PV Safety Detection With User Feedback to Cut False Alarms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing PV power station management methods fail to effectively address false alarms and lack real-time updating, leading to decreased detection accuracy due to inadequate user interaction and non-iterative equipment management.
Innovation Solution
A PV equipment safety detection method and system that performs status detection, generates natural language-based inquiry instructions, interacts with users to resolve abnormalities, and performs model switching or self-learning for iterative upgrading, using a large language model (LLM) to enhance detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If automated detection is implemented without user interaction, then detection speed is improved, but detection accuracy deteriorates due to false alarms
Solution Approach 1:
The system implements a feedback mechanism where detection results are sent to users for confirmation. When an abnormality is detected, the system queries the user to verify whether the detected issue truly exists, creating a closed-loop feedback system that improves detection accuracy while maintaining automated operation
Solution Approach 2:
The system enables users to self-verify detection results through simple interactions. Users can confirm or deny detected abnormalities through straightforward queries, allowing the system to leverage user knowledge for verification without requiring complex manual inspection procedures
2Measurement precision
If complex troubleshooting procedures are implemented, then abnormality resolution accuracy is improved, but operation complexity increases
Solution Approach 1:
The troubleshooting process is segmented into multiple simple interaction steps rather than presenting users with complex procedures all at once. The system divides abnormality resolution into sequential queries and actions, making each step manageable and reducing perceived complexity
Solution Approach 2:
The troubleshooting procedure dynamically adapts based on user responses. The system adjusts subsequent queries and actions according to the specific abnormality confirmed by the user, creating a flexible rather than rigid process that simplifies operation while maintaining accuracy
3Device complexity
If static detection models are used, then system simplicity is maintained, but detection accuracy deteriorates over time
Solution Approach 1:
The system collects detection data and user feedback in advance to build training datasets. This preliminary data accumulation enables subsequent model optimization without requiring immediate system redesign, allowing accuracy improvement while maintaining operational simplicity
Solution Approach 2:
The detection model parameters are optimized based on collected data, transforming the static model into an adaptable one. By changing model parameters rather than restructuring the entire system, the patent achieves accuracy improvement while preserving system simplicity
Data Source
Figure 1~2
Figure 3~4
AI summary
The present disclosure provides a photovoltaic (PV) equipment safety detection method and a PV equipment safety detection system. The method includes: performing status detection on a PV power station; determining whether there is an abnormality according to detection information, outputting an abnormality instruction if there is an abnormality, and parsing and processing the abnormality instruction to generate an inquiry instruction; displaying the inquiry instruction to a user, and providing, by the user, reply information; determining, according to the reply information, whether to send a further inquiry instruction to the user for an inquiry, or to send an execution instruction to the PV power station, such that according to the execution instruction, the PV power station performs an action to resolve the abnormality or conducts self-inspection and reports the abnormality; and feeding the execution instruction back to a preprocessing port, and performing model switching or model self-learning according to the execution instruction, thereby performing iterative upgrading to obtain an accurate detection result. Through effective iterative interactions with the user, a false alarm caused by the environmental condition is avoided, thereby improving the abnormality handling efficiency, saving human and material resources, reducing operation and maintenance costs, and effectively enhancing user experience.