Early Fault Detection in Wind Turbines

Multi-dimensional composite sensing technology and monitoring system for wind power to improve wind turbine reliability and reduce maintenance costs. The system uses a multi-sensor setup on the turbine body to monitor various components like bearings, gears, foundations, bolts, and oil. Sensors like accelerometers, inclinometers, gyroscopes, ultrasonic probes, and oil sensors provide comprehensive condition monitoring. Data is transmitted to a monitoring station and server for analysis and alarms. It allows early fault detection, prognosis, and optimization of wind turbine maintenance.

Source

Intelligent system for predicting wind turbine blade failures using deep learning to analyze blade temperatures, pressures, and images. The system extracts features from blade temperature and pressure time series, blade images, and correlates them to detect blade faults. It uses a neural network to fuse the features and a classifier to determine if a blade fault warning should be issued. The system trains the neural network with supplementary loss based on probability density consistency between features to improve feature fusion quality.

Source

Deep learning-based predictive maintenance system for wind turbines to improve reliability and reduce downtime compared to manual maintenance. The system uses sensors to monitor turbine parameters, processes the data, analyzes it with deep learning models to detect faults, and generates control signals to address issues before failures occur. The system sends the data to a central monitoring unit for analysis and decision making. It aims to provide more accurate, efficient, and automated maintenance compared to manual recording and inspection.

Source

Comprehensive online monitoring system for wind turbines to detect and diagnose issues with various components. The system uses a master station and four slave stations connected in sequence. The stations have sensors for blade load, main shaft torque, icing, atmospheric pressure, bearings, vibrations, currents, temperatures, and humidity at locations like the tower bottom, top, cabin, gearbox, generator, and motor.

Source

Wind farm monitoring system interaction method that improves fault diagnosis accuracy and efficiency to reduce maintenance costs. The method involves monitoring wind turbine component data, analyzing it using a basic diagnosis system to generate initial fault results, then allowing authorized personnel to customize the diagnosis criteria and repeat the analysis to get final results. Alarms are then provided based on multiple results and diagnostic reports are generated. This iterative customization improves fault diagnosis accuracy compared to fixed criteria.

Source

A real-time monitoring system for wind turbine tower tilt and sink based on multiple sensors that provides enhanced tower monitoring capabilities and alerts for wind turbines. The system uses sensors inside the wind turbine to detect tower tilt, settlement, internal environmental conditions, etc. The sensor data is transmitted wirelessly and processed by a central unit. Alarms are triggered if thresholds are exceeded and notifications sent via lights, sounds, SMS, calls, and web/app alerts. This multi-modal alerting improves response time and reliability compared to single methods. The cloud platform allows remote monitoring and analysis.

Source

Wind turbine fault diagnosis system that can efficiently monitor and diagnose faults in wind turbines with mixed equipment from different manufacturers. The system uses a modular design with a data acquisition module, signal acquisition card, signal processing module, and display module. The signal acquisition card can adapt to different data communication methods and formats of the wind turbine equipment. The signal processing module extracts fault features from the acquired signals. The display module presents the fault diagnosis results. This allows the system to monitor and diagnose faults across diverse wind turbine equipment types with mixed communication protocols.

Source

Wind turbine unit condition monitoring and fault diagnosis method using PHM technology that provides improved monitoring and diagnosis compared to traditional methods. The method involves monitoring wind turbine system health by fusing vibration and binary data using Copula functions. It also uses denoising and feature extraction techniques followed by SVM modeling to diagnose faults. This allows monitoring system-wide signals and binary variables, diagnosing both temporary and early faults, and improving fault detection rate and warning time.

Source

A wind turbine tower condition monitoring system that improves accuracy, reliability, automation, and safety compared to existing tower monitoring methods. The system uses sensors on the tower bolts and inclination angles to detect issues like loosening or leaning. A data acquisition module collects the sensor data, a router transmits it, and a data processing module analyzes it to detect problems. This allows real-time, automated monitoring and early warning of tower conditions.

Source

Lean management and control method for power grids using big data to improve accuracy and reliability of wind turbine control in wind farms. The method involves generating a sensor relationship diagram for the wind turbines and sensors, and leveraging big data techniques to analyze and model the sensor relationships. This allows predicting sensor faults and compensating for faulty sensors using data from healthy sensors. The steps include: 1) creating a sensor relationship diagram with units representing wind turbines and edges representing sensor connections; 2) generating initial features for each unit (wind turbine or sensor) using operating parameters; 3) training a relationship generation model to predict relationships between unit features based on the diagram; 4) using the model to predict faulty sensor behavior based on data from healthy sensors.

