Power Distribution Interruption Prediction Using Weather Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods fail to accurately predict power distribution system interruptions caused by common weather conditions, such as rain, wind, temperature, and humidity, which are overlooked in reliability analysis, and lack methods to assess interruption risk based on immediate weather conditions.
Innovation Solution
A method using daily and hourly weather data to predict the number of interruptions in a given region by modeling the combined effects of various weather conditions, including rain, wind, temperature, lightning, humidity, barometric pressure, snow, and ice, to assess interruption risk and improve reliability assessments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing reliability analysis methods are used, then catastrophic weather events are excluded from reliability indices, but common weather conditions are overlooked and fail to predict total interruptions accurately
Solution Approach 1:
The patent segments weather conditions into two categories: catastrophic weather events (excluded from traditional reliability indices) and common weather conditions (included in the new model). This segmentation allows the system to separately handle different types of weather impacts, with the new stochastic model specifically addressing common weather conditions like rain, wind, temperature, and humidity that were previously overlooked.
Solution Approach 2:
The patent merges multiple common weather conditions (rain, wind, temperature, humidity, barometric pressure) into a single integrated stochastic model. This combination allows the system to predict total interruptions by considering the combined effects of various weather parameters simultaneously, rather than analyzing them separately or excluding them entirely.
2Measurement precision
If methods predict daily or by shift interruptions based on combined weather effects, then interruption risk assessment improves, but no such methods currently exist for common weather conditions
Solution Approach 1:
The patent incorporates feedback mechanisms by using actual interruption data and weather data to continuously refine and validate the stochastic model. The system compares predicted interruptions with actual occurrences, allowing for model calibration and improvement over time, which enhances prediction accuracy while managing complexity through iterative optimization.
Solution Approach 2:
The patent changes the parameters of the prediction model by introducing stochastic variables that represent the probabilistic nature of common weather conditions. Instead of using fixed thresholds or deterministic models, the system employs probability distributions and statistical parameters to capture the variability and combined effects of multiple weather factors, enabling more accurate predictions despite increased model complexity.
3Reliability
If multiple common weather conditions are modeled simultaneously, then total interruption prediction improves, but the range of combinations is great and application complexity increases
Solution Approach 1:
The patent creates a universal stochastic model that can handle multiple common weather conditions and their various combinations through a single integrated framework. The model is designed to be multi-functional, accommodating different weather parameters (rain, wind, temperature, humidity, barometric pressure) and their interactions, while maintaining consistent methodology across diverse weather scenarios and geographic regions.
Data Source
AI summary
A method for predicting electrical power distribution interruptions based on common, immediate weather conditions. Daily, hourly, and bi-hourly weather data are used to predict the number of interruptions. Common weather conditions include, but are not limited to, rain, wind, temperature, lightning, humidity, barometric pressure, snow, and ice. The method includes compiling common weather data including a plurality of weather variables and the number of historical interruptions for a historical period, establishing model equations for the average value of the weather variables, combining the model equations for each of the weather variables into a composite model, and performing a regression analysis using the composite model to establish interruption prediction values. A computer program product for enabling said method and a computer system adapted to carry out said method are also included.