Electric vehicles (EVs) will gain more and more market share, eventually replacing internal combustion engine vehicles. DC fast charging stations will replace or integrate gas stations. Renewable energy sources such as solar and wind power will power them. People will want to be able to fully charge an electric car in less than 15 minutes, and they won’t want to wait in line for the only charging point.

Considering that there are multiple charging piles, the local charging peak power that the grid needs to provide exceeds 1MW. The grid could collapse at multiple points or require huge investments to improve transmission lines and centralized power plants to substantially increase base load. However, this load is pulsed and must be integrated with intermittent energy from renewable energy sources such as solar and wind.

Energy storage systems can solve this problem simply and elegantly. We use fuels like gasoline, natural gas, etc. to store energy and reuse it when needed, such as when refueling a car. In the same way, we can store electrical energy in batteries using Electronic and chemical methods. This energy can then be used to increase the amount of electric vehicle charging, by regulating power peaks, to keep the grid stable, or to provide power in the event of a power outage.

The automotive market has begun to shift. Nearly 3 million electric vehicles will be sold in 2020, with total vehicle sales exceeding 80 million. While 3 million may seem like a niche market, forecasts show that EV sales will grow rapidly, reaching 10 million in 2025 and more than 50 million in 2040, for a total of 100 million vehicles by then. This means that by 2040, 50% of vehicles sold will be fully electric. For all these cars, use a simple wall-mounted charger when at home, or a DC charger of several kilowatts in the case of a home with a solar power system and energy storage battery for slow charging overnight; on the street , it can be charged quickly through charging piles, or super-fast at the gas station of the future.

We are seeing good growth momentum in the renewable power generation market (which has recently seen a boom in solar photovoltaic (PV) systems) while the EV market is rapidly emerging, which is in line with solar system prices over the past 10 years A drop of about 80% is inseparable from strong decarbonization initiatives. Today, solar power accounts for just under 5% of global electricity generation and is expected to account for more than a third (33%) of global electricity generation by 2050.

In the context of the intermittent nature of electricity load in the future, electric vehicles to be charged and intermittent energy sources such as solar energy and wind energy will face some challenges, such as how to integrate these emerging players in the energy ecosystem around the grid. . Intermittent load demands such as electric vehicles require increased transmission line specifications to meet higher peak power demands.

Solar power will change the way centralized power plants operate, ensuring that the grid is not overloaded; people will demand more convenient ways to supply electricity, and more and more of their own electricity in their homes will be provided by residential solar power systems.

In order for all entities to work together smoothly and benefit from renewable energy and zero-emission electric vehicles, energy storage systems must be involved, ensuring that we can store and reuse electricity generated when demand is low (for example, using solar energy generated at noon at night), harnessing excess energy energy to balance the grid load.

Energy Storage Systems (ESS) are equivalent to oil tanks or coal warehouses in the field of electrical energy and can be used in a variety of applications on a residential and industrial scale. In residential applications, it is easy to connect a PV inverter to a battery to store and use energy at home, or to charge a car at night with the energy the sun produces during the day. In industrial or utility-scale applications such as grid-connected services, energy storage systems can be used for different purposes: from conditioning PV and wind to energy arbitrage, from backup support to black-starting (elimination of diesel generators), and most importantly From a total cost perspective, investment can be deferred. In the latter case, the energy storage system can be used to meet the peak load demand of the grid node, ensuring that there is no need to upgrade the existing transmission line at a high cost. Another relevant use case is off-grid installations, where energy storage systems make microgrids or islands self-sufficient in electrical energy.

Energy storage system boosts electric vehicle fast charging infrastructure

Figure 1. Integration of renewable energy, energy storage systems, and EV charging infrastructure

Taking into account all possible applications, the energy storage system market will exceed the threshold of 1000 GW of power generation / 2000 GWh of production capacity by 2045, which is a rapid growth compared to today’s 10 GW of power generation / 20 GWh of production capacity.

This article will focus on energy storage systems for EV charging infrastructure.

Private and public AC charging infrastructures, while simple, are limited in power. The Level 1 AC charger operates at 120 V and has a maximum output of 2 kW. The operating voltage and maximum output power of the Level 2 AC charger are up to 240 V and 20 kW, respectively. In both cases, the on-board charger requires the conversion of alternating current to direct current. Wall-mounted AC charging piles are not so much a charger as a metering and protection device. Due to cost, size and weight constraints, car onboard chargers are always rated below 20 kW.

