Solving the Challenges of Megawatt Chargers

Littelfuse executive details an alternative technique for fast and efficient high-power charging of commercial electric vehicles.

A forced air-cooled 12-pulse thyristor stack supports 2 MW of output power and has a smaller space requirement than a MOSFET configuration. (Littelfuse)

The most significant barrier to the growth of electric vehicles (EVs) is the battery charging infrastructure. The availability of charging stations and the time required to recharge vehicles have become the critical impediment limiting the adoption of EVs, which is particularly true for long-haul commercial electric vehicles (CEVs). Heavy-duty, fossil-fuel vehicles such as trucks and buses contribute about 25 percent of total vehicle emissions.

Dr. Martin Schulz, Global Principal, Application Engineering, Littelfuse. (Littelfuse)

Electrification of these heavy-duty vehicles poses a challenge for the charging infrastructure because the vehicles operate over long distances and have random routes. A charging station will require available power above 1 MW to recharge a long-haul CEV in under 30 minutes, the time a driver would find acceptable for a meal break.

A conventional circuit design for a high-power charging station employs wide bandgap semiconductors such as SiC MOSFETS. However, the highest possible efficiency must be the most important design objective since power losses can result in increased costs and wasted energy. An alternative design, detailed in this article, is based on topologies used for electrolysis, which substantially improves circuit efficiency, reduces system complexity and dramatically reduces energy costs.

Magnitude of power required

CEV delivery vehicles providing intra-city services during the day often have a central depot where overnight charging is practical. With at least eight hours of idle time, an 8-10 kW charger is sufficient to recharge 50-80 kWh into a vehicle’s battery. A typical electric bus will have a 250 kWh capacity and need 30-40 kW for a recharge during a 6- to 8-hour idle timeframe. A 500 kWh charger typically serves this charging requirement.

When the luxury of an overnight charge is not practical, such as for vehicles traveling long distances, chargers at stations along the road must be available. A long-haul CEV can use 500 kW delivered in under 30 minutes, and that level of energy delivery requires charging power greater than 1 MW. As a result, standards for high-power charging stations define power levels up to 2.2 MW and allow for options to upgrade to 4.5 MW in the coming years.

High-power charging technology

Figure 1. A four-station, high-power charger system using a medium-voltage transformer. (Littelfuse)

Current chargers on the market, such as passenger-car chargers, have power levels up to 350 kW and are often installed in groups of six to ten charging stations. The required power for the installation necessitates step-down transformers to convert 10-30 kV grid voltages to about 690 VAC. Figure 1 shows a charging station installation block diagram with four chargers. The AC-DC converters achieve 350 kW by paralleling units typically sized between 60-80 kW. The AC-DC converter consists of an input state with boost and power factor correction circuits and an output DC-DC buck converter stage. The output stage supplies a controlled voltage suitable for the charged vehicle battery.

Figure 2. Conventional high-power, battery charger topology using power MOSFETS for AC-DC conversion. (Littelfuse)

For 800 V-class battery systems, the chargers must operate at voltages up to 920 V to properly charge these batteries. Using 1200 V SiC-based MOSFETs in the power section’s design, as illustrated in Figure 2, the circuit can achieve an overall efficiency of up to 97 percent.

The space required for MOSFET-based, AC-DC converter design contains a volume of 35 liters/unit, which can fit into a 19-inch (483-mm) width rack with 31.5-inch (800-mm) depth and two height units (HU). Five of those 70-kW, AC-DC converters stacked to output 350 kW fill 175 liters – and that does not include any pumps and radiators needed for cooling the circuit. For a dual 350-kW charger, the minimum space consumed is a cube with 1.5-m (59-inch) sides and a volume of 3.4 m3 (120 ft3). This volume includes the power electronics, the cooling system and auxiliary systems.

For CEVs, new battery voltage packs will most likely reach 1.5 kV. One method to address the higher battery voltages would be to use MOSFETs with higher voltage ratings, which would consume even more space. However, a different approach can provide a better option for 2-MW and higher power chargers.

Efficient design based on electrolysis

Electrolysis, an electrochemical process used to create pure elements, uses DC voltage to drive the chemical reactions at a cathode and an anode. The process requires a quality DC voltage and precise current control. The technology used for electrolysis can work for battery charging, which also is an electrochemical process.

