Ronin AI
The first AI driven platform to evaluate health status
of large power transformers
What is RONIN AI?
RONIN AI is our cloud platform for the real-time calculation and health monitoring of large power transformers to prevent abrupt power outages due by machine faults and to predict manufacturing power needs.
RONIN AI is a unique digital predictive maintenance solution shifting popular health indexing approach into a scientific and validated machine learning model.
Power Transformers Health Index as Multi Purpose Strategy
CIGRE A2.44 (630) defines transformer health index as “an overall indicator describing the condition status of a transformer that the decision maker can use to define and prioritize the required action”.
CIGRE A2.49 (761) suggests that it can be used “to identify transformers that are most likely to fail on a timescale to plan intervention.”
A well-designed health index can convert strategic condition monitoring data in to useful remarks that can accelerate decision making. We use it to highlight and rank transformers on the basis of their health status within a fleet. It helps in determining asset reliability and intervention priority based on technical urgency.
A health index should not be used to mask alarms generated out of bad condition monitoring data.
CIGRE is a collaborative global community committed to the world’s leading knowledge development programme for the creation and sharing of power system expertise.
500kVA-800MVA range
Wide range of transformers applications
24/7 data availability
View, search, analyze, or download your data anytime
Fast Mass Reporting
Get bulk health index based recommendations
Only 14 indicators needed
Requires chemical data of main oil tank only
Why using Artificial Intelligence on power transformers?
An adequate asset maintenance and management strategy is the only key to avoid failures, unplanned downtime and, above all, extensive damage related to unexpected outages.
Just think of what happened in February 2021, when the great cold paralyzed the energy production and electricity grids of Texas, not only killing dozens and blocking most of the activities but revealing the vulnerability of essential infrastructures and the lack of investment towards a resilient network.
%
Prediction accuracy
Synthetic assets training dataset
Seconds to predict
Real assets trained
Features selected
Lines of code
Trial analysis
Is 80% accuracy level enough?
If you build a machine learning model that fits “too much” to the data you will probably get a model with a very high coefficient of determination. However, a model that adapts too much to a specific set of data, following every slightest variation, is then not very generalizable and with low predictive power.
In statistics, these are termed as over-fitting problems.
We trained our system to reach 80% of accuracy, because we simply do not want to overfit!
This is satisfactory for predictive modeling and very good for real case application because it’s not overfitting on a particular sample of power transformers.
Benefits of industrial transformers health indexing
Industrial transformers are strategic units that can improve the production capacities of metal processing industries such as steel, aluminium, copper, etc.
Industrial transformers are categorized as furnace and convertor transformers. A furnace transformer provides the electrical arc needed to melt the raw materials inside the furnance, such as in iron & steel industries. It is essential for an asset manager to identify and maintain the overall health and reliability of such transformers to avoid unplanned downtime and compromise production capacity.
Trend analysis and asset comparison
RONIN AI allows service engineers and asset managers a quick review of their asset management decisions. This assists them in prioritizing actions and recommendations for the best allocation of budget and resources.
Ronin AI provides health index score for each testing date of the same asset. This allows the asset manager to use the trend analysis feature for individual asset or for comparison purposes.
Type and range of asset
Ronin AI can analyse mineral oil filled transformers rated between 500 kVA and 800 MVA. Based on functionalities and voltage class, they can be power, substation, step-up, step-down, shunt-reactors, industrial, furnace, traction, and auto-transformers.
Smart Manufacturing
Manufacturers can use AI driven ranking tool for better completion of their asset management tasks with major budget saving on time and cost of operations. They can use such strategies to redefine the technical and economical parameters of their asset management actions with greater transparency. Conversely, this will improve their market performance, client satisfaction, and overall business expansion.
Satisfied Customers
Frequently Asked Questions
We collected some of the frequently asked question from our clients and customers during interactions on Ronin AI.
How is Ronin AI different from other software/tools?
Ronin AI is outstandingly distinct because of:
- A machine learning (ML) based algorithm
- Fully automated data management
- Less than 30 minutes of training
Can I use Ronin AI with my condition monitoring suite?
Ronin AI predicts transformer health index based on condition monitoring data using a simple .csv file format. Yes, you can use Ronin AI with your existing condition monitoring suite to rank and highlight your transformers that are most likely to fail within a fleet.
How can I buy Ronin AI predictions?
You can buy in 3 easy steps. Watch the demo, ask for a try out license, and lastly ask for a customized offer from our sales department.
Can I extend my licence or pause and resume the licence duration?
The licence duration is non-extendable and customers are not permitted to pause and resume the license duration once it has been activated.
What happens if I do not use all my predictions within the licence duration?
It’s recommended to utilize your predictions within the specified validity period to fully benefit from your license. Any unused predictions remaining at the time of termination (licence expiration date) will be permanently lost and cannot be carried over, cannot be refunded and cannot be redeemed thereafter.
Can I get a refund for my licence purchase?
Refunds are available within 14 days of payment if no predictions were used during this 14-days period.
Can I tranfer my licence instead of requesting a refund?
Licenses are non-transferable. If you require assistance with your license or have specific questions, please contact our support team for guidance.
What happens when my licence expires?
After the termination of your license, you will have access to your data in a “view-only” mode for a period of 1 month only. If you wish to retain access to your data beyond this period (referred to as data storage, which expressly excludes making predictions), you must request an offer for this specific data storage only service before your license expires.
Can I store my data? For how long? Is it safe?
We offer free 1 month storage to all our clients. For our long term clients, we provide an additional data storage (under payment) of safe, secure, and fully encrypted data storage facilities. During this time, you can share, view, and download your data unlimited!
Can I share my data with my colleagues?
Absolutely, yes! We believe sharing is caring. That is why we allow unlimited access to your data that can be downloaded and printed as per your convenience.
I purchased a licence with “100 predictions allowed within 30 days validity” can you explain exactly what does it mean?
This phrase indicates that upon purchasing the license, you are entitled to use our prediction services up to a maximum of 100 predictions within a period of 30 days from the licence delivery date.
Is there a limit on how many predictions I can make per day?
The license allows you to make a total of [number of predictions allowed] predictions within the [validity period] (e.g., 100 predictions within a 30-day validity period). There is no daily limit specified, so you can use the predictions at your discretion within the allowed timeframe.
How are predictions counted within this license?
Each time you make a prediction using our service, it will be deducted from the total allocation of predictions allowed under your license for the specified validity period.
Where can I find more information about my specific license terms and conditions?
For detailed information about your license terms and conditions, please refer to our offer, purchase order confirmation and our Seetalabs Customer Agreement
What is a CSV template?
A CSV (Comma-Separated Values) template is a file format used to store tabular data, where each line represents a row of data and values are separated by commas (or other delimiters).
CSV files can be opened and edited using common spreadsheet applications such as Microsoft Excel, Google Sheets, Apple Numbers, and others. Simply import the CSV file into the software to view and modify its contents.
Do you have a template to upload tabular data into your system?
Yes, you can download our template file following these steps:
1- Log-in your account
2- Go to “Make prediction” menu
3- Go to “Multi – Upload File” submenu
4- Search “Upload help” on top of the page and “Click to open” link
5- Download the CSV template and follow the wizard
Finally, open the file in a spreadsheet program like Microsoft Excel or Google Sheets, without modifying the first row containing the column header. Fill in the required data under each column, and save the file in CSV format. You are now ready to upload in our system the CSV file.