Research on the Icing Diagnosis ofWind Turbine Blades Based on FS–XGBoost–EWMA

In winter,wind turbines are susceptible to blade icing,which results in a series of energy losses and safe operation problems.Therefore,blade icing detection has become a top priority.Conventional methods primarily rely on sensor monitoring,which is expensive and has limited applications.Data-driven blade icing detection methods have become feasible with the development of artificial intelligence.However,the data-driven method is plagued by limited training samples and icing samples;therefore,this paper proposes an icing warning strategy based on the combination of feature selection(FS),eXtreme Gradient Boosting(XGBoost)algorithm,and exponentially weighted moving average(EWMA)analysis.In the training phase,FS is performed using correlation analysis to eliminate redundant features,and the XGBoost algorithm is applied to learn the hidden effective information in supervisory control and data acquisition analysis(SCADA)data to build a normal behavior model.In the online monitoring phase,an EWMA analysis is introduced to monitor the abnormal changes in features.A blade icing warning is issued when themonitored features continuously exceed the control limit,and the ambient temperature is below 0℃.This study uses data fromthree icing-affected wind turbines and one normally operating wind turbine for validation.The experimental results reveal that the strategy can promptly predict the icing trend among wind turbines and stably monitor the normally operating wind turbines.

Detecting Icing on the Blades of a Wind Turbine Using a Deep Neural Network

The blades of wind turbines located at high latitudes are often covered with ice in late autumn and winter,where this affects their capacity for power generation as well as their safety.Accurately identifying the icing of the blades of wind turbines in remote areas is thus important,and a general model is needed to this end.This paper proposes a universal model based on a Deep Neural Network(DNN)that uses data from the Supervisory Control and Data Acquisition(SCADA)system.Two datasets from SCADA are first preprocessed through undersampling,that is,they are labeled,normalized,and balanced.The features of icing of the blades of a turbine identified in previous studies are then used to extract training data from the training dataset.A middle feature is proposed to show how a given feature is correlated with icing on the blade.Performance indicators for the model,including a reward function,are also designed to assess its predictive accuracy.Finally,the most suitable model is used to predict the testing data,and values of the reward function and the predictive accuracy of the model are calculated.The proposed method can be used to relate continuously transferred features with a binary status of icing of the blades of the turbine by using variables of the middle feature.The results here show that an integrated indicator systemis superior to a single indicator of accuracy when evaluating the prediction model.

Heat Transfer Analysis on Wire Icing and the Current Preventing from Icing

This study concerns the heat transfer processes during ice accretion on wires. The steady state heat balance equation assumed to describe the thermodynamics at the surface of a current heated wire subjected to icing is obtained by analyzing and computing each terms of heat flux. The surface temperature of wire is derived from the heat balance equation, which gives out a proposed estimation of the current intensity to prevent the wire icing

Numerical Forecasting of Icing on Structural Components of Offshore Platforms in Polar Regions

The Polar Regions are rich in natural resources but experience an extremely cold climate.The surfaces of offshore platforms operating in the Polar Regions are prone to icing.To develop solutions to this problem of surface icing,the influence of both the liquid water concentration of the surrounding atmosphere and the average water droplet diameter on the formation of ice on two major structural components of offshore platforms was analyzed using a combination of Fluent and FENSAP-ICE.Results showed that at a wind speed of 7 m/s,as the concentration of liquid water in the air increases from 0.05 to 0.25 g/m3,the amount and thickness of the icing on the surfaces of the two structural components increase linearly.At a wind speed of 7 m/s and when the size of the average water droplet diameter is 20–30(30–35)μm,as the average water droplet diameter increases,the amount and thickness of the icing on the surfaces of the two structural components increase(decrease)gradually.