Source

A remote online monitoring system for wind power gearboxes that provides more accurate and reliable condition monitoring compared to vibration and temperature sensors alone. The system uses additional sensors like strain gauges at the gearbox input, pressure sensors at the oil inlet and outlet, and connects them to a computer via controllers. This allows monitoring of factors like torque, bending moment, oil pressure, in addition to vibration and temperature. The computer collects the sensor data and sends it over the network for analysis and alerts if needed. The system architecture includes firewall, router, and receiving terminal to securely transmit the monitored data offsite.

Source

A remote fault detection system for offshore wind turbines that allows monitoring and analysis of the turbine components to predict faults and improve maintenance. The system uses sensors to monitor blades, transmission shafts, and generators, transmits the data to a control center, performs calculations and analysis, and uses neural networks to predict faults based on operating conditions. This allows remote fault detection and diagnosis for offshore turbines where maintenance is difficult.

Source

Wind turbine fault monitoring system using current analysis to detect mechanical and electrical faults with higher accuracy compared to vibration-based methods. The system involves placing harmonic sensors on the main power cables of the wind turbine stator to measure current harmonics. By analyzing the harmonic content, it can diagnose deterioration and failures in components like generators, transformers, and cables. Harmonic energy is related to component frequencies and degradation levels. Increased harmonic content indicates energy loss, component failure signs, or overall equipment deterioration.

Source

A monitoring alarm system for wind turbines in power plants that provides real-time fault detection and quick fault localization to improve turbine reliability and uptime. The system monitors temperature and amplitude signals from critical points on the turbine. It processes the signals to extract diagnostic parameters that are insensitive to operating conditions but sensitive to faults. Threshold alarms are set for these parameters. If a parameter exceeds its threshold, the system alerts the operator and quickly traces the fault location to enable faster repairs. This allows early detection and isolation of faults before they propagate, reducing downtime and maintenance costs.

Source

Wind turbine condition monitoring system that uses wireless sensors to provide real-time monitoring of wind turbine components without the need for wired connections. The system has multiple wireless sensors placed at various locations in the turbine. These sensors collect data which is wirelessly transmitted to a central control module. The control module processes the data and provides feedback on any issues. This allows proactive maintenance and fault prediction by monitoring multiple components wirelessly instead of relying on wired harnesses.

Source

Monitoring and managing wind power generators using multi-sensor data to improve fault detection and diagnosis. The system collects sensor data from various systems inside the wind turbine like the wind measurement, pitch control, generator cooling, etc. It identifies abnormal fluctuations in the sensor data over time. By analyzing the characteristics of abnormal fluctuations, it can discover hidden faults and correlations between sensors. This allows identifying and diagnosing multi-component failures that may not be immediately apparent from individual sensor data.

Source

Wind turbine fault diagnosis and health monitoring system that uses sensors on the blades to comprehensively gather blade operating condition parameters. It analyzes real-time vibration, speed, and temperature data to identify wind turbine faults and monitor health. Preprocessing techniques like wavelet denoising and outlier removal are applied to the signals. Neural networks diagnose vibration faults. Anomaly detection determines critical early warning states compared to historical data.

Source

Fault protection system for wind turbine power generation systems that can detect and respond to faults in permanent magnet generators. The system monitors parameters like generator speed, converter status, and circuit breaker operation to determine the wind turbine operating mode. It then sets expected values for parameters like voltage and speed based on the mode. If actual values deviate, it identifies and responds to fault conditions. This allows more accurate and reliable fault detection in permanent magnet generators compared to just monitoring current thresholds.

Source

A fault diagnosis system for wind turbine drivetrains using a network of sensors in key components to provide more accurate and comprehensive fault detection compared to traditional methods. The system uses a mesh networking strategy with sensors in components like gearboxes, shafts, bearings, and lubrication systems that transmit data wirelessly. It collects parameters like vibration, temperature, lubrication quality, and audio. The sensor data is transmitted using 5G frequencies for stability. The collected data is analyzed to diagnose faults and prioritize maintenance. The system provides more detailed and accurate fault detection compared to just monitoring vibration or a few parameters.

