating point digitalWrite(dischargeOnPin, LOW); // Stop the application load // discharge Serial.print(((stabiliseVoltage*2)/dischargeVoltage)-2,3); // Calculate the cell internal // resistance and print to the // serial monitor Serial.println(); // Start a new line on the serial monitor

17

digitalWrite(depleteOnPin, HIGH); delay(10000); digitalWrite(depleteOnPin, LOW);

// Start the depletion discharge // Discharge at 2A for 10s // Stop the depletion discharge

} else { Serial.println("Discharge Complete"); while (true) { }

// If the cell has fully discharged // Print discharge complete to the serial // monitor // Enter an infinite loop waiting for reset

} }

18

6.3 Circuit Diagram The following circuit diagram shows the complete circuit utilised in the battery internal resistance measurement. The host PC undertaking the data recording and the connection to the Arduino ATmega328 microcontroller chip have been omitted for diagram simplification. Pins with no connection are not required for the operation of the circuit. Pins with the prefix PD are digital I/O ports used in the MOSFET switching process whilst pin AC7 is an analogue I/O port connected to an ADC for cell terminal voltage measurement.

19

7 Circuit Testing The following chapter highlights some testing procedure considerations and describes the testing method.

7.1 Testing Considerations As previously mentioned, the internal resistance of LiPo cells as well as the discharge resistances are dependent on temperature. In order to accrue consistent and accurate internal resistance measurements, the ambient temperature during testing must remain constant at twenty-one degrees Celsius. However, during the component selection process, discharge resistors with low temperature coefficients were selected and are therefore less affected by varying temperatures. Despite the load resistors having tolerances of 1%, to increase the reliability of internal resistance measurements the load resistances must be measured with an accurate ohmmeter to obtain their true resistances. As the cell internal resistance is also dependent upon the state of charge, the testing procedure will use a fully charged LiPo cell taken straight from the charger to equilibrate internal resistance measurements. Finally, in order to evaluate whether the hypothesis is correct, a method of showing increasing cell internal resistance with successive charge cycles is required. Therefore, the bulk of the testing procedure will consist of cycling the cell, measuring the internal resistance and discharging the cell whilst looking for an upward trend in the internal resistance against cell cycle graph. To facilitate this, an extra MOSFET switch and discharge resistor has been installed on the circuit as shown in the circuit diagram. This additional circuitry will discharge the LiPo cell at 1A until empty before it is recharged and the testing process repeats.

20

7.2 Testing Method 1. Ensure the LiPo cell is fully charged. 2. Ensure the ambient temperature is stable at twenty-one degrees Celsius. 3. Connect the circuit to a computer and initialise serial communications and data capture. 4. Remove the cell from the charger and connect it to the circuit. 5. Reset the circuit to start the internal resistance measurement process. 6. Monitor the discharge process until complete ensuring a consistent ambient temperature. 7. Remove the cell from the circuit and recharge for the next cycle. 8. Record the cell cycle number and transfer the internal resistance measurements to data processing software.

21

8 Test Results and Evaluation The following chapter presents the test data and results in graphical form before analysing them for accuracy and reliability. Sources of outlying data points are highlighted before conclusions are drawn as to the congruency of the results with the hypothesis, project aims and circuit specification.

8.1 Test Results After cycling the cell using the testing method described in chapter 7, the data collected has been aggregated and processed into a graph at figure 5 showing the variance of average LiPo cell internal resistance with cycle number. Due to variations in the fully charged cell voltage prior to the test procedure commencing, it was not possible to create a graph of fully charged cell internal resistance against cycle number. Therefore, figure 5 displays the average LiPo cell internal resistance calculated from internal resistance measurements taken every ten seconds during discharge.