Observational Study on the Supercooled Fog Droplet Spectrum Distribution and Icing Accumulation Mechanism in Lushan, Southeast China

A fog monitor, hotplate total precipitation sensor, weather identifier and visibility sensor, ultrasonic wind speed meter,an icing gradient observation frame, and an automated weather station were involved in the observations at the Lushan Meteorological Bureau of Jiangxi Province, China. In this study, for the icing process under a cold surge from 20–25 January2016, the duration, frequency, and spectrum distribution of agglomerate fog were analyzed. The effects of rain, snow, and supercooled fog on icing growth were studied and the icing and meteorological conditions at two heights(10 m and 1.5 m)were compared. There were 218 agglomerate fogs in this icing process, of which agglomerate fogs with durations less than and greater than 10 min accounted for 91.3% and 8.7%, respectively. The average time interval was 10.3 min. The fog droplet number concentration for sizes 2–15 μm and 30–50 μm increased during rainfall, and that for 2–27 μm decreased during snowfall. Icing grew rapidly(1.3 mm h-1) in the freezing rain phase but slowly(0.1 mm h-1) during the dry snow phase. Intensive supercooled fog, lower temperatures and increased wind speed all favored icing growth during dry snow(0.5 mm h-1). There were significant differences in the thickness, duration, density, and growth mechanism of icing at the heights of 10 m and 1.5 m. Differences in temperature and wind speed between the two heights were the main reasons for the differences in icing conditions, which indicated that icing was strongly affected by height.

Prediction Model of Aircraft Icing Based on Deep Neural Network

Icing is an important factor threatening aircraft flight safety.According to the requirements of airworthiness regulations,aircraft icing safety assessment is needed to be carried out based on the ice shapes formed under different icing conditions.Due to the complexity of the icing process,the rapid assessment of ice shape remains an important challenge.In this paper,an efficient prediction model of aircraft icing is established based on the deep belief network(DBN)and the stacked auto-encoder(SAE),which are all deep neural networks.The detailed network structures are designed and then the networks are trained according to the samples obtained by the icing numerical computation.After that the model is applied on the ice shape evaluation of NACA0012 airfoil.The results show that the model can accurately capture the nonlinear behavior of aircraft icing and thus make an excellent ice shape prediction.The model provides an important tool for aircraft icing analysis.

APPLICATION OF CFD TECHNOLOGY IN WIND TURBINE ICING PROBER DESIGN

A series of numerical methods,which are suitable to design the shape and configuration of the icing prober for the horizontal axis wind turbine,are presented.The methods are composed of a multiple reference frame(MRF)method for calculating flow field of air,a Lagrangian method for computing droplet trajectories,an Eulerian method for calculating droplet collection efficiency,and an arithmetic for fast computing ice accretion.All the numerical methods are based on the computational fluid dynamics(CFD)technology.After proposing the basic steps and ideas for the design of the icing detection system,the shape and configuration of the icing prober for a 1.5 MW horizontal axis wind turbine are then obtained with the methods.The results show that the numerical methods are efficient and the CFD technology plays an important role in the design process.

Quantitative assessment of flight safety under atmospheric icing conditions

A quantitative assessment method is proposed to sense the specific effects of atmospheric icing conditions on flight safety. A six degree-of-freedom computational flight dynamics model is used to study the effects of ice accretion on aircraft dynamics, and a pilot model is also involved. In order to investigate icing severity under different icing conditions, support vector regression is applied in establishing relationship between aircraft icing parameter and weather conditions. Considering the characteristics of aircraft icing accidents, a risk probability assessment model optimized by the particle swarm method is developed to measure the safety level. In particular, angle of attack is chosen as a critical parameter in this method. Results presented in the paper for a series of simulation show that this method captures the basic effects of atmospheric icing conditions on flight safety, which may provide an important theoretical reference for icing accidents avoidance.

