Made to Measure... Measurement Technology for Fuel Cell and Battery Systems

Wolfgang Schmid SMART Electronic Development, GmbH Stuttgart, Germany

The current very lively debate about alternative fuels once again throws the issue of electric drive systems into sharp relief. At the same time, particular importance is attached above all to the appropriate energy sources for electricity. Two recent reports illustrate this fact: first, the announcement of a new alliance to develop lithium-ion batteries between BASF, Bosch, Evonik, Li-Tec, Steag and VW and second, the founding of the Automotive Fuel Cell Cooperation with Daimler AG as the majority stakeholder.

Some of the most efficient energy sources for electricity include electrochemical systems such as fuel cells and batteries. Although these electrochemical cells form the very heart of energy converters and energy stores, a truly serviceable energy source is only achieved when a large number of other components have also been incorporated. In the case of a fuel cell, these are typically:

· Media treatment

· Waste heat management

· Voltage conversion

· Control electronics

It is only through the interplay of all of these components that we can achieve a fuel cell system that can provide a regular and dependable source of electrical power through the supply of hydrogen and oxygen.

Electrochemical Converters and Stores

Because all electrochemical systems have cell voltages of a few volts only, for any worthwhile use in energy applications they must be connected in series (also known as cascading or stacking) to bring the whole system up to a higher overall voltage. Consequently the performance of the system as a whole is determined by the weakest link in the chain (or stack). With robust, technically mature products such as NiCd or NiMH batteries this is not normally a problem. With these technologies, we only need to measure one or two basic parameters (e.g. total voltage, temperature, impedance) in order to be able to determine the control parameters required to operate the system.

The situation is very different with more sensitive, less mature technologies such as fuel cells or technologies that are run close to their limits (lithium-ion batteries). Here it may not be expedient or may even be dangerous to measure just overall parameters, since with typically 60 to 100 cascaded cells, the malfunction of a single cell is hard to detect. It is far safer to measure individual cell voltages (or groups of cell voltages) and to draw more detailed conclusions from their absolute and relative behavior to one another.

Requirements

As different as the complete systems may be, they usually have a number of common objectives:

1. To ensure stable operation under changing conditions: Irrespective of environmental conditions (e.g., temperature) or load conditions (e.g., increased power off-take) the cell stack should be operated at a favorable working point.

2. To prevent malfunctions: Even in exceptional circumstances and fault situations, the system should be secure, e.g., by initiating contingency action.

3. Diagnosis: The ability to analyze a data pool is desirable for warranty claims, servicing and further optimization.

The various different states are monitored by sensors that measure the input variables (in the case of fuel cells this is the media supply), the process variables (e.g., temperature) or the output variables (current, voltage).

If primarily feed-forward control procedures are used for system management, then any desired parameters can be selected, depending on the process model. With feedback control modes on the other hand, it makes sense to measure output variables and use them for feedback. Direct output variables, such as the lowest cell voltage for example, are best used to initiate safety functions. For diagnostic purposes it is advisable to record the load response (voltages, current and temperature), preferably in a compressed form as statistics.

Single-Cell Measurements

One of the most important measured variables is the output voltage, measured either as the total voltage of a stack or as single-cell measurements. Although it is far easier to measure the total voltage, single voltage measurements (or group voltage measurements) allow us to draw a number of conclusions (see the chart):

1. From the distribution of single measurements relative to one another we can infer the supply of medium and initiate system management actions accordingly (e.g., cell purging).

2. Only a single-cell measurement can indicate whether a cell falls below/rises above a critical voltage level, and where the cell is, so as to perform a safe shutdown, for example.

3. Time-resolved measurement can indicate trends. Statistical analyses enhance the quality of maintenance interval or service life projections.

Conventional (multi-function) measuring instruments can basically be used for these measurement tasks. However these are usually designed more for the measurement of parallel channels and less for the requirements of measuring cells connected in series. Care must therefore be taken to ensure that the high demands made on the input voltage range, common-mode rejection and insulation are met.

A cell stack of the type often used in automotive engineering can have up to 100 cascaded cells. If we wish to measure deviations in the mV range in the presence of voltages of up to around 100V with sampling rates of 1ms per channel, this equates to a common-mode rejection of 100dB at 1kHz and calls for special measurement technology.

System Requirements

Multiple stacks are often connected in series to obtain the working voltages of several hundreds of volts that are needed for power electronics. This means that the measuring equipment must also have insulation voltages of almost 1000V. Real systems have switched loads and peripherals that all cause electromagnetic interference. Consequently the electronic instruments must be able to reliably measure changes in wanted signals in the mV range even in the presence of high interference sources.

In addition, to these hardware requirements, there are a number of specific requirements for operating fuel cells systems and battery systems:

1. An alarm must be tripped when a cell voltage falls below a set threshold. This alarm must trip in real time, i.e., run on the measurement module.

2. The measured values must be sent to the system controller safe from interference. Field buses such as the CAN bus are ideal for this purpose.

3. The electronic measuring instrument should be robust and compact so that it saves space when housed in the system, preferably close to the stack to minimize induced EMI on the instrument leads.

A Practical Example

A proven device that has been used for many years to measure cascaded fuel cells is the Cell Voltage Monitoring System (CVMS, see Fig. 2) supplied by SMART Electronic Development GmbH. The robust, compact design of the measurement modules allows them to be directly incorporated into the final application, to equip experimental vehicles for example.

The measurement module has up to 90 measurement channels with different voltage ranges, and transmits the results to the system controller and system monitor on two CAN buses. The module is housed in a tough aluminum casing and is rated IP54 so is protected from splashing water.

A range of interface converters (e.g., CAN to USB) as well as software drivers and GUI's are available for data analysis on PC. The CVMS is used successfully in test stands, experimental vehicles and military systems as well as for monitoring electrolyzers. It stands as a well-engineered, robust and cost-effective instrument for the development, testing and quality assurance of electrochemical systems such as fuel cells or Li-ion batteries.