DC charging, on the other hand, allows charging EVs at higher power levels: Level 3 chargers have a maximum rated DC voltage and power rating of 450 V and 150 kW, respectively, and the latest superchargers (equivalent to Level 4) can Over 800 V and 350 kW. For safety reasons, the upper voltage limit is set to 1000V DC when the output connector is plugged into the vehicle. When using a DC charger, the energy conversion takes place in the charging pile, and the DC power output directly connects the charging pile to the car battery. This eliminates the need for an on-board charger, with the many benefits of reduced footprint and weight. However, during this transition phase, the EV charging infrastructure is still highly fragmented and varies by country, with most EVs using a small 11kW onboard charger that allows users to charge from an AC outlet when needed.

Increasing the charging power requires increasing the operating voltage, ensuring that the current remains within a reasonable range for cable size and cost, which means that the microgrid or sub-grid where the charging station is installed must be properly designed and sized.

Let’s imagine a charging station of the future (2030) where the fuel consists of electrons, supplied by pipes called transmission lines and connected to the medium voltage (MV) grid via transformers. Currently, fuel is stored in huge tanks underground and is regularly transported to gas stations by tanker truck. While supplying new fuel (electronics) all the time through the grid seems like an easy solution with no problems, we can see that if we want the driver to be able to fully charge an EV in less than 15 minutes, then This simple approach is not sustainable.

The charging station has five DC charging piles, each of which can output up to 500 kW of peak power. In the worst case, five charging stations simultaneously charge a completely depleted battery, which the charging station has to take into account. To simplify the calculations, we now assume zero losses in the power conversion stage and the battery charging path. Later in this article, we will see that even if the power loss in the entire power chain is small, the normal design can be affected.

Let’s assume there are five EVs, each with a 75 kWh battery (full EVs on the market today have batteries ranging from 30 kWh to 120 kWh) that need to be charged from 10% of state of charge (SOC) to 80%:

Energy storage system boosts electric vehicle fast charging infrastructure

This means that 262.5 kWh of electricity needs to be transferred from the grid to the EV in 15 minutes:

Energy storage system boosts electric vehicle fast charging infrastructure

The grid must supply a little more than 1 MW of power to these EVs for 15 minutes straight. The charging process of lithium batteries requires a constant current and constant voltage charging curve, so that the power required to fully charge the battery to 80% is greater than the power required to fully charge the last 20%. In our example, it is assumed that charging stops at 80% of maximum power.

The grid (preferably a sub-grid) where the charging station is located must intermittently maintain peaks greater than 1MW. Very efficient and highly complex active power factor correction (PFC) stages must be implemented to ensure that the grid remains efficient without affecting frequency or causing instability. This also means that very expensive transformers must be installed to connect low-voltage charging stations to the medium-voltage grid, ensuring that the transmission lines that carry electricity from power plants to charging stations are sized to meet peak power demands. If charging both cars, trucks and buses at the charging station, the power required is higher.

The simplest and most economical solution is to use electricity produced locally from renewable sources such as solar, wind, etc., rather than installing new transmission lines and large transformers. This allows users to connect directly to charging stations with excess power, rather than relying entirely on the grid. In practice, solar photovoltaic (PV) plants of 100 kW to 500 kW can be installed near charging stations or sub-grids connected to charging stations.

While photovoltaic power sources can provide 500 kW of electrical energy, reducing the power demand on the grid to 500 kW, photovoltaic power sources are intermittent and not always present. This creates instability problems for the grid, leaving EV drivers to charge their cars as fast as they can when the sun is shining. This is not what users want and is not sustainable.

Energy storage is missing from this puzzle of power electronics. Like the underground tanks at gas stations today, energy storage systems can be thought of as large batteries that store electricity from renewable sources and feed it to the grid, to a charging point, or back to the grid. The primary feature of energy storage devices is bidirectionality, which is on the low-voltage side of the grid. The new unit is designed to have a DC bus voltage of 1500V, connecting renewable energy sources, electric vehicle charging points and energy storage system batteries. The energy storage system must also be properly sized to ensure that the ratio between peak power and electrical capacity is optimized for the specific installation. This ratio depends largely on the amount of electricity generated locally through solar, wind or other energy sources, the number of charging points, other loads connected to the sub-grid and the efficiency of the power conversion system.

Energy storage system boosts electric vehicle fast charging infrastructure

Figure 2. Power transformation of future EV charging stations

In this calculation, the capacity of the energy storage system should be between 500 kWh and 2.5 MWh, with a peak power capacity of up to 2 MW.

Having identified the key components of the charging station above (source, load, energy buffer), we next analyze the four power conversion systems that form the energy paths in the charging station.