Source

Fault diagnosis device and method for wind turbine transmission systems that enables proactive maintenance and fault mitigation by monitoring and analyzing equipment conditions using sensors, wireless communication, and machine learning models. The device has a monitoring system with sensors for parameters like lubricating oil temperature, quality, and vibration, as well as audio monitoring. It wirelessly transmits the data to a central server using a mesh network. The diagnosis system extracts features from the sensor data and uses a causal inference model to predict faults and their causes. This provides specific fault mitigation actions and allows confirming the failure root cause by observing sensor changes.

Source

Intelligent monitoring system for large wind power components that uses electrical quantity analysis and diagnosis to predict and diagnose faults in wind turbine components like converters, generators, and blades. The system monitors wind turbine operation and environment, compares data, calculates component failure rates, and predicts faults in real-time. This allows remote analysis and diagnosis of complex faults before they escalate, reducing downtime and maintenance costs compared to waiting for failures.

Source

Wind turbine full state monitoring system using multi-source sensing to improve reliability and reduce maintenance costs. The system uses sensors like audio, vibration, video, and load sensors on the blades, tower, and nacelle to comprehensively monitor the wind turbine's core components. This allows early fault detection, predictive maintenance, and avoiding serious accidents compared to just structural sensors. The sensors are connected via wired/wireless hybrid communication using optical fiber. The integrated data acquisition device preprocesses the multi-type, multi-signal, multi-parameter data.

Source

Wind turbine rolling bearing fault monitoring system that enables remote and real-time detection of bearing failures in wind turbines. The system uses sensors to monitor vibration, oil quality, temperature, and pressure in the bearings. Data from these sensors is collected and transmitted via an optical fiber ring network to a central server. The server analyzes the data to detect bearing faults and sends alerts to the remote monitoring terminal. This allows proactive identification and prevention of bearing issues before they cause turbine downtime.

Source

A wind turbine online monitoring system that uses multiple sensor signals to provide better real-time monitoring and fault diagnosis of wind turbines compared to relying on just vibration or speed signals. The system has data acquisition, visualization, and transmission layers. Sensors like accelerometers, speed sensors, noise sensors, temperature sensors, humidity sensors, cameras, infrared sensors, and air sensors are used to collect comprehensive data from the turbine. The acquisition layer conditions, fuses, filters, and converts the signals. The visualization layer displays the turbine status and provides manual control assistance. The transmission layer sends the data to a central system for storage, analysis, and alarms.

Source

A wind turbine failure prediction system that uses sensor data from the turbine to forecast mechanical failures. The system involves installing sensor units on the turbine components, collecting the sensor data through a hub, analyzing it on a server, and using machine learning algorithms to predict component failures. The sensors are calibrated using a specific algorithm to compensate for variations and drift. The calibration involves calculating slopes and intercepts from sensor values to determine constants for updating the calibration function.

Source

Wind turbine condition monitoring and fault diagnosis system that uses multiple sensors to monitor vibration, partial discharge, temperature, and rotational speed of wind turbines. The sensors are located on the turbine and transmit data to a control cabinet. The cabinet communicates with a remote fault diagnosis server. This allows comprehensive, real-time monitoring of turbine operating conditions and diagnoses faults remotely. It improves wind turbine maintenance by providing detailed fault information to technicians.

Source

Real-time monitoring and fault prediction of wind turbine gearboxes using multi-sensor signal fusion. The method involves collecting signals from multiple sensors like acoustic emission, noise, vibration, oil temperature, and displacement. These signals are analyzed using a graph model that extracts principal components and calculates distance intervals between them. The intervals are used to judge the gearbox state for real-time monitoring and fault prediction.

Source

A fault trend diagnosis system for wind farm groups that can monitor and analyze the operating state of wind turbines to detect faults and predict failures. The system uses dual-axis magnetoresistive sensors placed on the top and bottom of the main shaft of each wind turbine to collect real-time operating state data. This data is compared to thresholds and analyzed by a central computer to diagnose faults as they occur. By continuously monitoring and analyzing turbine operating conditions, the system can predict and prevent failures in the wind farm group.

Source

An independent condition monitoring system for wind turbines that can be retrofitted to existing turbines to provide health monitoring without requiring modifications to the main controller. The system has a local sensor to acquire data like temperature and humidity near a device like a dehumidifier. This sensor connects to a signal processing unit that communicates with the main controller. The local monitoring can provide early warning of issues like dehumidifier failure before it affects the turbine. The system can also have an alarm function. It operates independently of the main controller and can be connected via Ethernet. This allows customization and expansion of monitoring beyond the standard turbine signals.

