Automating Substation Inspection: How Robotics Enhance Reliability and Reduce Costs

1.Project Background and R&D Necessity​

With technological advancements and the deepening of power system reforms, the automation level of power systems has significantly improved. Substations are evolving towards "unattended" or "fewer personnel on duty" operational models. Currently, substations primarily rely on the "Four Telemetry" functions (Telemetry, Telesignaling, Telecontrol, Teleregulation) and SCADA systems to monitor equipment electrical signals. However, this traditional approach cannot achieve real-time perception and awareness of the equipment's on-site physical status (such as appearance, temperature, abnormal sounds, etc.).

The current operation and maintenance model has obvious shortcomings: when an abnormality occurs in a substation, dispatchers must first notify remote substation operation teams to travel to the site, and then organize repairs. This process significantly delays defect elimination time, impacting power supply reliability and service quality. Furthermore, traditional remote video monitoring only realizes the digital transmission of audio and video, lacks intelligent analysis capabilities, and is limited by the fixed field of view of single cameras and limited network bandwidth, making it difficult to deploy on a large scale.

​2. Overall Robot System Structure​

This system adopts a two-layer "Base Station-Mobile Agent" architecture to achieve coordinated remote monitoring and on-site inspection operations.

​2.1 Base Station System​

The base station system is deployed at the remote monitoring center and serves as the human-machine interaction and command core of the entire system.

​2.2 Mobile Agent System (Robot Body)​​

The mobile agent is an intelligent terminal that performs on-site inspection tasks, possessing a high degree of autonomy and environmental adaptability.

• ​Mobile Chassis Design:​​ Utilizes a four-wheel differential drive structure. The two front wheels are independently driven wheels, each powered by a separate motor, enabling flexible differential steering; the two rear wheels are caster wheels. This structure offers advantages such as good straight-line motion stability, small turning radius (can pivot around the center point of the front wheels), strong road adaptability, no sideslip, and a simple, reliable structure.
• ​Motion Control Subsystem:​​ The hardware core is a PC104 mainboard, equipped with a PCL-839 motion control card and motor drivers. This subsystem is responsible for all motion behaviors of the robot. By receiving commands from the upper-level planner and integrating the vehicle dynamics model, it accurately decomposes velocity commands to each drive motor, achieving smooth and precise motion control.
• ​Task Execution Subsystem:​​ Serves as the robot's "senses" and "hands". Core functions include:
• ​Data Acquisition:​​ Integrates a visible-light CCD camera, an infrared thermal imager, and a high-performance directional microphone (MEMS) for collecting image (visible and infrared) and sound data from power equipment.
• ​Automatic Charging:​​ Capable of automatically returning to the charging dock for对接 and charging, ensuring 7x24 hours uninterrupted operation.

​3. Core Technologies and Functional Implementation​

​3.1 Intelligent Real-time Path Planning Technology​

• ​Global Path Planning:​​ Based on a pre-set substation electronic map, calculates the optimal sequence of equipment stopping points to visit during an inspection task and feasible paths according to strategies like "shortest path," "fewest turns," or "comprehensive optimum."
• ​Local Path Planning:​​
• ​Obstacle Avoidance:​​ Employs the VFF (Virtual Force Field Histogram) algorithm, combined with sensor data like LiDAR, to generate real-time avoidance commands, ensuring safe navigation in dynamic environments.
• ​Line Tracking:​​ Uses the classic PID control algorithm to ensure the robot follows predetermined routes accurately.
• ​Environmental Adaptation:​​ Applies the EM algorithm and clustering algorithms to process sensor data, effectively fitting road boundaries and overcoming positioning deviations.

​3.2 Multi-Modal Equipment Detection and Diagnosis System​

1. ​Remote Infrared Monitoring and Diagnosis System​
• ​Configuration:​​ Online infrared thermal imager, includes image acquisition, processing, display, storage, and report generation modules.
• ​Functions:​​ Automatically detects equipment surface temperature, compares it with preset thresholds, and triggers audible/visual alarms immediately upon detecting abnormalities; can generate equipment temperature gradient maps, temperature-time curves, etc., to assist in fault analysis; uses image compression technology to support simultaneous monitoring of real-time infrared feeds from multiple substations at the dispatch center.
2. ​Remote Image Monitoring and Diagnosis System​
• ​Configuration:​​ Visible-light CCD camera and video server.
• ​Functions:​​ The base station system performs intelligent analysis (e.g., difference image analysis, correlation analysis) on the returned visible-light images to automatically identify the appearance status of power equipment and instrument readings. Normally, it automatically switches monitoring points; it only stores images and triggers alarms when anomalies are detected, significantly improving channel utilization and monitoring effectiveness.
3. ​Remote Sound Monitoring and Diagnosis System​
• ​Configuration:​​ High-performance directional MEMS microphone.
• ​Functions:​​ Collects equipment operating noise in real-time, compresses it, and transmits it back. The system intelligently assesses the operating status and anomaly types (e.g., loosening, discharge) of equipment like transformers by comparing real-time noise with historical normal data, and provides an interactive interface for maintenance personnel to query and analyze.
4. ​Moving Object Intrusion Detection and Alarm System​
• ​Principle:​​ Based on video stream moving target detection algorithms, automatically identifies and extracts areas in the video containing objects moving relative to the background.
• ​Functions:​​ Once an abnormal moving target, such as illegal intrusion, is detected, the system immediately triggers an alarm and saves on-site images, providing evidence for security追溯, enabling true unattended security monitoring.

​4. Field Operation and Application Results​

​Core Application Value:​​ This robot system innovatively integrates "non-contact mobile detection" with the existing "contact-based fixed monitoring" in substations, forming a comprehensive monitoring system that covers both space and status, effectively compensating for the shortcomings of traditional inspection models.

​Operational Results:​​

• ​Significantly Enhanced Safety and Reliability:​​ Capable of promptly detecting potential faults such as thermal defects, surface foreign objects, oil leaks, and sound abnormalities in equipment, eliminating accidents in their infancy.
• ​Improved Operation and Maintenance Efficiency:​​ Replaces manual repetitive and tedious routine inspections, and provides dispatchers with real-time and accurate feedback on site conditions, offering crucial data support for emergency decision-making, substantially reducing fault handling time.
• ​Reduced Operational Costs:​​ Serves as key technological equipment for realizing the "unattended" substation model, helping power utilities optimize human resource allocation and reduce long-term operational costs.