Investigation of Heterogeneous Ice Nucleation on the Micro-Cubic Structure Superhydrophobic Surface for Enhancing Icing-Delay Performance

The aim of this study is to explore the heterogeneous ice nucleation behavior based on controllable micro-cubic array structure surfaces from the statistic perspective.To this end,we firstly constructed a group of micro-cubic array structures on silicon substrates by a selective plasma etching technique.After grafting low-free-energy substance,the as-constructed micro-cubic array structure surfaces exhibited higher non-wettability with the water contact angle being up to 150°.On this basis,500 cycles of freezing and melting processes were accurately recorded to analyze the instantaneous ice nucleation behavior according to the statistical results of freezing temperature.As a consequence,the statistical freezing temperature of the sample with micro-spacing distance of 40μm is as low as−17.13°C.This microstructure configuration(conforming to Cassie-Baxter wetting regime)not only could entrap more air pockets,but also achieved lower solid-liquid contact area,resulting in lower ice nucleation rate(~2–3 orders of magnitude less than that on the flat substrate).Furthermore,the gradually increasing micro-spacing distance to 60μm would induce the transition from CassieBaxter to Wenzel wetting state,leading to higher freezing probability and ice nucleation rate.The complete understanding on microstructure configuration improving the ice nucleation will lay the foundation stone for the microstructure design of ice-repellent materials.

Study on Characteristics and Climatic Causes of Guangxi Conductor Icing

The icing disaster data and field investigation data of conductor icing were analyzed,the results indicated that conductor icing is mostly mixed freezing of glaze and rime in Guangxi,it becomes more severe with the higher of latitude and the increasing of altitude;the thickness of windward side is larger than leeward side in the same elevation.The closer to reservoir and rivers,the more serious the icing degree is.When cold wave broke,southwest warm-moist airflow was transported like an endless stream to the upper air of South China and surface stationary front sustained in South China coast.The main reasons of coming into being conductor icing were a stable strong frontal zone sustained in the upper air in Guangxi and there was strong inversion layer in the middle and lower troposphere,at that time,it was easy to generate conductor icing.The terrain generated an important effect on conductor icing.

Influence of liquid water content on wind turbine blade icing by numerical simulation

In order to research the influence of liquid water content ( LWC ) on blade icing of wind turbine, a numerical simulation method for blade icing was established. The numerical simulation was based on low speed viscous N-S equation. The trajectory equation of water droplets was established by Lagrangian method. The mass and energy conservation equations of the water droplets impacting on the surface of the blade were solved based on control body theory. Three sections along blade span wise of a 1.5 MW wind turbine were decided to simulate icing. Five kinds of LWC were selected for simulation including 0.2,0.4,0.6,0.8 and 1.0 g/m^3 under two ambient temperatures of -10 ℃ and -20 ℃. The medium volume droplet diameter ( MVD ) was 30μm. The simulations included icing shape on blade surface, dimensionless icing area and dimensionless maximum stagnation thickness. Furthermore, the flow fields around both the iced blade airfoil and the original one were simulated and analyzed. Accor-ding to the results, the typical icing characteristics of icing shape, icing area and thickness were greatly affected by the difference of LWCs. This study can provide theoretical reference for the research on antiicing and deicing of wind turbine blade.

Expert System for Icing Prediction and Ice Melting Control

Atmospheric icing may have a great impact on the overall design and safety running of power transmission lines or railway contact wires. For engineers attempting to determine ice loads in a specific area,knowledge of historical ice events and ice type for that area would be extremely valuable information. However,there is no systematic ice accretion database that can be used to help engineers. The prolonged 2008 ice storm of central and southern China caused extensive damage to the electrical installations and railway contact wires. As a result,the transmission provider of the state has deployed substantial efforts to mitigate the effects of future ice storms. This paper describes the development of a knowledge-based decision support system for icing prediction and melting control strategy. The knowledge acquisition process is conducted based on the experimental and theoretical research. The system can predict if icing presents in the current condition,and estimate the possible icing types based on the meteorological parameters such as air temperature,relative humidity,wind speed,which are obtained by corresponding sensors. When icing presents,then the system will make a melting decision using the heat generated by the currents flowing in the wires based on the melting algorithm. Simultaneously the ice scene is monitored for security by real-time videos. This system can be used in power transmission lines or railway contact wires.

