1 of 10

ML service app “EnergyProdPredict”

by Engenir.ai

</

/>

Deeptech GigaHack 2023

1 0 1 1 0 1 1 0 1 1 0 1 1 0 0 1 1 0 1 1 0 1 1 0 1 1 0 1 1 1 0 1 1 0 1 1 0 1 1 1 1 1 0 1

2 of 10

</ Issue Description

Moldova's renewable energy sector faces several critical challenges:

  • Inaccurate Production Forecasting: Without precise predictions of energy generation from solar and wind power plants, producers struggle to make informed decisions regarding energy sales and market transactions, leading to inefficiencies and financial risks.

​

  • Risk of Adjustments: Inaccurate forecasts can result in costly adjustments through transactions on the balancing market, impacting the profitability and reliability of renewable energy production.

​

  • Market Streamlining: The absence of reliable production forecasts hampers the automation of bidding processes for electricity sales, hindering the growth and competitiveness of Moldova's renewable energy sector.

​

  • Sustainability: Effective management of renewable energy generation is vital for Moldova's sustainable energy future and reducing its carbon footprint.

​

1 0 1 1 0 1 1 0 1 1 0 1 1 0 0 1 1 0 1 1 0 1 1 0 1 1 0 1 1 1 0 1 1 0 1 1 0 1 1 1 1 1 0 1

3 of 10

</ Solution

Product demo

Machine learning prediction application

Integrates crucial Data Sources:

  • Location
  • Production technology (solar/wind)
  • Installed capacity of electrical installations
  • Technical specifications of generating equipment
  • Historical weather data
  • Historical energy production data
  • Other relevant factors

Predictive Algorithms

ML algorithms, such as time series analysis, regression, neural networks, are used to identify patterns and correlations in the data and make highly accurate hourly and daily forecasts of energy generation from renewable sources.

Advanced Forecasting

Accurate prognoses will empower producers to make well-informed decisions about the quantity of electrical energy available for sale on the Day-Ahead Market.

4 of 10

</ Competitive advantages

Enhanced Decision-Making

Producers will have access to accurate forecasts, enabling them to optimize energy sales, reduce market transaction risks, and streamline bidding processes.

Financial Efficiency

By reducing the need for costly adjustments in energy transactions will enhance the financial viability of renewable energy production in Moldova.

Sustainability

Accurate predictions will contribute to the sustainable management of renewable energy enterprises, supporting Moldova's commitment to green energy development.

Competitiveness

The project will increase the competitiveness of Moldova's renewable energy sector, attracting investment and fostering industry growth.

5 of 10

</Challenges approach

Time challenge

We developed the service rapid

Usability Challenge

We created an intuitive and easy-to-use multifunctional UI

Best Choice

We choose two best models and trained them ideal

Creativity challenge

We tried many models using different hyperparameters tune

Software Development Challenges

We developed our web-application using which user can retrain model with new data or make a needed prediction

6 of 10

</Revenue model

​

Consulting and Customization Services

Subscription Fees

Partnerships and Collaborations

Market Expansion

7 of 10

</Functional diagram

Web-app

Making prediction

Client Data

Training model

Web-app

Download data in csv format

8 of 10

Random Forest

9 of 10

</Tehnologies

10 of 10

</Team

Thank you for attention! >

Alexey Megley

Teamlider. Backend, ML

Mihai Coshlets

DA(pandas-profiling), making endpoints using Flask and Streamlit, Data Engineering

Serghei Safoshkin

KNIME, machine learning

Margarita Marfoi

DA, UI (Streamlit)