Ice Detection Research Articles

Underperformance analysis of wind turbines -Icing -

Wind energy systems are the most promising resources on green energy, it needs to follow the increase demand of electrical energy not only by the future technologies but also with improving the performances. In this work, we present a first performance study of the “Icing System” that monitors and controls the ice detection on blades, it stops the machines when the corresponded sensor detects the presence of ice on the blades, the problem is that the status “Ice Detection Stop” is activated out of consideration of the nacelle’s position and its direction, this stop cost a lot in terms of production and incomes. The turbine restart automatically when the ice is heated and after several hours of production losses. The main idea of this article is to set up a configuration on monitoring system that refers to available database and the daily registered data from the solar met mast to fix the nacelle direction taking into consideration the sunrise and sunset. Using the sun energy, the applied process will help to speed up the heating of ice on blades and minimise the impact of the stops. As result, the availability will improve as well as the annual production of energy.

A Deep Learning Architecture for Semantic Segmentation of Radar Sounder Data

During the last decades, radar sounders provided direct measurements (radargrams) of the Earth’s polar caps’ subsurface. Radargrams are of critical importance for a better understanding of glaciologic structures and processes of the ice sheet in the framework of climate change. This article aims to automatically extract information on basal boundary conditions given their substantial relevance for modeling the ice-sheet processes, such as the sliding. We introduce a novel automatic method based on deep learning to detect the basal layer and basal units in radargrams acquired in the inland of icy areas. Radargrams are segmented into englacial layers, bedrock, basal units, and noise-limited regions; the latter includes the echo-free zone (EFZ), thermal noise, and signal perturbation. The network is a U-Net with attention gates and the Atrous Spatial Pyramid Pooling (ASPP) module that automatically extract semantically meaningful features at different scales. Experimental results on two datasets acquired in north Greenland and west Antarctica by the Multichannel Coherent Radar Depth Sounder (MCoRDS3) indicate a high overall segmentation accuracy. The accuracy of basal ice and signal perturbation detection is high, and that of the other classes is comparable with the literature techniques based on handcrafted features. The results show the effectiveness of the proposed method in automatically extracting semantically meaningful features to segment radargrams and map the basal layer and basal units.

AMSR2 Thin Ice Detection Algorithm for the Arctic Winter Conditions

We have developed a thin ice detection algorithm for the AMSR2 radiometer data. The algorithm, denoted as AMSR2 thin ice detection algorithm—version 2 (ATIDA2), is targeted for the Arctic Ocean. The detection of thin ice with a maximum thickness of 20 cm is based on the classification of the 36-GHz polarization ratio ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$PR36$ </tex-math></inline-formula> ) and H-polarization 89–36-GHz gradient ratio ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$GR8936H$ </tex-math></inline-formula> ) signatures with a linear discriminant analysis (LDA) and thick ice restoration with <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$GR3610H$ </tex-math></inline-formula> . The thick ice restoration removes erroneous thin ice detections due to thin and thick ice <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$PR36$ </tex-math></inline-formula> and <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$GR8936H$ </tex-math></inline-formula> signature mixing. ATIDA2 is applied only when sea ice concentration (SIC) is ≥70% and the air temperature is ≤−5 °C to decrease misclassification of thick ice as thin ice. For the AMSR2 L1R brightness temperature data, an atmospheric correction is applied following an EUMETSAT OSI SAF correction scheme in SIC retrieval algorithms. ATIDA2 is applied to L1R swath datasets, and then, the results are combined to a daily thin ice chart. ATIDA2 was developed and validated using MODIS ice thickness charts over the Barents and Kara Seas. The average probability for misclassification of thick ice as thin ice in the daily chart is 8.7%, and 37.0% for vice versa. The comparison of the ATIDA2 chart and the SMOS ice thickness chart over the Arctic showed rough correspondence in the thin versus thick ice classification. The ATIDA2 chart is targeted to be used together with SAR imagery for various sea ice classifications.