The four power conversion systems are based on the main DC bus and are rated at 1000V to 1500V DC. The higher the required power, the higher the DC bus voltage. 1500V DC represents the industry standard today and for the next 20 years. While it is possible to impose higher voltage requirements, this would complicate safety regulations, power components and system design, making existing technology inefficient. But this is not to say that in 10 years, new technologies such as power switches, protection systems, etc., will not be able to achieve DC voltages of 2000 V or higher.

Taking the photovoltaic inverter as an example, we see that it has dual functions, one is a DC-DC converter (for the power path from the photovoltaic panel to the DC bus), and the other is a DC-AC inverter (for the Power path from PV panel to AC bus to grid). The DC-DC conversion stage is the most important here, as the AC-DC stage can also be integrated into the main bidirectional power factor correction (PFC) inverter from the DC bus to the AC grid. For the latest power electronics designs, the highest efficiency is achieved with converters based on silicon carbide (SiC) power MOSFET designs. Comparison with silicon insulated gate bipolar transistors (IGBTs) shows an efficiency improvement of 5% (maximum load) to 20% (part load). In our example, using a PV inverter rated at 500 kW, a 5% increase in efficiency means a 25 kW reduction in losses, or a 25 kW increase in power output, equivalent to the energy consumption of five houses or a large heat pump Energy to produce hot water or cool charging station buildings in summer.

Highly similar calculations can be made for both DC charging piles and energy storage system chargers. In both cases, two design approaches are possible: parallel use of large monolithic power converters rated greater than 100 kW or multiple small converters rated 25 kW to 50 kW. Both solutions have their advantages and disadvantages. Today, thanks to economies of scale and simplified design, costs are falling, and multi-connection of small converters has become the mainstream of the market. Of course, an intelligent energy management system must be employed.

Even for these DC-DC converters, switching from silicon IGBTs to SiC MOSFETs brings huge efficiency advantages, as well as space and weight savings, but at a slight price increase—currently a 25% increase, expected in the next five years Annual meeting dropped to 5%. The efficiency gains alone are more than enough to offset the slightly increased costs (assuming a 5% increase in costs at maximum load):

Energy storage system boosts electric vehicle fast charging infrastructure

Finally, in a PFC inverter, 5% of 1 MW is 50 kW, and a total of 250 kW can be saved simply by using more efficient SiC instead of IGBTs. This is equivalent to adding a charging pile, or it may be possible to better balance the overtime energy consumption with the actual load demand.

As we said, getting these results requires SiC MOSFETs, but they don’t solve the problem alone. How the SiCMOSFET is driven is key to achieving the desired switching frequency, which determines the optimal balance between system design cost (influenced by MOSFETs, coils, and inductors) and efficiency. The designer has set a target switching frequency range of 50 kHz to 250 kHz. The requirements for gate drivers are getting higher and higher, mainly in terms of shorter propagation delay and better short-circuit protection.

ADI’s ADuM4136 is an isolated gate driver using the latest iCoupler® technology. This isolation technique achieves a common-mode transient immunity (CMTI) of 150kV/µs, driving SiC MOSFETs at switching frequencies in the hundreds of kHz. Coupled with fast fault management features such as desaturation protection, designers can properly drive single or parallel SiC MOSFETs up to 1200 V.

Isolated gate drivers must be powered, and in ADI application note AN-2016 we show how the combination of the ADuM4136 gate driver and LT3999 push-pull controller can be a noise-free and efficient building block for proper SiC management MOSFETs. The LT3999 is used to control the bipolar isolated power supply of the ADuM4136. The LT3999 isolated power supply features an ultra-low EMI noise design with switching frequencies up to 1MHz, enabling a cost-effective, compact solution.

The total propagation delay (including dead time and propagation delay) is 226 ns when on and 90 ns when off. The delay time of the driver is 66 ns when on and 68 ns when off, and the dead time is 160 ns when on and 22 ns when off.

Ultra-high power densities can be achieved in power converters without sacrificing efficiency.

Figure 3. ADuM4136 and LT3999 gate driver unit

While the power converter is the foundation of the power conversion path, in an energy storage system, the key component to ensure the best total cost of ownership is the battery management/monitoring system (BMS). We found by splitting prices that for megawatt-scale energy storage systems, more than half of the cost comes from the battery rack: currently around $200/kWh, which is expected to drop to $100/kWh by 2025. Having a reliable and accurate BMS solution can extend battery life by 30%, save huge costs and simplify the operability of the entire charging station. Less maintenance means longer working hours, no problems for users, and fewer risks associated with repairs, resulting in a higher level of safety.