Research on the Change of Airfoil Geometric Parameters of Horizontal Axis Wind Turbine Blades Caused by Atmospheric Icing

Icing can significantly change the geometric parameters of wind turbine blades,which in turn,can reduce the aerodynamic characteristics of the airfoil.In-depth research is conducted in this study to identify the reasons for the decline of wind power equipment performance through the icing process.An accurate experimental test method is proposed in a natural environment that examines the growth and distribution of ice formation over the airfoil profile.The mathematical models of the airfoil chord length,camber,and thickness are established in order to investigate the variation of geometric airfoil parameters under different icing states.The results show that ice accumulation varies considerably along the blade span.By environmental temperature drop,the minimum and maximum extents of ice accumulation are observed near the blade root(0.2 R)and the blade tip(0.95 R),respectively(R represents the blade length).The icing process steadily increases the chord length and decreases the airfoil curvature,reaching the largest value at the blade tip region.Furthermore,the maximum curvature is reduced to 41.50%of the original curvature.The maximum camber position of the airfoil moves towards the trailing edge,and the most prominent position occurs at the middle blade region(0.6 R),where it moves back by 19.43%.Ice accumulation steadily increases airfoil thickness.It leads to the maximum thickness growth of 53.40%that occurs at the blade tip region and moves forward to the leading edge by 10%.The research results can provide the required theoretical support for further monitoring the blades operating conditions to ensure reliable wind turbines’operation.

A Hybrid Model Based on Back-Propagation Neural Network and Optimized Support Vector Machine with Particle Swarm Algorithm for Assessing Blade Icing on Wind Turbines

With the continuous increase in the proportional use of wind energy across the globe,the reduction of power generation efficiency and safety hazards caused by the icing on wind turbine blades have attracted more consideration for research.Therefore,it is crucial to accurately analyze the thickness of icing on wind turbine blades,which can serve as a basis for formulating corresponding control measures and ensure a safe and stable operation of wind turbines in winter times and/or in high altitude areas.This paper fully utilized the advantages of the support vector machine(SVM)and back-propagation neural network(BPNN),with the incorporation of particle swarm optimization(PSO)algorithms to optimize the parameters of the SVM.The paper proposes a hybrid assessment model of PSO-SVM and BPNN based on dynamic weighting rules.Three sets of icing data under a rotating working state of the wind turbine were used as examples for model verification.Based on a comparative analysis with other models,the results showed that the proposed model has better accuracy and stability in analyzing the icing on wind turbine blades.

Observation and Analysis of Meteorological Conditions for Icing of Wires in Guizhou, China

Icing of wires is a product of rain, fog, and freezing rain, and is a common meteorological disaster in winter in Guizhou Province of China. It is extremely harmful to facilities such as power transmission and communication lines, and has caused huge economic loss up to 48.9566 billion dollars a year. Based on the meteorological records of Guizhou from 1967, we analyze the meteorological characteristics during the icing of wires, and obtain the temperature, wind speed and direction conditions of the ice accident. The icing of wires is carried out by supercooling raindrops, freezing of the clouds, freezing and spreading on the wires. Different types of supercooled raindrops and cloud freezing and freezing processes will form different types of ice accretion;wind direction and wind speed will affect the growth of ice accretion by changing the speed of sub-cooling raindrops and cloud falling. The weight of rain-type ice accretion is between 24 and 152 g, and the weight of smog-type ice is between 40 - 76 g. The average ice density of these two places can be calculated to be 0.2 - 0.5 g/cm3. The longer the icing of wires, the higher the ice disaster rate.

