What are Dynamic Line Ratings?
Dynamic Line Rating (DLR) is a method for determining how much power a line can safely carry depending on forecasted weather conditions. DLR accounts for all atmospheric cooling and heating factors: wind speed, wind direction, ambient temperature, and solar radiation.
Gridraven’s DLR is unique because it is fully sensorless. Gridraven uses machine learning to make ultra-precise weather forecasts in every span that account for hyper-local features such as wind sheltering.
Hyper-Local Wind Cooling: The Core Advantage
The most significant factor in a line’s capacity is the cooling effect of the wind. Even a modest wind speed of 2 m/s can cool a conductor so effectively that it allows for twice the energy transmission compared to standard static assumptions.
- Probabilistic Forecasting: Gridraven determines confidence intervals for weather inputs. Grid operators require at least 98% certainty in line ratings.
- 1-Meter Resolution: Gridraven’s DLR is powered by a proprietary machine learning model that predicts wind conditions on the basis of 1-meter landscape features.
- Landscape-Awareness: Our model accounts for the surrounding landscape—such as hills, valleys, and forests—which can shelter conductors and reduce cooling effects by over 50% compared to open weather stations.
- Span-Level Granularity: Ratings are calculated for every individual span (the segment between two towers). This ensures that "hotspots" or critical segments with low wind are never overlooked.
This visualization clearly shows a sheltered area (forest) where the wind speed is significantly reduced. Consequently, the spans in this segment exhibit lower cooling and a reduced ampacity (indicated by the red line), whereas the segments over open terrain benefit from higher wind speeds and substantially greater capacity (indicated by the green line). Identifying the specific "weakest link" at any given moment defines the safe operational limit for the entire circuit.
Probabilistic Forecasting and Risk Management
Energy markets and grid operations take decisions in advance. Gridraven Claw provides forecasts for up to 10 days ahead with hourly resolution.
To ensure reliability, we employ a probabilistic approach:
- Gaussian Mixture Models (GMM): Instead of a single "best guess," our model produces a probability distribution of possible outcomes.
- Confidence Intervals: We provide ratings at various confidence levels—50%, 75%, 90%, 95%, 98%, and 99%.
- 98th Percentile Default: By default, Gridraven operates at the 98th percentile, providing grid operators with an extremely high level of statistical certainty for safe operation
This graph demonstrates how predicted DLR (even at high confidence levels) consistently stays above traditional Ambient Adjusted Ratings (AAR), significantly increasing available capacity without compromising safety
Safety and Reliability
Safety is the cornerstone of Gridraven’s DLR implementation. The system aggregates the thermal capacity of every individual span and automatically selects the lowest value as the line's overall rating. This "weakest link" logic ensures that no single segment of the line ever reaches its maximum allowable temperature (e.g., 95°C), preventing excessive sagging and potential ground clearance violations.
By reducing wind speed prediction errors by over 50% compared to raw numerical weather predictions, Gridraven provides a stable, software-only solution that is as reliable as hardware sensors but far more scalable.