Artificial Intelligence for Hydropower-plants

Empower your decision-making capabilities for operations and maintenance, by exploiting the insights within your data.

What is Tyde

Tyde is a cloud-based software product running on your preferences. It is a home for your data and a place where your data makes sense, in the context of Hydropower. You can use the insights produced by Tyde to plan maintenance smarter, reduce downtime and avoid manual inspection rounds.

How It Works

Using state-of-the-art Artificial Intelligence techniques, operators gain insight into the states of the various assets that the Hydropower-plants consists of.

Sensor Monitoring

Asset Diagrams

Machine Learning

Data Aggregation

Sensor Monitoring

Tyde connects to sources such as SCADA-systems , local control equipment or third-party IoT telemetry devices. Many protocols are supported, for example OPC-UA , AMQP and MQTT. The data typically consists of time-series , control-signals and source-metadata. This is all taken care of by Tyde, it is stored and put into a context of what they are measuring and controlling. You can visualize both the raw-data directly, and also the aggregates and analytical results that are being calculated on the fly, using the Tyde web-portal.

Asset Diagrams

Hydropower-plants consists of various equipment and tools, which we refer to as assets. The assets have a hierarchical relationship with each other, which defines the context of the Hydropower-plant. The context can be built manually using tools available from the web-portal, you can use already-made power-plant templates or upload Excel-files containing the necessary information. The resulting asset-diagram can then be seen in the web-portal , and associations between data-sources and the assets can be made. Support for the IEC-81346 standard, which RDS-Hydro is based on, is included. In particular, the functional-aspect.

Machine Learning

Our sophisticated machine-learning algorithms is continuously monitoring the asset-diagrams and the realtime data entering the system. So-called “online-learners” are keeping our neural-network based asset-models up to date with the newest data. Data-scientists or other advanced operators may use the portal to create tailor-made models by overriding the default learning parameters for a selection of assets. Models are under version control, and you can jump back and forth between various available models for any given assets. The output of the models is a metric which describes the degree of anomaly which any given asset is currently in. The anomalies, their strength and their frequencies are further analysed in order to create actionable decision-points, which can be connected to a notification mechanisms so the operator can receive reports and event-information as e-mails.

Data Aggregation

Aggregated time-series data can often be easier to consume and understand, rather than the raw-data itself. Therefore, our automated statistical engines will create aggregates to make it easier to understand long-term-trends , variations and drifts. The operator can drill-down from the eagle-eye view all the way down to the raw-data itself, to reach the preferred level of detail. If you want to look at various sources of data in the context of each other, you can create ‘sensor-groups’. These groups can be stored for easy access. The statistical engines will see them, and then create correlation matrices to give some indication on how the sources appear to be varying with each other.

How We Do It

Tyde is a web-based system aimed at strengthening your decision-making foundation. But it can also be seen as an ecosystem of various micro-services that work together in a cluster, in the cloud. They collaborate on turning the information in the data that are entering the system in real-time and the historical data with information about the past, into actionable insights about the Hydropower plants.

How We Help

Tyde can be set up to run on anything, anywhere. We will assist you to set up, install and configure it based on your own needs and preferences. We recommend using Microsoft Azure, although it is not a strict requirement. To make the low-level arrangements on-site to make sure we can map the data safely all the way from the analogue sensors to the Tyde deployment in the cloud, our partners are eager to contribute.

Want to know more? Sign-up here and we’ll get in touch.

Why Tyde

Tyde is built from the bottom up based entirely on real business needs. We have assumed nothing here, everything is based on dialogue and feedback from real customer on real deployments. Chances are, many if not all your hydropower needs, matches those that our pilot customers has and that means they have already been covered. Therefore, the threshold to get Tyde to run for you is low, both in terms of price and time.

How We're Different

Tyde is built by the Broentech Solutions, a Norwegian deep-tech company. We are a small and agile team, consisting of developers and data-scientists that work in an environment where staying up to date on the bleeding-edge is a core feature. This, in combination with very tight relationship to our customers within Hydropower-production has lead to what is now Tyde, the go-to solution for predictive maintenance focusing entirely on Hydropower.
Tyde has an autonomous A.I. engine, developed in-house. It adapts to the context of the Hydropower-plants, as defined by the operator. The engine itself supports various learning-algorithms, although it is the neural networks that is currently giving the best results. The online-learners have a good default set of parameters, that can be overridden by an operator. Models are under version control, so that the operator can go back and forth between available models.
Tyde is analyzing the performance of the assets. The operator may define rules that maps how SCADA systems initializes sequences of various tasks. Using these rules, Tyde listen for the corresponding signals and create trends that describe drifts in performance. Such drifts in performance are often correlated to drift in asset health.
Previous
Next

Latest News

Join our newsletter to get the latest in your inbox.

Artificial Intelligence in Tyde

This is a post, with a video, that explains how we are using AI in Tyde.

A general presentation

This is a video-blogpost containing a general presentation of Tyde

The Tyde portal – an introduction

A brief introduction to the things you can do with the Tyde portal.

Send us your contact info and we'll get back to you with more details.