Categories
Electric Vehicles

Measuring EV Efficiency

Sometimes industry standards don’t make sense. It may be a requirement by a specific industry (specs for some connectors come to mind), obscure specifications (such as GM’s long-lived wheel arch clearance for extreme snow), and sometimes it’s a federal requirement created to provide a measure of something technical (I think we can all agree non-technical people should NOT be making decisions on technical issues!).

Enter MPGe, or Miles Per Gallon “Equivalent.”

The problem originated a few years ago when the EPA had to provide a measure of efficiency for the Nissan LEAF when it was introduced. You may not know what a “Monroney” label is, but if you have ever looked at a new car you’ve seen one; it is the EPA label displaying a vehicle’s fuel efficiency.

"Monroney" Efficiency Label for an EV
“Monroney” Efficiency Label for an EV

As engineers, we have a tendency to overcomplicate things. We like to analyze things quantatively, using formulas and empirical methods to measure the quantifiable and extrapolate trends and break things down to the lowest common denominator. I think we can all agree, though, the best “technical” method to determine efficiency of any system: take the ratio of the unit output or end product to the amount of energy or power input. Right?

While the familiar “miles per gallon” may not appear on the surface to fulfill this, consider it taken in a more literal form:

where η=Wout/Win

ƞ represents System Efficiency (Greek letter “eta,” I admit I had to look this one up)

Wout represents output power (the amount of energy to move a car a specified distance, in miles)

Win represents input power (the amount of energy contained in a specified quantity of fuel, in gallons)

While we may not think of this commonly used term in this way, it is essentially what it represents. So why not use the same formula for an EV, in commonly accepted units? You would find that:

Wout represents output power (the amount of energy to move a car a specified distance, in miles)

Win represents input power (the amount of energy utilized by the EV, in KWh)

So- you simply end up with miles per KWh. Simple and descriptive, right?

Well, a few early vehicles did get out with this on their efficiency label. But, as with any good regulatory and political organization, the EPA shortly thereafter decided to conduct a variety of focus group studies to determine what the public would accept as a standard definition. While a variety of definitions were provided, MPGe was identified as the clear winner in this battle. The EPA settled on that, and that figure now is provided on every EV sold in America.

Measuring EV Efficiency 2

The MPGe figure was originally created as a measure of the efficiency of a non-gasoline alternatively-fueled vehicle for Progressive’s X-Prize back in 2007. The EPA currently defines this as the efficiency of a vehicle, assuming an equivalency of 33.7 KWh per gallon of gasoline, measuring the energy consumed when driven on one of their standard test cycles (as with a normally-fueled vehicle).

I am not sure what is more disappointing; that we now have a technically accurate but very non-descriptive definition of efficiency for EV’s, or that I expected a different result from a bunch of bureaucrats. One thing is for certain, I will continue to use miles/KWh as a measure of efficiency for my own EV. It makes more logical sense, and it doesn’t make my engineer brain swarm with questions… Just really how much energy is in a gallon of fuel? What sort of fuel is it on the test? What if a vehicle calls for 91 octane Premium Unleaded but most Americans will just put in the cheapest available? What if you have a Flex Fuel vehicle and you run E85? What if…

So, what silly specifications have you had to deal with lately? What odd specifications do you have to convert back to “your” specification in your head?

Categories
Electric Vehicles

Don’t Forget About System Efficiency

As engineers, we can be obsessed with solving problems. If a circuit doesn’t work right, we debug it to find the root cause. If a component fails, we look for the root cause to prevent future failures. If you are given a specific improvement to make in a product, you focus on that specification or feature to improve.

I was recently tasked with improving the range in an EV. The vehicle featured a 14.4 KWh, 72 Volt (nominal) Lithium Iron Phosphate (LiFePO4) traction battery (24 cells, each 200Ah and nominally 3.0V). Vehicle weight was approximately 2100 lbs. Using a standardized test cycle, this vehicle routinely achieved an average of just over 60 miles range (driven on the same course, at similar speeds, throttle position/load, and environmental conditions). The target was to improve the range by at least 30%, to approximately 80 miles.