Toward the Optimal Antenna-Based Wireless Sensing Strategy: An Ice Sensing Case Study

Remote ice detection has emerged as an application of Radio Frequency (RF) sensors. While antenna-based &#x201C;RFID&#x201D; sensing can detect various measurands, antenna-based sensors are not currently designed based on a systematic methodology, and in most cases may have a low sensitivity requiring specialist hardware or broadband interrogation signals, incompatible with spectrum regulations. Here, we develop a systematic methodology for designing an antenna-based sensor, applicable to measurands inducing a dielectric change in the near-field of the antenna. The proposed methodology is applied to designing printable antennas as highly-sensitive sensors for detecting and measuring the thickness of ice, demonstrating best-in-class sensory response compared to more complex antenna designs. Antenna design is investigated systematically for wireless interrogation in the 2.4 GHz band, where it is found that a loop antenna outperforms a dipole owing to its more distributed capacitance. The antenna&#x2019;s realized gain was identified as the optimum parameter-under-test, with &#x201C;positive&#x201D; sensing proposed as a method of improving linearity and immunity to interference. The developed loop antenna sensor exhibits resilience to interference and applicability to different real-world deployment environments, demonstrated through over 80% average ice thickness measurement accuracy and at least 5 dB real-time sensitivity to ice deposition.

Assessing the Effects of Induced Field Rotation on Water Ice Detection of Tianwen-1 Full-Polarimetric Mars Rover Penetrating Radar

China’s first Mars probe Tianwen-1 has successfully landed on the southern Utopia Planitia of Mars on May 15, 2021. The Zhurong rover is first equipped with a full-polarimetric Mars Rover Penetrating Radar (FP-RoPeR) system, aiming to map the subsurface fine structure and to find the potential underground water ice. However, different from the previous water ice detection of orbital radar, the FP-RoPeR signals will be affected by the induced field rotation (IFR) if electromagnetic (EM) waves propagate through rough interfaces. Therefore, in this article, we assess the IFR effects from rough interfaces on the circular polarization ratio (CPR) response of FP-RoPeR data, which is a significant parameter for water ice detection. The theoretical computation and numerical validation indicate that the depth, the number of rough interfaces, and relative permittivity are three vital parameters that affect the IFR effects; depth plays a more important role than the other two for FP-RoPeR system. The CPR estimation result will be with greater error in the shallow region (0–1 m). The relative error in the region of depth greater than 1 m can be guaranteed to be under 10%.

Ice Detection of Wind Turbine Blades Based on Lstm Networks and Discrete Wavelet Transform

Ice Detection of Wind Turbine Blades Based on Lstm Networks and Discrete Wavelet Transform

A Novel Ellipsoidal Semisupervised Extreme Learning Machine Algorithm and Its Application in Wind Turbine Blade Icing Fault Detection

The conventional semi-supervised extreme learning machine (SS-ELM) algorithm can provide a solution to the lack of labeled samples in wind turbine blade icing fault detection, but its performance is limited by the irrationality of the spherical nearest neighbor graph (SNNG) calculation strategy. To solve this problem, a novel ellipsoidal semi-supervised extreme learning machine (ESS-ELM) algorithm is proposed in this paper and applied to wind turbine blade icing fault detection. In this study, we creatively propose a novel ellipsoidal nearest neighbor graph (ENNG) calculation strategy that considers the distribution information of the labeled samples to construct the ESS-ELM algorithm. Different from the conventional SNNG, the ENNG can adaptively assign corresponding calculation weights to each feature in the data space according to the Fisher criterion, which allows it to better match the smoothness assumption of the manifold regularization learning framework. In addition, the two adjustable parameters of the enhancement factor and the degradation factor are ingeniously introduced into the ENNG, which further ensure the applicability and security of the proposed ESS-ELM. The superiority of the ESS-ELM algorithm is verified by extensive benchmark industrial fault datasets and the real-world icing dataset of two wind turbines. It is demonstrated that the proposed ESS-ELM algorithm achieves better performance than the ELM, its existing variants, and some state-of-the-art wind turbine blade icing detection methods.

Wind Turbine Blade Icing Detection Based on Random Forest

In this paper, a multimodal random forest method is proposed to solve the problem that the traditional linear method has low accuracy in predicting the icing of generator blade. This paper uses the random forest algorithm to model and analyze the icing event of wind turbine blades, and describes in detail the process of using the C4.5 algorithm to generate a decision tree, and then randomly selecting samples and sample features to train to generate a random forest. The effectiveness of the method in this paper is experimentally verified by SCADA data, and the results show that the multi-modal random forest algorithm reduces the error rate to 1.97% in wind turbine blade icing prediction, which is more accurate than the traditional method.

