Transportes Delgado
Areas of expertise:- Data Analysis
- Machine Learning
Use of machine learning to implement an algorithm for predicting system failures in heavy-duty trucks
Achievements:- Cleaning of the database
- Data analysis and visualization
- Implementation of machine learning algorithms
- Comparison of the implemented algorithms
- Advice for the future management of the database
- Scikit-learn
- XGBoost
- Statsmodels
- Pandas
- Plotly
Private Accommodation
Areas of expertise:- Home Automation
- Internet of Things
Connection of smart devices and automation of tasks
Achievements:- Setting up of a Raspberry Pi 4 with Ubuntu Server
- Installation of the Home Assistant and ESPHome software as Docker containers on the server
- Implementation of NodeMCUs to use sensors and relay modules, and integration of these Home Assistant with ESPHome
- Implementation of ConBee II to enable communication with Zigbee devices and integration with Home Assistant
- Implementation of smart devices from several brands (TP-Link, Tuya, Aqara, etc.), and integration of these with Home Assistant
- Implementation of IoTaWatt on electrical installations to measure energy consumption and consumption and integration of the same with Home Assistant
- Installation of InfluxDB and Grafana as Docker containers on the server
- Implementation of InfluxDB time series databases to save the values of the devices and Grafana for the visualization of these values
- Raspberry Pi
- Ubuntu Server
- NodeMCU
- ConBee II
- Docker
- Home Assistant
- ESPHome
- IoTaWatt
- InfluxDB
- Grafana
Institut National de l'Énergie Solaire (INES)
Areas of expertise:- Machine Learning
- Deep Learning
Addition of deep learning to a power consumption prediction tool
Achievements:- Implementation of deep learning models for electricity consumption forecasting
- Comparison of deep learning models with existing XGBoost models
- Optimization of features for forecasting
- Analysis of the features importance
- Optimization of model hyperparameters
- Analysis of the hyperparameters importance
- Experimentation of the transfer learning method for scenarios with scarce training data
- Pytorch
- TensorFlow
- XGBoost
- Scikit-learn
- Optuna
- Captum
- SHAP
- Pandas
- Plotly
Laboratoire de Génie Electrique de Grenoble (G2Elab)
Areas of expertise:- Energy Efficiency
- Renewable energy
Simulate several restructuring scenarios of an electrical microgrid in order to compare the energy efficiency
Achievements:- Cleaning of the database
- Data analysis and visualization
- Implementation of an object-oriented approach to simulate different scenarios
- Simulation of the integration of energy storage into the microgrid in order to compare the energy consumption from the main grid
- Simulation of the integration of electric vehicles into the microgrid in order to compare the energy energy consumption from the main grid
- Simulation of the integration of both energy storage and electric vehicles into the microgrid to compare the energy consumption from the main grid
- Simulation to find the optimal charging time for electric vehicles
- Comparison between the simulated scenarios
- OMEGAlpes
- multiprocessing
- Pandas
- NetworkX
- Plotly
Grenoble INP-ENSE3
Areas of expertise:- Energy Efficiency
- Renewable energy
Design of an electrical distribution network and simulation of the implementation of new renewable energy sources
Achievements:- Data analysis and visualization
- Simulation of solar energy production
- Simulation of wind power generation
- Modeling of the electrical distribution network
- Power flow calculation for the modeled electrical distribution network
- Financial analysis of the cost of electrical components for the modeled electrical distribution distribution network
- Pandapower
- pvlib
- windpowerlib
- Scikit-learn
- Pandas
Laboratoire d’Informatique de Grenoble (LIG)
Areas of expertise:- Hardware-in-the-Loop Simulation
- Electrical faults
Hardware-in-the-loop simulation for the testing of intelligent electronic devices
Achievements:- Modeling of an electrical distribution network in Python
- Test the connection of the simulation to intelligent electronic devices using TCP/UDP communication protocols
- Running of the simulation in "Multiprocessing" mode to manage the modeled electrical distribution network distribution network and monitor the signals of the intelligent electronic devices simultaneously
- Implementation of "Queues" to share variables between multiprocesses
- Implementation of a graphical interface to manage the simulation
- Pandapower
- socket
- multiprocessing
- Tkinter
Sistema Tecnológico de Monterrey
Areas of expertise:- Energy Efficiency
Ensure that the electrical consumption of several buildings meets the forecast
Achievements:- Collection of meteorological variables necessary for the forecasting of electricity consumption
- Update the number of people visiting the buildings daily in the databases to obtain more accurate forecasts
- Notification to building staff regarding significant deviations from expected electric consumption
- Pandapower
- socket
- multiprocessing
- Tkinter