Source

Monitoring wind turbine components using machine learning models trained on different turbines. The method involves mapping vibration signals from the component being monitored to the signals of a comparable component from a different turbine type. This allows using a pre-trained ML model for fault detection on the new turbine, since the mapped signals have equivalent fault patterns. The mapping is based on kinematic parameters of the components.

Source

A method for detecting faults in wind turbines using a data acquisition device and server that reduces the amount of wind turbine operating parameter data transmitted between the devices. The method involves generating configuration parameters and collection instructions for the data acquisition device based on sensor parameters. This controls the acquisition state of all sensors, reducing the number of collected parameters and wireless data transmission overhead. The server selectively configures sensors to collect parameters and sends instructions to the device to do so. This improves data transmission efficiency between the acquisition device and server.

Source

Detecting sensor failure in wind turbines to prevent unprotected operation. The method involves continuously monitoring the data collected by sensors in a wind turbine. It calculates the difference between consecutive readings to check for changes. If the difference stays within a threshold for a certain time, it indicates the sensor is failing. This allows proactive detection of sensor issues before they cause turbine faults.

Source

Sensor fault detection for wind turbines using a sliding mode observer to improve accuracy compared to traditional methods. The method involves constructing a sliding mode observer based on the fan's linear variable parameter model. Sensor data and fan controller output are fed into the observer to generate a sensor fault compensation signal. This signal is then used to correct the fan controller output, allowing continued operation even with sensor faults. The observer accounts for model uncertainty and performance degradation.

Source

Intelligent integrated condition monitoring system for offshore wind turbines that provides comprehensive monitoring of the wind turbine components to detect faults early and improve availability. The system uses sensors like accelerometers, eddy current displacement sensors, and inclinometers at the blades, root flanges, hub, transmission, bearings, and cabin to collect vibration, deformation, and position data. This data is transmitted to a central receiving module in the cabin for analysis and alerting of fault conditions. It allows real-time monitoring of all bolts and components versus selective monitoring of individual bolts. The system aims to reduce costs by eliminating the need for multiple auxiliary monitoring systems and providing complete coverage of all components.

Source

Protection management system for wind power equipment that enables predictive maintenance and fault diagnosis of wind turbines to improve reliability and reduce downtime. The system has modules for monitoring turbine health, predicting failures before they occur, and diagnosing faults when they do occur. It uses sensors and data analysis to detect issues like bearing degradation, gearbox problems, and blade damage, allowing proactive maintenance and repair.

Source

Monitoring wind turbine operation using probability density functions to provide more accurate and predictive warning and alarm thresholds. Instead of fixed thresholds, the method collects parameter values over a window, determines the actual probability density function, and compares it to a reference function. If the similarity exceeds thresholds, an alarm is triggered. This allows detecting trends and anticipating dangerous situations rather than just exceeding thresholds.

Source

Wind turbine gearbox condition monitoring system using STM32 microcontroller to detect and transmit gearbox health data. The system has a front-end module with sensors, amplifiers, ADC, DSP, and STM32, connected to a network module. The front-end module acquires vibration, temperature, and speed signals from the gearbox, converts them digitally, processes them, and sends them over the network to a monitoring center. The STM32 manages communication and data handling. This allows remote monitoring and diagnosis of gearbox conditions using standardized STM32 components.

Source

Online monitoring device for wind turbine blades to detect blade damage and faults in real-time during operation. The device has sensors attached to the blade tips, roots, and leading edges to measure parameters like ice accumulation, crack propagation, screw loosening, vibration, and stress levels. These sensors continuously monitor blade health and transmit data for remote analysis to detect blade damage and failures. This allows early diagnosis and prevention of blade issues compared to traditional periodic inspections.

Source

Fault monitoring system for wind turbines that reduces false positives compared to traditional two-channel redundant systems. The system uses three independent monitoring devices instead of two, and determines a fault only if at least two devices indicate one. This prevents single device failures causing false alarms. The outputs from the three devices are combined using logic like interconnected switches or a PLC to generate a fault output.

Source

A compact, remote-monitoring wind turbine fault detection device that can be mounted on the turbine nacelle and provides real-time data on turbine health. The device has a modular design with a housing, sensor acquisition board, processing board, power supply, and communication unit. The sensor board connects to turbine sensors like vibration, temperature, and electrical signals. The processing board analyzes the sensor data. The communication unit transmits the processed data wirelessly or via Ethernet. The device uses Power over Ethernet (PoE) to provide power and data transmission over a single cable. This allows remote monitoring of turbine health without needing dedicated cabling or power sources on the turbine.