Simple Method to Extract Lake Ice Condition From Landsat Images

Ice plays key roles in regulating hydrological, ecological, biogeochemical, and socioeconomic functions of lakes. Long-term <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">in situ</i> lake ice phenological records indicate that lake ice is trending toward later freeze-up, earlier breakup, and a shorter ice duration. Parallel to study of lake ice using <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">in situ</i> records and process-based models, satellite remote sensing can expand our understanding of lake ice change over large spatial scales. However, most remote sensing studies have focused on large lakes or short periods of time, which may not robustly represent changes over multidecadal time periods or in the much more numerous small lakes. Here, we present a random forest model, Sensitive Lake Ice Detection (SLIDE), to accurately extract ice conditions from Landsat TM, ETM+, and OLI images. We trained the model using a manually labeled lake ice dataset (1089 labeled areas over 995 lakes globally). Our results show that our model achieves accurate classification between ice/snow and water (accuracy: 97.8%, kappa coefficient: 95.5%). Comparing Landsat-derived ice cover with <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">in situ</i> ice conditions, we show that our model produces less bias, lower RMSE, and higher kappa than does the Landsat snow/ice flag from the quality assessment band. This is especially true during the transitional period surrounding the ice on and off dates reported from <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">in situ</i> (mean bias −7.3% from our model, −17.3% from the Landsat quality band). Our results demonstrate the feasibility of mining the rich Landsat archive to study lake ice dynamics and of better flagging ice-affected lake observations.

Arctic Sea Ice Classification Based on CFOSAT SWIM Data at Multiple Small Incidence Angles

Sea ice type is the key parameter of Arctic sea ice monitoring. Microwave remote sensors with medium incidence and normal incidence modes are the primary detection methods for sea ice types. The Surface Wave Investigation and Monitoring instrument (SWIM) on the China-France Oceanography Satellite (CFOSAT) is a new type of sensor with a small incidence angle detection mode that is different from traditional remote sensors. The method of sea ice detection using SWIM data is also under development. The research reported here concerns ice classification using SWIM data in the Arctic from October 2019 to April 2020. Six waveform features are extracted from the SWIM echo data at small incidence angles, then the distinguishing capabilities of a single feature are analyzed using the Kolmogorov-Smirnov distance. The classifiers of the k-nearest neighbor and support vector machine are established and chosen based on single features. Moreover, sea ice classification based on multi-feature combinations is carried out using the chosen KNN classifier, and optimal combinations are developed. Compared with sea ice charts, the overall accuracy is up to 81% using the optimal classifier and a multi-feature combination at 2°. The results reveal that SWIM data can be used to classify sea water and sea ice types. Moreover, the optimal multi-feature combinations with the KNN method are applied to sea ice classification in the local regions. The classification results are analyzed using Sentinel-1 SAR images. In general, it is concluded that these multifeature combinations with the KNN method are effective in sea ice classification using SWIM data. Our work confirms the potential of sea ice classification based on the new SWIM sensor, and highlight the new sea ice monitoring technology and application of remote sensing at small incidence angles.

A flexible surface-mountable sensor for ice detection and non-destructive measurement of liquid water content in snow

Ice and snow accumulation on structures can cause serious safety issues due to falling pieces of accumulated ice as well as overloading of the structure. It is crucial to detect ice and the degree of wetness of snow when selecting a mitigation strategy such as applying anti-icing/de-icing techniques and/or notifying personnel to take necessary action. In this study, a sensor was developed to detect the ice presence and determine the liquid water content (LWC) by mass of wet snow accumulated on a surface through measuring the electrical resistance and temperature of the medium covering the surface, namely, the Ice Presence and Snow Liquid Water Content (ISLWC) sensor. This is a lightweight and flexible sensor that can be installed on surfaces such as bridge cables and provides a real-time and non-destructive way to detect ice or measure the LWC of snow in layers as thin as a few millimeters. The detection algorithm used was developed using the thresholds obtained from numerous indoor experiments data. Outdoor tests verified the detection accuracy and reliability as well as the ISLWC sensor durability under extreme weather conditions. Both indoor and outdoor experiment results showed that this sensor is an efficient tool for real-time monitoring and mitigation of icing problems on structures.