There are many ways to affect range in an EV. The primary target for most designers is the battery pack. Increase the capacity (KWh) of the pack, and you increase range (generally at a 1:1 ratio). Weight in an EV is range, and generally  proportional as well. You can also work to eliminate system-level inefficiencies, which may range from connector and cable resistance, mechanical inefficiency, control electronics (such as FET or IGBT) resistance, etc.

If you have ever taken on a weight reduction program you will know that reducing weight by 30% (or a major portion of that) is a task of Titanic proportions. In this case, the time and budget was not available to undertake this. Likewise, space for a larger battery pack was very limited. With battery packs, capacity (KWh) can be increase by physically larger cells (increased Ah capacity) or by adding additional cells to increase voltage. The former is a very transparent change, other than physical space needed. The latter is one that requires a review of system capability (maximum voltage of components, number of cells the battery management system can handle, etc.).

In order to maximize the space available for an increased battery pack it was decided to increase the battery pack capacity by adding cells (larger Ah cells were not available to fit within existing space constraints). 8 additional cells could be located in the existing battery compartment, bringing the total number of cells to 32, system voltage to 96V (nominal), and total pack to 19.2 KWh. As it would happen, the system was designed with sufficient overhead to tolerate the higher voltage with minimal parts changes.

The expectation was that the increase from 14.4 to 19.2 KWh would yield a 33.3% increase in capacity. This should provide a direct increase in range after compensating for the weight increase. Each cell weighs 12.8lbs, resulting in a 308lb 72V pack and a 410lb 96V pack. The ~100lb increase in pack weight is an approximately 5% increase in vehicle weight, which should translate to a ~ 28% overall increase in range. Actual testing resulted in a very different story, however.

Tests using the same standardized test cycle resulted in ranges of 103.9 and 109.0 miles on one test vehicle, and 94.2 and 102.1 on another. Further testing is ongoing, though with an average of approximately 100 miles per charge, I can say comfortably that a 40 mile or 67% increase in range has been achieved. What has caused this nearly 2x increase over the expected improvement in range?

With the vehicle and motor in this application, 72V puts the motor at the low end of its efficiency. This becomes clear when looking at the efficiency of different operating voltages as a function of motor RPM:

Don’t Forget About System Efficiency 1
Figure 1 Motor Efficiency Comparison

As much as a 15% improvement in motor efficiency can be had just by upping the system voltage. When driving a set course with pre-defined speed and acceleration, only a certain KW (power output) is needed. As only that level of power is needed, the resulting improvement in efficiency translates directly to additional range. In addition to greater efficiency, the motor torque for a given RPM is achieved earlier in proportion to the higher voltage. With a greater low-end torque, even less energy is required to operate the vehicle, translating to further improvements in efficiency and thus range. Finally, with a higher system voltage, lower current is needed to achieve these power levels, resulting in reduced system losses (principally wasted power in the form of heat, directly proportional to current flow).

All of these items translate to additional range- not just the increase in energy capacity. It’s important when designing a system, or a component part of a larger system, to consider all the possible ways in which it and the system affect the efficiency for the end product you are designing.

Categories
Automotive Electric Vehicles

Connecting EV Batteries

You may have seen some of my previous blogs regarding Lithium batteries. In my blog Lithium Battery Fires Why I highlighted some of the various fires that have occurred in the last few years. In blogs {doclink 259971} and The Care and Feeding of a Lithium Battery I discuss some of the important properties of these batteries, and in effect, how to avoid the catastrophes detailed in that first link.

Connectorization of cells is absolutely critical. Why? Aside from heat and other losses (more below), arcing due to poor contacts can cause some massive problems with control electronics and circuit protection, not to mention safety.  As we all know, poor, arcing contacts can cause voltage spikes, just imagine what those spikes are like at hundreds of amps with a nominal DC voltage in the hundreds of volts.

In the EV world, connectors are large. For the most part, you will find cells inside of a pack (most traction battery packs are made up of dozens, possible hundreds of cells in series or a series-parallel combo) bolted together with bus bars. Why? Quite frankly, it is simple and effective. Due to the risk of arcing, the mechanical reliability of a bolted joint, and the high current requirements (again, typically hundreds of amps), large copper bus bars are typical. Cells also are usually installed once at the factory and not touched again for many years, possibly ever.