Source

Method to diagnose and simulate hardware signals of wind turbines to quickly and accurately determine the cause of turbine hardware failures and perform signal simulation. The method involves comparing measured sensor signals against pre-set standard signals. If a sensor's signal deviates from the standard, it indicates a fault. The method involves collecting sensor signals, comparing them to standards, and diagnosing faults. It also involves simulating signals on lines to test components. This improves diagnosis efficiency and maintains wind turbine operation reliability.

Source

A method for fault detection and fault tolerance in wind turbine measurement and control systems that improves availability and reliability. The method uses fault detection based on set membership estimation to determine if the system is normal. If the detection indicates a fault, virtual sensors and actuators are used to correct the fault state for fault tolerance. This avoids shutdowns and performance degradation from letting the controller handle faults. By isolating and correcting faults earlier, it reduces major system failures.

Source

Accurately determining the working state of a wind turbine to enable earlier fault detection and avoid misjudgments compared to just using unit temperatures. The method involves using a component temperature estimation model based on parameters like external temp, working conditions, and component temps. By detecting abnormal components from the model, it accurately reflects wind turbine faults versus temperature increases due to environmental or operating factors.

Source

A wind turbine vibration fault state monitoring and intelligent diagnosis system to detect and diagnose vibration faults in wind turbines. The system uses sensors on the turbine components like the impeller, gearbox, and generator to measure vibration, temperature, and speed. The sensor data is acquired and integrated using a central control room switch. A diagnosis server analyzes the integrated data to detect faults and diagnose issues in the turbine components. This allows remote monitoring and diagnosis of wind turbine vibration faults.

Source

A PSO-ANFIS-based wind turbine sensor fault diagnosis system that accurately detects faulty wind turbine sensors. The system uses particle swarm optimization (PSO) and adaptive network-based fuzzy inference system (ANFIS) techniques to analyze sensor data and identify faulty sensors. It involves a sensor prediction unit with layers like input, mapping, bottleneck, de-mapping, and output. The PSO optimizes parameters for training the ANFIS. The system compares predicted and actual sensor values, detects discrepancies, and diagnoses sensor faults.

Source

Wind turbine gearbox fault warning and diagnosis system that uses sensors, modules, and wireless communication to detect and diagnose gearbox faults in wind turbines. The system has sensors for vibration, temperature, and bearing vibration in the gearbox. The signals are converted and transmitted wirelessly to a central processing module. It classifies the faults and sends alerts via wireless communication. The system provides early warning and diagnosis of gearbox faults to enable proactive maintenance.

Source

A system for monitoring the state of wind turbine blades without using strain sensors or accelerometers. The system determines blade state parameters using 3D scans of the blades collected from a LiDAR system. It compares the scans to reference models to identify blade deformations, deflections, and other parameters. This allows monitoring of blade conditions like deflection, twist, vibration, and pitch without installing many sensors inside the blades.

Source

Wind turbine failure early warning system to monitor vibration and impact conditions of wind turbine components to improve reliability by detecting and diagnosing component faults in real time. The system uses sensors on critical parts like blades, towers, and transmission chains to collect vibration and impact data. The sensor signals are preprocessed, then analyzed by a signal processing device to generate fault diagnoses. These diagnoses are communicated remotely in real time to enable remote monitoring and fault detection.

Source

An intelligent monitoring system for wind turbines that uses sensors to detect faults in wind turbine components like gears and bearings. The system has sensors mounted on the turbine to measure vibrations and other parameters. The sensor data is wirelessly transmitted to a central computer that analyzes it to detect faults. If a fault is found, an alarm is triggered to alert the operator. This allows early identification and isolation of faults before they escalate and cause failures.

Source

Method to monitor gearbox faults in wind turbines that enables reliable transmission of fault data to a remote monitoring center even when the turbine itself fails. The method involves using a local monitoring device with sensors, data acquisition, storage, analysis, and communication capabilities. It continuously monitors the gearbox and compares data with nearby turbines. If fault thresholds are exceeded or data discrepancies are large, fault information is sent to the remote center. If the main turbine fails, the local device relays fault data using nearby turbines as relays. This ensures fault data is transmitted even if the main turbine goes offline.

Source