Smart embedded system for prediction of the state of charge of the battery of electric vehicles, scheduling of recharging and integration into buildings

Abstract

The present invention relates to a smart embedded system for prediction of the state of charge of the battery of electric vehicles, scheduling of recharging and integration into buildings, comprising: means for collecting, in real time, data from the vehicle such as GPS position, road traffic and weather conditions; means for monitoring the consumption and production of electricity by renewable resources in real time; means for processing these fed-back data in order to train a machine learning algorithm to predict the state of charge of the battery of the electric vehicle; actuators for controlling the active equipment or electricity-consuming devices in line with the prediction result. The invention also relates to a method for implementing this system, comprising the steps of collecting data, in particular vehicle data, building data and weather data; processing the collected data using a prediction algorithm and making a decision in respect of management of electrical equipment in the building and of charging of the batteries of the vehicle.

Description

[0001] An intelligent embedded system for predicting the state of charge of electric vehicle batteries, scheduling charging and integrating them into buildings.

[0002] Field of invention

[0003] The present invention relates to the field of battery charging methods, specially adapted to electric vehicles and to the exchange of energy storage elements in electric vehicles.

[0004] Background to the invention

[0005] Currently, electric vehicles are considered a building load. The vehicles are charged overnight and used during the day for various use cases. Furthermore, green buildings or microgrids rely primarily on renewable energy sources to cover their energy needs and use batteries for storage. Given the large battery capacity designed to power the vehicle's electric motors, EVs can be used as a secondary storage medium in buildings.

[0006] Indeed, there are bidirectional charging stations that allow the transfer of electrical energy between the building and the EV in both directions, but this mechanism remains manual or semi-automatic. Thus, the user must decide whether to use the vehicle as a load or a storage medium. This decision is based on the current state of energy production / demand in the building and does not take into account how these parameters may evolve in the future.

[0007] Summary of the invention (abbreviated)

[0008] The invention aims to integrate the electric vehicle into buildings, particularly those equipped with renewable energy sources. The proposed solution consists of an embedded system to be deployed at the electric vehicle level to collect data relating to the operation of the EV and ensure real-time data feedback to a remote server. The building also has a system that allows it to receive data from the electric vehicle in real time from the server. This information will allow the building to optimize the use and storage of the energy produced at the building level. In this way, the building can use the electric vehicle battery as a means of storing the excess energy produced. However, it can use the vehicle battery to mitigate consumption peaks during periods characterized by low efficiency of renewable energy sources.The operation of this auxiliary storage source takes into consideration the constraints linked to the state of health of the battery and the planning of vehicle use.

[0009] The architecture of this invention is based on two systems: one embedded in the electric vehicle and another system installed in the building. The first system offers two main functionalities:

[0010] • Real-time collection of vehicle, traffic and weather data. The data is then pre-processed, stored locally and sent to the server.

[0011] • The use of data collected to train a machine learning algorithm to predict the state of charge of the electric vehicle battery. This information is subsequently used in applications to minimize energy consumption, increase vehicle range and improve driving conditions.

[0012] The second system deployed in the building offers three functionalities:

[0013] Monitoring energy production and consumption in the building. To achieve this, several sensors are deployed to monitor the energy demand of various machines and the efficiency of renewable energy resources. This data flow is recorded in a local database.

[0014] • Training a machine learning model based on stored production / consumption data to predict future consumption peaks and periods of energy overproduction.

[0015] • Combining prediction and electric vehicle operating information to optimize vehicle integration into the building and use it as an auxiliary energy storage medium.

[0016] Identification of the invention

[0017] The present invention aims to overcome these constraints and improve energy management in buildings by using a system that integrates new technologies in the field of Internet of Things and automatic machine learning.