Figure 1 Prius Battery Cells
Figure 1 Prius Battery Cells

How do you size bus bars? This question can be a complicated one, based on your allowable cost, voltage drop, and power loss. I have yet to find a good tool for this, so have had to create my own in the past to answer this question based on the variables I cared about. I would suggest building a spreadsheet to calculate your size based on what matters to you. Using system voltage, peak (and sustained) current, use the following voltage drop formula to determine the Circular Mils of cable you need:

CM = (2*K * I * D) / VD

Where:

CM = Circular Mills needed

K = 11.1 (conductivity of copper, multiplied by 2 for both legs of circuit)

I = Max load current in amps

D = 1-way cable distance in feet

VD = Voltage Drop desired

 

Using this, and a chart to convert wire gauge (AWG or mm) and Circular Mils to cross sectional area, you can then determine the cross-sectional area of a bus bar that would be equivalent. For the sort of power levels that are experienced in EVs, I have typically used 0.5% voltage drop (as a % of source voltage) as a maximum value. Any higher than this and you start to dissipate a lot of watts (you can start to dissipate hundreds of watts quickly) in the bus bars. Always err on the high side, and I highly recommend testing. As you may know, loads can be created cheaply enough using light bulbs, heater elements, or even spools of wire.

Interestingly, once the battery pack is assembled, most EV makers switch to a more traditional connector. These may be rather large in size (contact wiping surfaces measuring in the square inches), but nonetheless are connectors that can be fairly quickly disconnected. This is due primarily to the need to disconnect a pack from a vehicle (electrically) for service or inspection procedures. Many vehicles will also feature a purpose-built emergency disconnect (independent of cable connectorization) and may also incorporate a fuseholder in such an emergency disconnect. Such a product is available from TE, the AMP+ Disconnect.

Figure 2 AMP+ Connector
Figure 2 AMP+ Connector
Figure 3 AMP+ Disconnect
Figure 3 AMP+ Disconnect

This product line offers a variety of connector types for the EV world, in various circuits and ampacity, all with a low contact resistance and ease of operation, in the SAE designated “safety orange” color used to designate high voltage connections in EVs.

Figure 4 LEAF Battery Cells, Bus Bars, and Output Connector
Figure 4 LEAF Battery Cells, Bus Bars, and Output Connector

As with any connector, identifying your critical parameters is vital to proper selection. EV connectors are no different. You have to understand your system and quantify your needs before you can begin the selection process.

Categories
Automotive Electric Vehicles

The Care and Feeding of a Lithium Battery

You have probably been hearing a little about Lithium Ion battery fires as of late. Boeing has gotten a lot of press for it lately. In my blog I highlighted some of the various fires that have occurred in the last few years. Only some of these fires can be attributed to the batteries themselves, and the majority of them to various issues regarding the packaging and handling of the batteries. What many people, including design engineers, don’t know is that there are many different Lithium battery chemistries available, all with differing precautions for handling and use. So what are the major types?

One of the most prevalent and also earliest Lithium battery introduced is the non-rechargeable (disposable) type that are used in watches and small electronics. These use lithium or lithium compounds as an anode. These vary in voltage from 1.5V to almost 4.0V, and can be found in higher voltages (such as a replacement for the common 9V battery) via the stacking of lower voltage cells internally. There is a dizzying array of exact chemistries, though the majority of these cells are Li-MnO2 (Lithium Manganese Dioxide). These are some of the hardiest of lithium batteries, though typically are found in lower-cost and low-drain applications. The environment and handling is not much of a concern, though they do self-discharge in high heat, and can overheat or explode if short-circuited or too rapidly discharged. Overall, they are fairly robust.

The rechargeable types of lithium batteries come in a similarly wide range of chemistries. Most handheld electronics use LiCoO2 (Lithium Cobalt Oxide). These are relatively inexpensive and offer excellent capacity as a function of cost. These are particularly sensitive, though, to rapid discharge and faults (short circuits, physical damage) and can create tremendous heat and burst under these conditions. Lithium Cobalt Oxide also generally has a much shorter cycle life (number of charge/discharge cycles). This can vary but will generally be in the hundreds of cycles (<500).