Ice imaging in aircraft anti-icing fluid films using polarized light

During qualification endurance tests of aircraft de−/anti-icing fluids, their failure is evaluated by a naked-eye visual inspection and occasionally poking freezing areas with a toothpick. For ice detection improvement, image processing techniques have been developed in the visible and infrared spectra. Such techniques have issues with low ice contrast or combining with the visual inspection. Here we show how to enhance ice contrast in the visible spectrum by using ice birefringence and polarized light reflection. The method can be used for both visual inspection and automatic ice detection systems.

Magnetically coupled planar microwave resonators for real-time saltwater ice detection

Buildup of saltwater ice can cause serious issues in many industries if not identified. This work presents a comparison between electrically and magnetically coupled planar microwave resonators for saltwater ice detection applications. Due to the highly coupled nature of both sensors, the frequency splitting phenomenon occurs resulting in each sensor having two resonant frequencies. The proposed electrically coupled resonator has operational frequencies of 4.726 GHz and 4.942 GHz whereas the magnetically coupled resonator has operational frequencies of 4.647 GHz and 4.960 GHz. The quality factor of the low and high frequency for the electrically coupled resonator were 260 ± 2 and 175 ± 2 respectively and were 195 ± 2 and 260 ± 2 respectively for the magnetically coupled resonator. The sensitivity of each sensor was investigated through simulation. Experiments were conducted using reverse osmosis (RO) water and saltwater with 5% and 10% salt concentrations. These experiments demonstrated the magnetically coupled sensor to be sensitive to the formation of both RO ice and saltwater ice, whereas the electrically coupled sensor could only detect the presence of RO ice due to the increased conductivity of saltwater. This magnetically coupled sensor is a proof of concept for saltwater ice sensing and is well suited for the application due to its high sensitivity, low-cost, and real-time sensing capabilities.

Unsupervised transfer learning for condition monitoring of a fleet of wind turbines

The condition monitoring and health status prediction of a fleet of wind turbines are essential for the safety of wind turbines. At present, the Supervisory Control And Data Acquisition (SCADA) system has been widely used in wind turbines, which can monitor and collect various physical information and sensor information of wind turbines in real-time. Due to the fact that the amount of data obtained by SCADA systems is extremely large, developing an intelligent decision-making system based on deep learning is a very valuable research. Therefore, this paper is committed to exploring a health status prediction algorithm of wind turbines based on deep learning and SCADA systems. However, yet in actual industrial applications, it is very time-consuming and expensive to obtain a large amount of labeled data. In addition, as failures rarely occur, there is a serious sample imbalance problem in the datasets. More importantly, due to the difference in working environment and physical parameters, there are significant differences in the feature distribution of different wind turbines data, which lead to a significant drop in the performance of the deep learning model on unknown wind turbines.&#x0D; Therefore, we propose an unsupervised transfer learning algorithm based on Generative Adversarial Networks for wind turbine health status prediction (WT-GAN). WT-GAN can not only remove the domain shift between wind turbines, but also it is an unsupervised learning method. This practically means that only the unlabeled data for the target domain is required, which solves the problem of labeling data. In order to evaluate the effectiveness of WT-GAN on the condition monitoring of a fleet of wind turbines, we apply this method to one dataset about blade icing detection of wind turbines. The experimental results prove that the proposed method can predict the health of the wind turbine well. In addition, it can significantly reduce the domain shift among different wind turbines, thereby achieving excellent performance on unknown wind turbines.

Application of a Convolutional Neural Network for the Detection of Sea Ice Leads

Despite accounting for a small fraction of the surface area in the Arctic, long and narrow sea ice fractures, known as “leads”, play a critical role in the energy flux between the ocean and atmosphere. As the volume of sea ice in the Arctic has declined over the past few decades, it is increasingly important to monitor the corresponding changes in sea ice leads. A novel approach has been developed using artificial intelligence (AI) to detect sea ice leads using satellite thermal infrared window data from the Moderate Resolution Imaging Spectroradiometer (MODIS) and the Visible Infrared Imaging Radiometer Suite (VIIRS). In this new approach, a particular type of convolutional neural network, a U-Net, replaces a series of conventional image processing tests from our legacy algorithm. Results show the new approach has a high detection accuracy with F1 Scores on the order of 0.7. Compared to the legacy algorithm, the new algorithm shows improvement, with more true positives, fewer false positives, fewer false negatives, and better agreement between satellite instruments.