Correlation of icing events with meteorological variables,and two statistical models to discriminate freezing days

In this paper, the characteristics of meteorological variables are statistically correlated with icing events （i.e., glaze and rime） in China, using daily observations of air temperature, relative humidity, wind speed, and weather phenomena from 700 stations in China from 1954 to 2008. The weather conditions most favorable for icing events are investigated and two statistical models are developed to discriminate potential freezing days. Low air temperature, high relative humidity, and low wind speed are shown to be important conditions for occurrence of icing events; also, the favorable daily mean air temperature is shown to have a decreasing trend from north to south in China, while the favorable relative humidity and wind speed varies little across the country. The statistical model developed with the daily mean temperature combined with precipitation, fog, and mist weather phenomena proved to be well able to determine the possible occurrence of freezing days. The accuracy of model outputs is well above 60% for northwestem Yun- nan, Guizhou, northern Guangxi, southern Hunan, and southern Jiangxi, among other regions where icing events are more fre- quent, and the average false alarms are few. Using observations or forecast products of conventional meteorological variables, this model has high performance and is practical and applicable for early warning and monitoring of icing events.

An anti-icing scaling method for wind tunnel tests of aircraft thermal ice protection system

The efficiency of the aircraft Ice Protection Systems(IPSs)needs to be verified through icing wind tunnel tests.However,the scaling method for testing the IPSs has not been systematically established yet,and further research is needed.In the present study,a scaling method specifically designed for thermal IPSs was derived from the governing equation of thin water film.Five scaling parameters were adopted to address the heat and mass transfer involved in the thermal anti-icing process.For method validation,icing wind tunnel tests were conducted using a jet engine nacelle model equipped with a bleed air IPS.The non-dimensional surface temperature and runback ice closely matched for both the reference and scaled conditions.The validation confirms that the scaling method is capable of achieving the similarity of surface temperature and the runback ice coverage.The anti-icing scaling method can serve as an important supplement to the existing icing similarity theory.

Development and Implementation of Physics-Informed Neural ODE for Dynamics Modeling of a Fixed-Wing Aircraft Under Icing/Fault

Accurate dynamics modeling is crucial for the safety and control offixed-wing aircraft under perturbation(e.g.icing/fault).In this work,we propose a physics-informed Neural Ordinary Differential Equation(PI-NODE)-based scheme for aircraft dynamics modeling under icing/fault.First,icing accumulation and control surface faults are considered and injected into the nominal(clean)aircraft dynamics model.Second,the physics knowledge of aircraft dynamics modeling is divided into kinematics and kinetics.The former is universally applicable and borrows directly from the nominal aircraft.The latter kinetics knowledge,which hinges on external forces and moments,is inaccurate and challenging under icing/fault.To address this issue,we employ Neural ODE to compensate for the residual between the aircraft dynamics under icing/fault and the nominal(clean)condition,resulting in a naturally continuous-time modeling approach.In experiments,we benchmark the proposed PI-NODE against three baseline methods in a dedicated flight scenario.Comparative studies validate the higher accuracy and improve the generalization ability of the proposed PI-NODE for aircraft dynamics modeling under icing/fault.

Research progress on construction strategy and technical evaluation of aircraft icing accretion protection system

The impact of unstable supercooled water droplets suspended in the cloud on the solid will cause its surface to freeze,and the flight safety of the aircraft will be seriously affected when flying in this environment.Aircraft icing protection system is an important device to reduce icing accidents and improve aircraft safety performance,which is of great significance to ensure flight safety.Based on the energy source,this paper proposes a general strategy for constructing an aircraft icing protection system,including Active Anti-icing and De-icing(AAD)system,Passive Antiicing and De-icing(PAD)system and Composite Anti-icing and De-icing(CAD)system.The principle,scope of application,advantages and disadvantages of aircraft anti-icing and de-icing technologies such as electric pulse de-icing,low-frequency piezoelectric de-icing,and hydrophobic material anti-icing are explored in detail,and the corresponding improvement measures are proposed.The future development of aircraft anti-icing and de-icing technology is prospected,and some new ideas are provided for the improvement of aircraft anti-icing and de-icing technology.