Growing in use are Lithium Iron Phosphate (LiFePO4), Lithium Manganese Oxide (LiMn2O4), and Lithium Nickel Manganese Cobalt Oxide (LiNiMnCoO2). These have found their way into a wide variety of products, but are becoming the darling of the transportation industry. These chemistries trade a little capacity for a dramatic increase in life (>2000 cycles). Additionally, these are some of the most stable chemistries available, being less reactive in nature, being less reactive with water (most Lithium batteries are very reactive), and “failing” in a much less pyrotechnic manner if shorted or breached (a “nail” test where a nail or spike is driven through the cell is common).

How do rechargeable lithiums compare to “traditional” rechargeable batteries? There are a few key differences, some known to anyone who has worked with other rechargeable chemistries, and some that are new:

Cell Life: The lifespan of a rechargeable lithium cell is generally expressed by the manufacturer as the number of charge-and-discharge cycles before degradation of the cell life. This typically is expressed as a discharge from 100% of cell capacity to 20% (described as an 80% “Depth of Discharge” or DOD), then a recharge back up to 100%. Degradation is typically defined as 80% of cell capacity. So, for example, a 1000 mAh cell with a cycle life of 500 can be discharged to 20% and recharged to 100% 500 times and at the 500th charge the cell will still deliver 800 mAh. Do you understand now why your cell phone battery seems to give up the ghost after 1-2 years?

Charging: Lithiums, like other rechargeable chemistries, require some intelligence while charging. Unlike other chemistries, they are much less forgiving to improper charging (risk of internal damage or leakage). Due to charge rate limitations (see below) the cells must first be charged in a constant current mode. As the cell voltage increases to near the fully charged state, the charger must switch to a constant voltage mode. When multiple cells are used or charged in series, cells must be balanced individually (see more below).

Charge and Discharge Rates: Typically rated in “C” (the Ah capacity of the battery), this is a measure of what current level the cell should be charged or discharged at and is specified by the manufacturer. For example, a 10Ah Lithium Iron Phosphate cell may allow for charging at 1C (10Ah), continuous discharging at 2C (20Ah), discharging for 10 seconds at 5C, and a pulse (less than 1 sec) of 10C.

Operating Voltage: As mentioned, a supplier will provide a recommended operating range for cell voltage. This is critical to cell life and safety. As such, active systems need to be employed to monitor the cells and manage charging. For small individual cells, there are many charging ICs available for engineers to select from that incorporate these functions on a single component with a few support components. For applications with many cells (such as EVs) a Battery Management System is required (more below).

Determining % Charge: For lead-acid and other chemistries, cell voltage is a good indicator of cell % charge. The cell voltage is fairly linear with this % value and is typically just translated by a simple voltage measurement. Not so with lithiums. Lithium chemistries can be somewhat linear in the middle of their voltage range, but with very little margin. For instance, a lithium cell might be considered fully charged at 3.6V and fully discharged at 3.2V. At 0.4V difference between the two, that’s less than a 10% swing in voltage that has to translate to a 0-100% scale. As a result, most charging or battery management ICs or systems usually resort to “coulomb counting” to accurately track battery %. A cell (or cells) will be fully charged based on max voltage, then an IC in the device or management system will track and report coulombs (discharging or then charging) to indicate % charge.

Temperature: Like most components, lithium cells have recommended storage and operating temperatures. It is important to note that this can be a huge impairment in operation. Most lithium chemistries lose a lot of capacity near or below freezing (as much as 20%). Interestingly, this is a function of the chemistry and a fully charged cell brought down in temp and then back up will still retain a full charge. A heating system may be required to realize full capacity at low ambient temperatures. Heat is also generated at high discharge rates (typically only an issue in automotive traction batteries or other high-draw applications) and will need to be dealt with to keep cells within their recommended temperature range.

Battery Management System (BMS): As mentioned, there are many factors to consider when using lithium cells. As a result, charge controller ICs (for one or a few cells) or BMSs (for larger arrays) have been employed to manage charging, voltage and temperature monitoring, and charge status. In some cases these incorporate shunts to allow an array of cells in series to shunt charge current around them once they are full (to better balance the pack). As you may know, in a series array of cells your total charge capacity (in Ah) is only as great as the weakest cell. If you continue to draw current or run a series array past what the lowest common denominator can deliver, you may damage that cell.

In all, lithium chemistries offer some great advantages over others, but at a cost of greater care and management. Even lead acids need the above precautions to maximize their life, but with their cost so low it is far less a concern in most consumer applications if there is a failure.