Jungle AI - Toucan

Accuracy meets automation

Jungle’s forecasting model reduces the cost of variability of the power output of your wind and solar farms, producing highly accurate and reliable power forecasts based on asset telemetry and numerical weather forecasts.

Accuracy

Our state-of-the-art AI models leverage SCADA data on top of various numerical weather models to deliver accurate probabilistic power forecasts for your renewable energy assets.

SCADA data

Jungle’s tailored models are trained with contextual operational data directly from your farms.

Numerical weather forecasts

Several weather forecasts are ensemble to ensure that the best power forecasts are obtainable.

AI models

We developed next-gen deep learning AI models that can leverage and learn from very large quantities of data.

Automation

Automation is required to provide maximum flexibility for our users, whilst offering our services at competitive cost levels. Our personalised models and predictions are made available through a programmatic interface (API), generated by Jungle for each of your wind and solar farms.

Precise power predictions

Experience the transformative potential of Toucan's state-of-the-art deep learning models, guiding your success in open energy markets. Jungle’s AI models uncover intricate patterns in your data in order to generate reliable forecasts, allowing you to benefit from reduced imbalance costs.

Get started

"Working with Jungle allowed us to validate and obtain a more robust power forecasting estimation for a challenging wind project in the Indian Ocean that EDF Renouvelables is working on."

Noé Le Guerrannic

Chef de projet ingénierie_, EDF Renouvelables

Power forecasts delivered how, where and when you need them

Fully adaptable modelling strategy, learning and relearning behaviour at any time, for any farm around the world, at your request.

Get started

Higher precision at competitive levels

Cost-effective, reliable and accurate power forecasting models for individual renewables assets, which reduces the imbalance costs, saves penalties and improves the profitability of your farms.

Backtesting

Using our API to request historical forecasts, you can easily backtest the performance of the model and the tailored behaviour for your farm.

Forecasting windows & resolutions

Request different forecasting windows and resolutions, which can go from intraday to several days ahead and resolutions that can go down to minutes.

Probabilistic Predictions

Our probabilistic predictions allow you to have deeper insights and greater flexibility when bidding. Query different quantiles to adjust for custom penalisation schemes or curtailment implications.

How to get started

Your farm's details
and data

Start by providing us with your asset’s details (name, size and location) and SCADA data, and define the forecasting requirements.

AI models and API
setup

Behind the scenes, we run our tailored AI models and set up your API. After that, you receive an API key with the proper access credentials and parameters.
‍

Accurate forecasts in a heartbeat

All you have to do now is query the forecasting window and resolution, and get your power forecasts in a matter of seconds.

Frequently asked questions

What are your AI models doing, how are weights defined, etc?

We use state-of-the-art deep learning strategies. In the context of power forecasting, this allows the discovery of complex patterns in the data that are often missed by domain experts and simpler models, leading to more accurate forecasts.

What kind of weather models do you use?

We ensemble several weather forecasts from multiple forecaster locations around the farm to ensure that our models have enough context of the weather around the entire project.

What information do you need to get started?

Before getting started, we need some information about the project, most importantly the location of the farms. Next we need to understand your requirements on forecasting frequency, intervals, horizon and if you require quantile predictions. Reach out if you would like to receive the complete requirements list.

How much historical SCADA data is required for your forecasts?

We use SCADA data to tailor our models. For precise power forecasts, the most important signal is the production values. As seasonality plays a big role in forecasting energy production, our models perform better when they can learn from multiple seasons. Two years of data is ideal, but don’t hesitate to get in touch in any case.

What’s the advantage of using your API solution?

Our API allows you to integrate our forecasts in your existing processes. It’s easy to configure and once everything runs, you don’t have to worry about fetching forecasts.

Do you offer free trials?

Yes, we do! Get in touch with our sales team to discuss how and when to get started.