Kavaken empowers renewable energy asset owners with artificial intelligence to overcome the challenges they are facing
Renewable energy asset owners face decreasing revenues, aging assets, and increased downtime. At the same time, their assets generate a significant amount of data that can help them overcome these challenges. However, they lack the capabilities to exploit these data efficiently.
Kavaken is an Artificial Intelligence-powered solution that increases revenue by optimizing asset management with a data-centric approach. In this solution, production forecasting, asset management, and predictive maintenance modules work in unison to drive operational decisions with a focus on revenue optimization.
The capable modules of Kavaken, which brings Artificial Intelligence to renewable energy asset management, are:
- Revenue Cockpit: Use the powerful dashboard to watch critical metrics such as “Lost” Revenue, Revenue-Based Availability, Energy-Based Availability, Capacity Factor
- Forecast+: Use advanced analytics to better forecast short-term production, reduce balancing costs and increase profits.
- Power Curve Analyzer: Analyze turbine power curves, maximize turbine output and revenue by recommendations for shifting the power curve up.
- Maintenance Planner: Plan and track all maintenance activities in one place; see real-time progress and avoid potential delays via advance notifications.
- Operations Tracker: Achieve peace of mind from miles away by knowing the turbines are operating within normal behavior limits and costly errors are eliminated.
- Predictive Maintenance: Reduce unplanned maintenance and downtime by harnessing large amounts of data and alerts produced ahead of time to prevent breakdown and loss of revenue.
1. Revenue Cockpit
Revenue Cockpit provides a live dashboard to view the wind power plant critical information and perform reporting and analysis. Main features of the Revenue Cockpit:
- A general overview of all wind power plants on a live map
- A live map with information such as wind speed, production, temperature, and a summary of last days production and loss revenue
- A summary table reflecting the summary of KPIs of all sites focused on key revenue metrics such as Production, Capacity Factor, Revenue-Based Availability, Energy-Based Availability, Time-Based Availability, Power Efficiency of the turbines and sites.
- The site page will represent a live map of the site representing each turbine.
- Each turbine will show live information such as wind speed, production, temperature.
Kavaken’s Forecast+ module uses an AI-based approach to help wind energy clients with forecasts that minimize uncertainty. Wind energy predictions can negatively impact grid operations and profitability for intraday markets.
Kavaken’s unique approach benefits from combining the advantages of its machine learning model with additional input from commercially available forecast models. Each model has its own strengths for specific scenarios, and by combining all, Kavaken can provide higher accuracy prediction to end-users. Kavaken provides the client with Day Ahead (GÖP) and Intraday (GİP) production forecasts.
AI-powered Forecast module’s advantages start with the power of historical data available for training machine learning models. There are three key steps to Kavaken’s proprietary model. The first layer is getting inputs from Plant, SCADA, and Weather along with physical weather models. The second stage is the optimization of Weather and physical models using machine learning. The final stage combines ML-based meteorological results with turbine-specific information. After all three steps, the final deliverable provides power forecasts for the farm. For Kavaken users, the model also takes other forecasts as input for enhanced accuracy.
The client will have access to the dedicated online portal of Kavaken and will have access to Kavaken Forecast and the comparison of historical model accuracy of each forecast model.
3. Power Curve Analyzer
Kavaken’s Power Curve Analyzer model uses AI and machine learning to help wind energy clients analyze their power curves and provides a clean power curve model that would enable the detection of any unoptimized area in the power curve model of each turbine.
Kavaken’s unique approach benefits from using data science tools and applications of AI, which outlines any unoptimized points in the power curve. It identifies the cause, improves the power curve, and increases turbine outputs. The Power Curve Analyzer presents customers with a baseline power curve that acts as a benchmark for turbine performance enabling detection in shifts in the power curve, non-performance analysis (icing, curtailment, hysteresis, Etc.), and yaw performance analysis.
4. Maintenance Planner
Kavaken’s Maintenance Planner is a maintenance management tool where high-level plans are entered by the user and then visualized for tracking which maintenance activities are on track, including commenting and documentation attaching capabilities. After the integration, the product is ready for use. No training data is required, and the user only needs to enter maintenance plans and update the system regularly.
Screens show on target, delayed, planned maintenance activities, and it is a one-stop portal for accessing comments and documents associated with maintenance activities.
5. Operations Tracker
Kavaken’s Operations Tracker uses an AI-based approach to create a standard behavior model for turbine parameters and enables the user to identify anomalies. The classical approach to maintenance relies on the service teams to identify faults following the defect; the modern approach integrates predictive maintenance that allows avoiding any fault before occurrence. Kavaken Operations Tracker works 24/7 reviewing the behavior of all critical signals and alerting the user of an anomaly.
6. Predictive Maintenance
Kavaken’s Predictive Maintenance generates a warning (flag) if it predicts a major component (i.e., generator, gearbox) may have a breakdown in the near future by using ML-based techniques (primarily anomaly detection).
Predictive maintenance systems, in general, maximize wind turbines’ uptime by early detection of potential failures of various turbine components with the aim of reducing unplanned maintenance and downtime by harnessing large amounts of data to produce alerts ahead of time to prevent the breakdown of costly components such as gearboxes, main bearings and generators and loss of revenue. By analyzing both high-frequency vibration and SCADA data, the product utilizes a wide range of statistical analyses and machine learning techniques such as normal behavior model development, clustering, and anomaly detection methods to predict potential failures.
While the initial focus is on wind energy, the capabilities are easily transferable to other renewable energy sources, making it a solution available to an $80B global market. With the ever-increasing importance of renewable energy for our planet, we believe our solution can play an essential role in making the transition smoother by using Artificial Intelligence.