Internal Resistance / Ohms

A graph showing the variance in average LiPo cell internal resistance with cycle nymber 0.12 0.1 0.08 0.06 0.04 0.02 0 0

5

10

15

20

25

30

35

40

45

Cell Cycle Number

Figure 5

A graph showing the variance in the average LiPo cell internal resistance with

cell cycle number. Studying figure 5, the linear trend line clearly indicates an upward trend in the internal resistance with cycle number, which is in agreement with the hypothesis and expectations. However, it is clear from figure 5 that the internal resistance measurements become significantly more spurious and unreliable with a large deviation beyond 35 cycles. As a result of this spread, the weighting of the trend line beyond 35 cycles gives a false

22

indication of increasing internal resistance with cell cycle. In order to allow reliable conclusions to be drawn, figure 6 presents the same graph as figure 5 with the spurious results beyond 35 cell cycles omitted. Prior to the 36th cell cycle, the LiPo cell was left fully charged, off the charger for several days which appears to have introduced large variations in the cell internal resistance measurements for later tests.

Internal Resistance / Ohms

A graph showing the variance in average LiPo cell internal resistance with cycle number 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

5

10

15

20

25

30

35

Cell Cycle Number

Figure 6

A graph showing the variance in the average LiPo cell internal resistance with

cell cycle number with data beyond 35 cell cycles omitted. As previously mentioned, the intended plot of internal resistance measurements taken at the fully charged cell voltage 4.1V against cell cycle number has not been possible to draw. For comparison with figure 6, figure 7 shows the variation in internal resistance measurements taken at the fully discharged cell voltage 3.0V against cell cycle number.

23

Internal Resistance / Ohms

A graph showing the variance in fully discharged LiPo cell internal resistance with cycle nymber 0.1 0.08 0.06 0.04 0.02 0 0

5

10

15

20

25

30

35

Cell Cycle Number

Figure 7

A graph showing the variance in the fully discharged LiPo cell internal

resistance with cell cycle number with data beyond 35 cell cycles omitted. As expected, figure 7 shows the fully discharged cell internal resistance is higher than the average cell internal resistance shown in figure 6. This confirms that internal resistance is a function of cell charge state and agrees with common theory that says battery internal resistance increases with decreasing charge level. The internal resistance measurements are presented with an uncertainty of ±0.02Ω, calculated from the discharge resistor tolerance ±1%. It should be noted that this is a large uncertainty with respect to the maximum working cell internal resistance 0.16Ω as detailed in the cell data sheet. Other sources of uncertainty are difficult to quantify as they have not been monitored throughout the project, but they are highlighted in the test results evaluation.

8.2 Test Results Evaluation It is difficult to draw conclusions regarding the agreement of the test results with the hypothesis due to insufficient testing data. The lack of data is due in part to the time consuming process of charging and discharging the LiPo cell, but the elimination of unreliable data due to various sources of uncertainty is the largest contributor. As previously highlighted, careful control of cell temperature and charge level during internal resistance measurement was required to ensure consistently accurate results. In practice however, these are difficult variables to control without great levels of attention and large

24

amounts of scientific equipment. The testing procedure was carried out in a converted church, which possessed poor thermal control capability due to the nature of the building. Therefore, large uncontrollable temperature variations occurred during the internal resistance measurement procedure and due to the aforementioned dependency of internal resistance on cell temperature, affected the consistency and reliability of the test results. Despite the aforementioned sources of error and inconsistencies with the test data, the circuit appears to be measuring cell internal resistance correctly. According to the cell data sheet, the normal working cell internal resistance should be under 160mΩ for 300 cycles. Forecasting the trend lines of figures 6 and 7 forward, according to the data the cell internal resistance is approximately 120mΩ at 300 cycles, which is a surprisingly accurate forecast considering the internal resistance measurement uncertainty of ±0.02Ω and lack of data. The gentle upward slope of the trend lines displayed in figures 6 and 7 shows increasing internal resistance with successive cell cycles and is therefore in agreement with the hypothesis. A significantly larger amount of test data would display the upward trend more clearly and allow the conclusion that the test results are in line with the hypothesis to be stated with much greater confidence. The project aim - to develop a system for RC vehicle users that facilitates informed and unambiguous battery replacement decisions – has been partially fulfilled by creating an internal resistance measurement circuit, but will be completed with further development into SOH quantification as detailed in later chapters. The circuit does present scalability to work with multiple cell LiPo batteries thanks to the multitude of analogue input and output ports on the Arduino microcontroller. The use of balance leads found on many commercially available LiPo batteries facilitates multiple cell internal resistance measurements. A simple circuit expansion and a small amount of program code change is required to adapt the circuit to measure multiple cell LiPo battery internal resistance.