Figure 3. High-power, 12-pulse bridge rectifier-based charger topology. (Littelfuse)

High-power electrolysis systems often use a circuit topology based on thyristors in a controlled 12-pulse bridge rectifier configuration. Figure 3 gives an example of such a high-power thyristor electrolysis system. The design exhibits outstanding efficiency and reliability since it represents a single-stage AC-DC energy conversion. Thyristor-based designs have been in use for decades, and the components have superior power- and thermal-cycling capabilities.

The transformers used in a thyristor-based design are the same size as those used to power the MOSFET-based designs that output equivalent power. The space savings derive from the thyristor topology, which can be as small as 10 percent of the size of a scaled-up MOSFET design.

Figure 4. Battery-buffered charger using two series-connected 6-pulse bridge rectifiers. (Littelfuse)

Incorporating a battery pack in the system becomes an option to relieve the power grid from the high-power demand of battery charging systems. A thyristor-based design can charge the buffer battery pack. The buffer battery pack has a higher voltage than CEV batteries, so the charging circuit requires a buck DC-DC converter to adapt the charger output to the CEV battery. Depending on the choice of the transformer’s winding technique, designers can use a circuit with either parallel thyristor bridges, as can be seen in Figure 3, or series bridges, as depicted in Figure 4.

With the thyristor-based, battery-equipped circuit topology, the charger can either relieve the grid of peak demand or supply stored energy to the grid. Unlike the thyristor designs, the rectifier design of the MOSFET-based chargers is not capable of returning power to the grid. In contrast, the firing angle on the thyristors determines whether the circuit dispenses DC power to a battery or AC power back to the grid. If the firing angle is less than 90°, the thyristors are in the rectifier mode. The thyristors are in inverter mode when the firing angle is between 90° and 180°.

Benefits of the thyristor design

Figure 5. Guidelines for the operating area of a 2.2-MW battery charger. (Littelfuse)

The CharIN organization, a non-profit organization that facilitates collaboration among companies in the e-mobility business, has published guidelines on high-power charging for commercial vehicles. The guidelines provide recommended maximum values for charging voltage, current and total power. Figure 5 illustrates the recommended charger operating area for chargers with up to 2.2-MW capacity.

Using an appropriate power thyristor that can deliver over 1000 A with a conducting state resistance of around 1 mΩ, a B6-configuration bridge circuit can output 1.1 MW with only 2200 W of internal losses. The resulting efficiency exceeds 99.7 percent.

Using the MOSFET topology, two chargers in parallel could boost efficiency from 97 to 98.5 percent. Even then, that configuration does not approach the efficiency of the thyristor-based topology. Furthermore, the two-charger MOSFET configuration carries a much higher cost and lower reliability.

The thyristor-based design also has a much smaller space requirement. A 2000 V B12C stack capable of outputting 1700 A has the dimensions shown in the lead image. Two of the air-cooled units require 0.4 m3 (14 ft3). The thyristor-based design, eliminating liquid cooling with corresponding pumps, tubes and chillers, reduces the installation space from about 6 m3 (212 ft3) to less than 1 m3 (35 ft3), saving 83 percent of the space needed for installation.

A 2.4-MW charger can transfer 400 kWh into the battery pack of a long-haul CEV within 10 minutes. Assuming the charger provides a 10-minute charge for three vehicles in one hour, over 24 hours, the charger could supply 72 CEVs every day. In a year of 7 days/week of operation, the charger could service over 26,000 vehicles.

The total energy supplied by the charger would exceed 10 million kWh. A 97-percent-efficiency charger would lose around 300,000 kWh as heat. Using the cost for electricity at $0.11/kWh, a 2.4-MW charger would incur over $33,000 due to the lost energy. For every 1000 charger installations, lost energy would cost $33 million.

A 2.4-MW charger based on a thyristor topology that can achieve 99.7 percent will reduce the losses incurred by a 97-percent-efficiency charger by 90 percent. In addition, lower cooling demands on the charger power electronics further reduce energy consumption.

In a converter offering 97-percent efficiency, 60 kW of losses must be managed by an active cooling system. Chillers and pumps necessary to dissipate 60 kW easily consume 20 kW. When operating over the same 4300 hours that the charger operates, the chillers and pumps consume an additional 86,000 kWh. The additional power requirement adds $9.5 million to the energy bill for each 1000 chargers installed.

Every kWh saved eliminates 0.5 kg of CO2 in power generation. The more efficient approach thus contributes to save an additional 190 tons of CO2 per charger every year.

Dr. Martin Schulz, Global Principal, Application Engineering, Littelfuse, submitted this article to SAE Media.