To achieve these effects, the energy management system responsible for controlling the energy flow of the charging station must have a very accurate knowledge of the SOC and state of health (SOH) of the energy storage battery. Accurate and reliable SOC and SOH calculations can extend battery life by up to 10 to 20 years, typically by 30%, without increasing the cost of BMS-related electronics. Due to the extended battery life, operating and owning costs can be reduced by at least 30%. Coupled with more accurate SOC information, we can use all the energy stored in the battery to charge the battery in an optimal way, eliminating overcharge or overdischarge; overcharge and overdischarge problems can be drained in a very short time battery power, causing short circuit, fire and other dangers. To enable predictive maintenance, ensuring that energy and power flow are properly managed, knowing battery SOC and SOH means predicting and adjusting grid stability, EV charging processes, and vehicle-to-grid (V2G) connectivity (where vehicles are also considered storage devices) ) used in various algorithms.

The way to achieve accurate monitoring is to use a multi-cell (up to 18 cells) battery monitoring IC with a total measurement error of less than 2.2 mV. All 18 cells can be measured in 290µs, with a lower data acquisition rate selected for noise reduction. Multiple stack monitoring devices can be connected in series to monitor long strings of high voltage cells simultaneously. Each stack monitor has an isolated serial peripheral interface (isoSPI) for high-speed, RF-immune, long-distance communication. Multiple devices are daisy-chained with a single host processor for all devices. The daisy chain operates in both directions, ensuring communication integrity even if the communication path fails. The battery stack can power the IC directly, or it can be powered from an isolated power supply. The IC features passive equalization and separate PWM duty cycle control for each cell. Other features include an on-chip 5 V regulator, nine general-purpose I/O lines, and sleep mode (in which power consumption is reduced to 6 μA).

For BMS applications with short-term and long-term accuracy requirements, the IC uses a buried Zener-converted reference rather than a bandgap reference. This provides a stable low drift (20 ppm/√kh), low temperature coefficient (3 ppm/°C), low hysteresis (20 ppm) primary voltage reference, and excellent long-term stability. This precision and stability is critical and underlies all subsequent cell measurements, and these errors have a cumulative impact on the confidence of the acquired-data, algorithm consistency, and system performance.

While a high-accuracy voltage reference is a necessary feature to ensure excellent performance, it is not sufficient on its own. The AC-DC converter architecture and its operation must comply with the electrical noise environment, which is a result of the pulse-width modulation (PWM) transient characteristics of the system’s high-current/voltage inverters. Accurate assessment of a battery’s SOC and SOH also requires associated voltage, current, and temperature measurements.

To mitigate system noise before it affects BMS performance, the converters used inside the stack monitor use a sigma-delta topology with six user-selectable filter options to assist in handling the noisy environment. The sigma-delta method reduces the effects of EMI and other transient noise because of its nature to use multiple samples per conversion, with averaging filtering.

In ADI’s product portfolio, the LTC681x and LTC680x families represent the state-of-the-art in battery stack monitors. The 18-channel version is the LTC6813.

In conclusion, to meet the challenges facing the future DC fast charging infrastructure, power conversion systems and energy storage systems are key. We give two examples combining the ADuM4136 isolated gate driver with the LT3999 power controller (for power conversion stages designed with SiC MOSFETs) and the LTC6813 battery monitoring device (for energy storage batteries). In fact, there are many more areas of these systems to focus on, including from current metering to fault protection devices, from gas detection to functional safety, they are extremely important and can bring many benefits, and Analog Devices is currently actively developing all of them. These subsystems ensure that we can sense, measure, connect, interpret, protect and drive all physical phenomena, resulting in reliable and robust data. High-end algorithms will use this data to ensure that most of the energy is converted from renewable sources to load (in this case, electric vehicles).

About the Author:

Stefano Gallinaro joined Analog Devices in 2016 and works in the Renewable Energy business unit. He is responsible for managing strategic marketing activities in solar, electric vehicles, charging and energy storage, with a particular focus on power conversion. Based in Munich, responsible for global business.

Stefano is pursuing a bachelor’s degree in electrical engineering at the Politecnico di Torino, Italy. He started his career as an application engineer at STMicroelectronics Srl-DORA SpA in Aosta, Italy. Prior to joining ADI in 2016, he worked for two and a half years as Product Marketing Manager at Vincotech GmbH in Anderach, Germany.

The Links:   MG100Q2YS51 LQ150X1LGN2E