Towards standards in the analysis of wind turbines operating in cold climate – Part A: Power curve modeling and rotor icing detection

Analyzing the performance of wind turbines operating in cold climate conditions is a complex matter. The lack of any widely used standards adds to the challenge of comparing turbine performance between and within a wind farm. It proves to be an even bigger issue when alternative operational strategies, including ice protection systems (IPS) are implemented and their performance has to be assessed. Before making any comparison between turbines, it is important to build robust power curve models that are not influenced by secondary factors (e.g. temperature and wind orientation). An accurate power curve model allows to detect more precisely any ice formed on the rotor. This paper details a process that allows to build a simple power curve model while minimizing the influence of temperature and wind direction. The piecewise function used to represent the power curve yields a precise parametric representation while being more robust to outliers. Different tools being used to evaluate the performance of the model are presented. With these tools it was possible to observe that standard air density corrections were not able to completely compensate the effect of temperature on the power output. The analysis also provides comprehensive definitions of the different periods composing an icing event. A simple and efficient ice detection threshold using only the power curve is proposed and validated. Using this threshold, algorithms defining significant production loss periods and icing events are presented. The different tools introduced in this paper can decrease significantly the uncertainty in the available power estimation, which proves to be crucial when comparing the performance of multiple turbines.

Sea ice detection from SSMI data based on an improved U-Net convolution neural network

ABSTRACT It is of great significance to accurately obtain the polar sea ice for the study of global hydrosphere and polar climate. The U-Net convolutional neural network (CNN) can achieve end-to-end pixel-level semantic segmentation, and it can learn comprehensive image information as much as possible through the training process. The vertical polarization at 37 GHz and 19 GHz of Special Sensor Microwave/Imager (SSM/I) is used as the input of the improved U-Net CNN, and the improved U-Net CNN is proposed to detect the sea ice. In order to improve the accuracy of sea ice detection, anew feature fusion structure based on the improved U-Net CNN is constructed, which further refines the features of different semantic levels. By using batch normalization (BN), dropout regularization and modifying the loss function to improve the U-Net convolutional neural network, the accuracy of sea ice detection is improved. The results were compared with the NASA Team algorithm, which showed that the improved U-Net CNN is afeasible solution. The sea ice obtained from MODIS data is real result, and our comparison with NASA team algorithm and U-Net CNN showed that the modified U-Net CNN improves the accuracy of sea ice detection to 75%. Therefore, the improved U-Net CNN significantly improves the accuracy of sea ice detection. The improved U-Net CNN had the ability to combine multiple nonlinear transformation functions to solve the problems of sea ice identification, which improved the accuracy and robustness of sea ice detection.

Active hyperspectral imager using a tunable supercontinuum light source based on a MEMS Fabry-Perot interferometer.

We have developed an active hyperspectral imager based on a tunable near-infrared supercontinuum light source. Non-dispersive wavelength selection of the supercontinuum laser source is achieved with a microelectromechanical Fabry-Perot interferometer. The tunable light source enables the use of any monochromatic imaging sensor with a suitable spectral sensitivity for hyperspectral imaging. The imager is characterized and demonstrated in the laboratory for remote detection of ice.

A PCC-Ensemble-TCN model for wind turbine icing detection using class-imbalanced and label-missing SCADA data

Blade icing problems are ubiquitous for wind turbines located in cold climate zones. Data-driven indirect icing detection methods based on supervisory control and data acquisition system have shown strong potential recently. However, the supervisory control and data acquisition data is annotated through manual observation, which will cause the data between normal condition and icing condition to be unlabeled. In addition, the amount of normal data is far more than icing data. The above two issues restrict the performance of most current data-driven models. In order to solve the label missing problem, this article proposes a Pearson correlation coefficient–based algorithm for measuring the degree of blade icing, which calculates the similarity between the unlabeled data and the icing data as its label. Aiming at the class-imbalance problem, this article constructs multiple class-balanced subsets from the original dataset by under-sampling the normal data. Temporal convolutional networks are trained to extract features and make predictions on each subset. The final prediction result is obtained by ensembling the prediction results of all temporal convolutional network models. The proposed model is validated using the actual supervisory control and data acquisition data collected from a wind farm in northern China, and the results indicate that ensuring the consecutiveness and class-balance of the data are quite advantageous for improving the detection accuracy.

RETRACTED ARTICLE: Ice detection and voice alarm of wind turbine blades based on belief network

RETRACTED ARTICLE: Ice detection and voice alarm of wind turbine blades based on belief network