25

9 Conclusion The ability of RC vehicle users to decide confidently when to replace their batteries is hampered by a lack of battery health measurement products available to them, creating unnecessary wastage and expenditure. Internal resistance is a key indicator of battery performance and can be measured to provide RC vehicle users with an overview of the state of health of their batteries, permitting better informed replacement decisions and minimising cost. This project set out to find a solution to this real problem through a design, build and test exercise. By performing thorough research of the issue at hand, a clear set of aims and objectives was defined, permitting the design of a solution which best serves the needs of RC vehicle users. Despite numerous project scope changes during the design and build stages caused by technical ability and time limits, the overarching test results suggest the main aims and objectives have been fulfilled. With further refinement and development, the project promises to provide a legitimate marketable solution to save RC vehicle users money.

26

10 Future Development This chapter highlights areas for future project development and briefly discusses the reasons for developments and their practical implementation.

10.1 State of Health Quantification As previously mentioned, battery SOH is a battery health metric encompassing parameters which contribute to performance reductions. Although internal resistance measurement has been identified as a key battery performance indicator, various other battery measurements including charge acceptance and charge rate are required to yield accurate and consistent battery SOH measurements. Furthermore, raw internal resistance readings are difficult for the average hobbyist battery user to interpret, strengthening the case for development of a system that produces SOH health measurements as a percentage – significantly easier for users to understand. Through further circuit development, the production of a complex algorithm and the implementation of battery heath trend monitoring the circuit will deliver SOH measurements to the user, fulfilling the aim of the project.

10.2 Charger Integration Typically, consumers prefer to purchase ‘all in one’ units as opposed to separate items. Despite the inability of the project to deliver a combined battery charger and SOH measurement system, the user requirements for ease of use and low cost are best fulfilled by combining the two systems to enable the purchase and use of a single product. Further development will reveal the best solution for integration with a charging system. Options include utilising an ‘off the shelf’ charging IC solution, or striking a deal with manufacturers to include the SOH measurement system with current charging solutions available on the market.

10.3 Temperature Control The importance and difficulty of maintaining careful temperature control during internal resistance measurement has been continually stressed throughout this report. A significant amount of further development is required in order to mitigate the difficulties of temperature control for remote control vehicle users to facilitate the ease of use requirement. Further development options include implementing an active thermal control system or using modelling to provide some form of temperature compensation.

27

11 Project Management This chapter compares the actual project progress with the expected progress as detailed by figures 8 and 9 and provides a summary of the overall project progress.

11.1 Gantt Charts

Figure 8

The first project Gantt chart produced 21/11/15

Figure 9

The revised Gantt chart produced 22/04/16

11.2 Project Progress Overall, the project progress has been consistent and in line with the expectations as planned and displayed in figure 7. With the exception of the circuit testing objective, all parts of the project have remained on track and have been completed before or on the

28

planned deadline. The extensive amount of time required to charge and discharge the LiPo cell during testing was not foreseen and as such, twice the amount of planned time was required to complete the testing objective.

29

12 Self Review This chapter provides some insight into my progression and performance as a final year student and highlights areas that I have developed throughout the project.

