HVDC Return Element Fault Detection via Electromagnetic Simulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Fault detection in return elements of high voltage direct current (HVDC) power transmission systems is difficult due to the absence of voltage and current flow in normal operation, making it challenging to assess their health state.
Innovation Solution
A method involving an electromagnetic transient simulation using initial and altered electromagnetic models, machine learning, and input pulses to detect and locate defects in return elements, utilizing a sensing system and computing device to generate a trained model for state determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional sensing systems are used in HVDC power transmission systems, then the system operates normally without voltage and current in return elements, but fault detection becomes difficult due to absence of voltage and current flow
Solution Approach 1:
The patent applies preliminary action by performing electromagnetic transient simulations before actual fault detection. Multiple altered electromagnetic models are created in advance representing various fault conditions, and a trained model is generated beforehand to enable subsequent fault detection without requiring voltage or current flow in the return element during normal operation.
Solution Approach 2:
The patent introduces an intermediary approach by using electromagnetic transient simulations as a mediator between the sensing system and the return element. The simulations create virtual fault scenarios that allow fault detection capability to be developed and stored in a trained model, which then serves as an intermediary tool for actual fault detection without requiring direct voltage or current measurement in the return element.
2Measurement precision
If electromagnetic transient simulations with multiple altered electromagnetic models are performed, then fault detection precision is improved, but computational complexity and processing time increase
Solution Approach 1:
The computationally intensive electromagnetic transient simulations and model training are performed in advance during a preliminary phase. The trained model generated from multiple altered electromagnetic models is stored and reused for actual fault detection, avoiding repeated complex computations during operational fault detection while maintaining high precision.
Solution Approach 2:
The patent creates multiple copies of electromagnetic models with different fault conditions (altered electromagnetic models) to train a comprehensive fault detection model. This copying approach allows the system to learn from various fault scenarios without requiring complex real-time computations during actual fault detection, as the trained model encapsulates the knowledge from all simulated scenarios.
3Productivity
If manual intervention is used for fault detection in return elements, then system complexity is reduced, but troubleshooting efficiency and productivity decrease
Solution Approach 1:
The patent implements self-service by enabling the sensing system to automatically detect and locate faults in the return element without requiring manual intervention. The trained model processes sensing data autonomously to identify fault conditions and determine fault locations, significantly improving troubleshooting efficiency while maintaining system simplicity through automated decision-making.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables efficient and precise fault detection and localization in return elements, reducing reliance on manual intervention and improving troubleshooting efficiency.
Implementation Method 1
The sensing system is configured to provide measurement information of the return element. The measurement information can comprise voltage information and/or current information.
Implementation Method 2
The initial electromagnetic model is exemplarily used for performing an electromagnetic transient, EMT, simulation. The transients refer to short-duration events such as switching operations, faults, and/or sudden changes in the electrical network of the electrical components.
Implementation Method 3
The analyzing computing device is configured to provide a trained model, wherein the trained model is generated dependent on a plurality of responses to the provision of the plurality of predetermined input pulses. The state of the return element is detected dependent on some of the predetermined input pulses applied to the return element and dependent on the trained model.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method for detecting a state of a return element (1) connected to a sensing system (2) of a high voltage direct current, HVDC, power transmission system (3) is provided, comprising: - providing an initial electromagnetic model characteristic of the return element (1) and the sensing system (2), wherein the return element (1) has no defective parts, - providing at least one altered electromagnetic model based on the initial electromagnetic model, wherein the return element (1) has at least one defective part, - providing a plurality of predetermined input pulses (4), which are provided to the initial electromagnetic model and the altered electromagnetic model, wherein some of the predetermined input pulses (4) are different from one another, - providing a trained model of the return element (1) and the sensing system (2), wherein the trained model is generated dependent on a plurality of responses to the provision of the plurality of predetermined input pulses (4), - detecting the state of the return element (1) dependent on some of the predetermined input pulses (4) applied to the return element (1) and dependent on the trained model. Additionally, a system (14), a computer program product and a computer-readable storage medium are provided.