EV charging - digital architectures

From Electrical Installation Guide
HomeElectric Vehicle chargingEV charging - digital architectures

EV Charging Power Management

EV charging is a new kind of electric load, with unique characteristics in terms of where and when it is used, but also in terms of variable power and energy demands. It is worth noting that its connection to new or existing electrical installations may have a significant impact on the overall power distribution system.

Power management of EV charging stations is therefore essential, in order to minimize the impact on the existing or new electrical infrastructure while distributing available energy between all connected loads.

There are 3 main levels of power management, depending on the set objective:

These power management levels are detailed in this section. You can also check this example with 3 EV infrastructure scenarios: it is the same application, implemented without Load Management System (LMS), with static LMS and with dynamic LMS, to illustrate how it can impact the electrical installation sizing.

Fig. EV45 – The different levels of EV charging power management

EV charging power management via static setpoint

In static mode, the Load Management System regulates and distributes energy evenly and in real time between all connected vehicles, so as not to exceed the general STATIC setpoint for the vehicle loads.

Example: In a building, a static setpoint of 100 kVA is defined as the power available for the EV chargers, and there is a requirement to install 10x 22 kVA charging points. With the energy management system, regardless of the number of terminals (EV chargers) that are being used simultaneously, it ensures that the 100kVA limit is never exceeded, and that any risk of tripping is avoided.

The current setpoint for each of the charging points is transmitted in real time to the electric cars, which have 5 seconds to apply it. If this instruction is not applied by the car, then the charge point contactor will be instructed to open the circuit.

This allocation method allows you to:

  • Evenly distribute available energy between all vehicles that are being charged
  • Sequence loads between the vehicles that are connected simultaneously
  • Optimize occupant comfort by ensuring that the main power supply does not trip as a result of an influx of vehicles requiring a recharge
  • Reduce cost and dimensions for electrical panel dedicated to the power supply of the electric vehicle charging network (100 kVA in this example)
Fig. EV46 – EV load management with static setpoint

EV charging power management via dynamic setpoint

In dynamic mode, the Load Management System allocates available onsite energy in real time to the electric vehicle charging network. In doing so, it also temporarily limits the charging power, to meet the energy constraints imposed by the rest of the electrical installation. Conversely, the power allocated may be higher at times when energy consumption for the rest of the electrical installation is low.

Example: The total power of the building is 250 kVA and the objective is to install 10x 22kVA charging points. With this system, whatever the load of the building and number of terminals (EV chargers) used at the same time, total consumption must never exceed 250kVA by instructing the terminals to adapt in real time to the other loads of the building.

The current setpoint for each charger is transmitted in real time to the cars, which have 5 seconds to apply it. If this instruction is not applied by the car, then the charge point contactor will be instructed to open.

This allocation method allows you to:

  • Evenly distribute available energy between all vehicles being charged
  • Sequence the loads between the connected vehicles simultaneously
  • Optimize occupant comfort by ensuring that the main power supply does not trip as a result of an influx of vehicles requiring a recharge
  • Control energy costs by subscribing the optimal energy contract from the energy supplier (may not be applicable in countries where the energy contract has no limit)

To determine, in real time, the DYNAMIC setpoint allocated to the charging infrastructure, the system must measure the available power at building level.

Fig. EV47 – EV load management with dynamic setpoint

Advanced (Artificial Intelligence-based) EV charging management

Advanced AI-based EV charging management is used to generate an optimal dynamic setpoint based on a number of criteria. These include electric vehicle planning, grid energy tariffs, prediction of building consumption, and prediction of local energy sources production, if any.

Advanced AI-based EV charging management usually relies on two techniques: forecast and model predictive control.

Fig. EV48 – Advanced AI-based EV charging management with forecasting and predictive control

Forecasting energy consumption, EV charging needs and local production

A forecasting component is used to predict local energy demands and local energy production. This enables short-term energy resource planning and optimized local energy use.

The forecasting component uses supervised machine learning techniques to learn the relationship between the variables at hand and the variable we intend to forecast.

Photovoltaic production forecasting is similar to the solar radiation forecast provided by a weather forecast service.

Building energy consumption and EV charging needs can be forecast based on historical energy consumption that identifies recurrent patterns. This forecast can be improved by adding additional drivers like weather forecast information or EV charging planning.

The accuracy of the forecast is critical for an optimal Model Predictive Control.

Model Predictive Control

Model Predictive Control (MPC) techniques can be used to optimize energy usage over the following 24 hours, by anticipating energy demands (EV charging and other loads) as well as local renewable production.

The Model Predictive Controller relies on:

  • A model with the description of the electrical network, the assets’ characteristics and constraints that should be respected.
  • Forecasts over the following 24 hours including the energy consumption of the installation, EV charging demand, photovoltaic production, and grid energy tariffs.
  • Knowledge of the assets’ current state, for example, the state of charge of the electric cars or Energy Storage System.

By updating the local controller based on latest site measures, and updating forecast information every 15 minutes, the AI-based management can continuously adapt to prediction and model errors, to ensure optimal closed-loop control performance.

EV charging integration in the Microgrid

Microgrids are integrated energy systems consisting of a group of interconnected Distributed Energy Resources within clearly defined electrical boundaries that act as a single controllable entity with respect to the grid.

Fig. EV49 – Elements of a microgrid

New EV loads need to be managed according to the operational activity of the building, and at the same time they have to be coordinated with the other energy sources like photovoltaic or battery storage, in order to optimize energy as well as energy-related costs.

In such a scenario, two systems are required to efficiently manage this group of Distributed Energy:

  • A Load Management System is needed to control and split the EV chargers power demand and to manage the charging priority of the fleet.
  • A Microgrid Advisor Solution is required to manage the energy of the different sources, based on the forecast of the fixed loads of building energy and to manage the flexible loads (such as EV and HVAC), to optimize the contract with the utility.

EV Charging diagnostics and maintenance

According to a recent study from AVERE, in France, 96% of respondents said they were satisfied with their electric car, but 68% of them expressed dissatisfaction with the accessibility of public charging points, (mainly due to faults or because non electric cars had occupied the EV dedicated parking space).

Charging stations with advanced features or communications systems like data or payment collection may require more periodic maintenance than a basic unit, simply because there are more components with the potential to malfunction. In many cases, a local electrician can troubleshoot problems with the units. Charge point operators also offer remote options that can reduce long-term maintenance and repair costs.

Therefore, having a plan to protect your EV charging assets is key to the success of your charging infrastructure. This plan includes various activities which can be done on site or remotely:

  • Error diagnostics and troubleshooting
  • Upgrading the charging station with the latest firmware and benefitting from additional features
  • Restoration of factory default settings
  • Changing spare parts

Next level of maintenance is to prevent any unplanned downtime and expensive costs as a result of unexpected equipment failure. Preventive maintenance requires careful planning and scheduling of maintenance on equipment before a problem arises, as well as keeping accurate records of past inspections and servicing reports. This mode of maintenance would be even more efficient when the charging station is connected to a remote maintenance platform.

EV charging supervision

A Charge Point Owner operates a pool of charging points and is responsible for making sure that the network runs smoothly. This role encompasses driver management, diagnostics and device maintenance. Yet the most important feature for operating a charging station is its capacity to connect to a cloud-based service.

Today, the most common protocol for connecting a charging station to a Cloud platform is OCPP. OCPP, Open Charge Point Protocol, is the de facto standard in the EV charging industry, for enabling a connection between hardware and software.

The following figures provides some examples of functions that may be included in a platform for supervising an EV charging infrastructure.

Fig. EV50 – Examples of functions that can be provided by EV charging supervision