[0018] Indeed, the integration of an intelligent system would allow the complete automation of the decision-making mechanism as well as the optimal use of energy resources. The collected data (from the electric vehicle or the building) can be used to train a machine learning model to predict the rate of energy production and consumption and make a decision to ensure good energy management in the building. The system in the vehicle sends the current data and the predicted state of charge of the battery to the system installed in the building. The latter, based on the data received from the car and the data collected locally, makes a decision to minimize energy consumption whether in the building or in the electric vehicle.When the vehicle's onboard system receives the decision to minimize energy consumption, it will recommend new actions to drivers such as optimal route selection, speed control or others to reach the destination with the minimum energy consumed.

[0019] Description of the drawings and embodiment of the invention

[0020] The method used to carry out the invention is as follows: In the building:

[0021] • Total household or building electricity sensor.

[0022] • Sensor of electricity produced by different renewable energy sources.

[0023] • Microcontroller which allows the data collected by the various measuring instruments to be collected.

[0024] • Actuators for controlling active equipment or electricity-consuming devices. • Integrated compact graphics card for embedded systems for deploying artificial intelligence applications, data analysis, storing and sending data to servers, and also for real-time prediction of future electricity consumption and production.

[0025] • Graphical interface allowing users to visualize their electricity consumption in real time.

[0026] In the electric vehicle:

[0027] • A module for reading CAN messages from the OBDII port of the electric car in order to read the values ​​of the various sensors installed in the car.

[0028] • A GPS sensor to collect position data from the electric vehicle.

[0029] • An embedded system that allows the various sensor data to be read and processed, recorded and sent in real time.

[0030] • A graphical interface to display data to the user.

[0031] • A modem to be able to send data and collect weather and traffic data from the internet.

[0032] At the building level, the various electricity sensors used are used to track current electricity consumption or production. This data is then preprocessed and stored and subsequently used to train a machine learning model. This model is used to predict the future electricity consumption and production profile in order to anticipate consumption peaks and power outages and make a decision to properly manage the energy stored in the storage batteries used.

[0033] At the vehicle level, the on-board system collects vehicle status information via the OBDII port using a CAN bus transcoder. Other information such as weather and traffic conditions is provided by online services using APIs. The collected data is then pre-processed and recorded to train a machine learning model. This model is used to predict the state of charge of the traction battery. The collected and predicted data is then sent to the server to make comprehensive decisions.

[0034] The role of the components of the developed solution:

[0035] At the server level (intelligent system), the received data is stored and analyzed in order to make appropriate decisions.

[0036] At the building level, decisions can be in the form of suggestions for:

[0037] • Choose the mode and planning of use of certain electrical devices, encourage the recharging of batteries in the event of a predicted drop in electricity production.

[0038] • Schedule the charging of electric vehicles, and use the vehicle as an energy source to mitigate consumption peaks. At the vehicle level, the system can make decisions to minimize the car's electrical consumption and thus ensure that the user will arrive at their destination with the minimum possible energy.

Claims

Claims:

1. Embedded intelligent system for predicting the state of charge of electric vehicle batteries, scheduling charging and integrating them into buildings, including: Means of collecting, in real time, vehicle data such as GPS position, road traffic and weather conditions. Means to monitor electricity consumption and production from renewable resources in real time Means of processing this data collected for training a machine learning algorithm for predicting the state of charge of the electric vehicle battery. Actuators for controlling active equipment or electricity-consuming devices following the prediction result.

2. Intelligent on-board system for predicting the state of charge of the battery of electric vehicles, characterized in that the vehicle data collection means comprise: A module for reading CAN messages from the OBDII port of the electric car in order to read the values ​​of the various sensors installed in the car.

3. Intelligent on-board system for predicting the state of charge of the battery of electric vehicles characterized in that a graphical interface for displaying the data to the user.

4. An intelligent on-board system for predicting the battery charge status, characterized in that the weather and road traffic status are provided by online services using APIs, 5. Method for predicting the state of charge of the battery comprising the steps: Real-time vehicle and weather data collection Collection of data on consumption and production of electricity from renewable sources Processing of collected data using a machine learning and state of charge prediction algorithm. Choose the mode and schedule for using certain electrical appliances, and encourage battery charging if a drop in electricity production is predicted. Plan the charging of electric vehicles and use the vehicle as a source of energy to mitigate consumption peaks.

Images