12.1 Project Progress Overall, I am very pleased with how the project has progressed and developed. After some initial teething problems defining the project scope and aims, the project rapidly took shape and allowed me to formulate some definitive objectives to work towards. The failure to build a functioning battery charging circuit was disheartening and frustrating to begin with, but in hindsight, was a useful mistake to make as it helped me to review my technical capability and reduced the workload down to a more manageable size.

12.2 Personal Development Personally, I feel this project has been thoroughly useful in developing my confidence when posed with a challenge I know little about. It has helped me to not feel intimidated when faced with a task requiring significant time dedication and technical ability, and to break it down into smaller objectives with a meticulously planned approach to complete it. My professional time and crisis management skills have benefitted significantly by facing all of the challenges during this project and will translate well into a commercial environment as a graduate. My technical writing and formatting skills have improved dramatically and enabled me to approach future technical projects and papers with increased confidence. The importance of correct and consistent referencing to maintain project integrity has been clearly demonstrated and stands me in good stead to produce similar high quality, reputable pieces of work in the future. Crucially, my research and analysis skills have improved as I have learned to locate alternative sources of information such as library textbooks and technical papers. I now understand how to perform a literature analysis, extracting key information and assessing the source reliability and objectivity. Overall, this final year project has been thoroughly enjoyable and invaluable in highlighting the level of responsibility and professionalism required to undertake large research and development tasks. I would not hesitate to use the experience I have gained as guidance for future projects in further education and a foundation upon which to build my professional career.

30

13 References Albér, G. (n.d.) Predicting Battery Performance Using Internal Cell Resistance. Florida, Albércorp. Battery University. (2011) BU-902: How to Measure Internal Resistance [online]. Canada, Cadex. Available from: http://batteryuniversity.com/learn/article/how_to_measure_internal _resistance [Accessed 10th April 2016]. Buchmann, I. (2001) Batteries In A Portable World. Canada, Cadex Electronics. Energizer. (2005) Battery Internal Resistance. Technical Bulletin, [online], 1(1), 1 – 2. Available from: http://data.energizer.com/PDFs/BatteryIR.pdf [Accessed 10th April 2016]. Linden, D and Reddy, Thomas B. (2011) Linden's Handbook Of Batteries. New York, McGraw-Hill. Mizushima, K. et al. (1980) “A New Cathode Material For Batteries of High Energy Density”. Materials Research Bulletin [online], 15(6), 783 – 789. Available from: http://ac.els-cdn.com/0025540880900124/1-s2.0-0025540880900124- main.pdf?_tid=6703 360a-ff37-11e5-9214-00000aab0f27&acdnat=1460305092_f85ead6d3b879a95bb2a422f61 3f9484 [Accessed 10th April 2016]. Schweiger, G. H. et al. (2010) Comparison of Several Methods for Determining the Internal Resistance of Lithium Ion Cells [online], Sensors. 10(1), 5604 – 5625. Available from: www.mdpi.com/1424-8220/10/6/5604/pdf [Accessed 10th April 2016]. Sony Energy Devices Corporation. (2016) “Keywords to Understanding Sony Energy Devices” [online]. Japan, Sony Energy Devices Corporation. Available from: http://www.s onyenergy-devices.co.jp/en/keyword/ [Accessed 10th April 2016]. Turnigy. (n.d.) Instruction Manual [online]. Hong Kong, Turnigy. Available from: http://www.hobbyking.com/hobbyking/store/uploads/832792292X250282X36.pdf [Acces sed 10th April 2016]. Whittingham, M. S. (1976) “Electrical Energy Storage and Intercalation Chemistry”. Science [online], 192(4244), 1126 – 1127. Available from: http://science.sciencemag.org/c ontent/192/4244/1126.f ull-text.pdf+html [Accessed 10th April 2016].

31

View more...

Comments

Copyright © 2017 KUPDF Inc.
SUPPORT KUPDF SitemapRead more | Ojousama wa Oyomesama | The Walking Dead No Man's Land pro