Dataset Of Measurements Research Articles

A deep learning framework for solving forward and inverse problems of power-law fluids

We for the first time leverage deep learning approaches to solve forward and inverse problems of two-dimensional laminar flows for power-law fluids. We propose a deep-learning framework, called Power-Law-Fluid-Net (PL-Net). We develop a surrogate model to solve the forward problems of the power-law fluids, and solve the inverse problems utilizing only a small set of measurement data under the assumption that boundary conditions (BCs) can be partially known. In the design of the methods, we incorporate the hard boundary condition constraints to accelerate the iteration of stochastic gradient descent methods for minimizing loss functions. For the forward problems, by incorporating the constitutive parameters into the input variables of neural networks, the PL-Net serves as a surrogate model for simulating the pressure-driven flows inside pipes having cross sections of varying shapes. We investigate the influences of the BC type, activation function type, and number of collocation points on the accuracy of numerical solutions. For the inverse problems, the PL-Net infers the physical quantities or constitutive parameters from a small number of measurements of flow field variables. The BCs of the inverse problems can even be partially known. We demonstrate the effects of BC type, number of sensors, and noise level on accuracy of inferred quantities. Computational examples indicate the high accuracy of the PL-Net in tackling both the forward and inverse problems of the power-law fluids.

Massive MIMO Channel Measurement Data Set for Localization and Communication

Channel state information (CSI) needs to be estimated for reliable and efficient communication, however, user location information is hidden inside and can be further exploited. This article presents a detailed description of a Massive Multi-Input Multi-Output (MaMIMO) testbed and provides a set of experimental location-labeled CSI data. We first focus on the design of the hardware and software of a MaMIMO testbed for gathering multiple CSI data sets. We then show this data can be used for learning-based localization and enhanced communication research. The presented data set is made fully available to the research community. We illustrate that a CSIbased joint communication and sensing processing pipeline can be evaluated and designed based on the collected data set. Specifically, the localization output obtained by a convolutional neural network (CNN) trained on the data sets is used to schedule users to improve the spectral efficiency (SE) of the communication system. Finally, we pose promising directions for further exploiting this data set and creating more data sets.

New Equations for Rate of Energy Dissipation of a Stepped Spillway with Slope less than Critical and Specific Step Height

Stepped spillways disperse energy in floodwater by utilizing their stepping nature. Several researchers investigated the hydraulic and geometric relationships of the stepped spillway with a dam slope above 26.6o, which resulted in energy dissipation. But a few studied stepped spillways with a dam slope of 26.6o with a step height of 0.61 m. This move led to a lack of knowledge and rules for the designers of stepped spillways with slope of 26.6o and a step height of 0.61 m. The authors reviewed the researchers' publications on horizontal stepped spillways with dam slopes of 26.6o with step height of 0.61 m conducted in transition and skimming flows in large-size facilities with phase-detection intrusive probes. They retrieved data sets from them, which they took into account to create two novel equations for energy dissipation in stepped spillways: one for nappe flow and one for skimming flow. The results showed that the measured data sets compared well with the data sets calculated from the two empirical models.

Cross-Level Dependability Assessment With a Distributed Split Mechanism for Wireless Communication Systems

Wireless techniques are playing an increasingly important role in the Industrial Internet of Things. However, the dependability of wireless communication systems is challenged facing high requirements for industrial applications. Dependability assessment models especially supported by machine learning somehow ease the issue, while a cross-level model for distributed wireless devices is lacking. Therefore, this paper proposes a cross-level dependability assessment model with a distributed split mechanism. The model includes device-level assessment sub-models and a system-level assessment sub-model which are trained together but are split and deployed separately by the distributed split mechanism. We implemented this model on a realistic measurement dataset. Comparative experiments indicate that the proposed model not only completes cross-level assessment but also has better performance than past single-level assessment models. Further, experiments on execution time show that the model based on the distributed split mechanism is faster than a centralized mechanism because the split sub-models can process data locally.

A LOF-IDW based data cleaning method for quality assessment in intelligent compaction of soils

Intelligent compaction (IC) is attracting increasing interest in construction engineering involving earthwork compaction. However, outliers may exist in data sets collected during IC, which can lead to misinterpretation of the compaction status of soils and may further result in erroneous results for compaction quality assessment. This study proposes a novel method for cleaning the measured data sets by combining the density-based local outlier factor (LOF) and inverse distance weighted (IDW) method. In the proposed LOF-IDW based method, the effect of spatial variation in soil properties is taken into account, while a boxplot-based method is proposed to dynamically determine the suitable threshold for outlier diagnosis. The capability and performance of the proposed method are verified against three data sets collected from construction sites. The results indicate that compared with the commonly used 3σ criterion, the proposed method not only exhibits a better performance in outlier diagnosis but also can rehabilitate the relevant data. In addition, the proposed method is demonstrated to be capable of significantly reducing the coefficient of variation of the measured data sets, which provides a more reliable support for the quality assessment and decision-making in IC.

Thunderstorm detection from GPM IMERG rainfall: Climatology of dynamical and thermodynamical processes over India

AbstractThis study aims to (i) prepare a premonsoon thunderstorms database, (ii) understand the thunderstorm frequency, duration and intensity and (iii) composite analysis of dynamic and thermodynamic processes related to thunderstorms over India. The thunderstorm associated rainfall varies across India. Hence, a percentile‐based approach is implemented with the integrated multi‐satellite retrievals for global precipitation measurement (IMERG) dataset at 0.1° resolution to identify thunderstorms for 2001–2021. The 93rd percentile appears to be better for thunderstorm detection, with a success ratio of 82% (642 events are confirmed out of 786 detected) in eastern India. Further analysis indicated that 84% of the detected thunderstorms in eastern and northeastern India are associated with lightning activity. Based on this long‐term (2001–2021) thunderstorm data, the highest frequency of thunderstorms (40–45 events·year−1) is observed over the western foothills of the Himalayas, the northeast region, and the west coast of Kerala. The thunderstorm duration in the eastern and northeastern regions and the southwest coast of India is mostly 0.5–2.5 h, producing heavy rainfall (>7 mm·h−1) due to more moisture content and stronger updrafts. The composite structure of thermodynamic indices exhibits significant spatial variations over India and can be used to differentiate the regions of high thunderstorm activity. The minimum (maximum) convective available potential energy (convective inhibition) value required for thunderstorm development is not uniform throughout the country. However, the composites of K index and total totals index during thunderstorms are mostly uniform. This study highlights the benefits of IMERG rainfall in thunderstorm detection over India and helps to understand the local forcings and the effect of thunderstorm activity on different sectors like aviation, agriculture and so forth.

A gridded air quality forecast through fusing site-available machine learning predictions from RFSML v1.0 and chemical transport model results from GEOS-Chem v13.1.0 using the ensemble Kalman filter

Abstract. Statistical methods, particularly machine learning models, have gained significant popularity in air quality predictions. These prediction models are commonly trained using the historical measurement datasets independently collected at the environmental monitoring stations and their operational forecasts in advance using inputs of the real-time ambient pollutant observations. Therefore, these high-quality machine learning models only provide site-available predictions and cannot solely be used as the operational forecast. In contrast, deterministic chemical transport models (CTMs), which simulate the full life cycles of air pollutants, provide predictions that are continuous in the 3D field. Despite their benefits, CTM predictions are typically biased, particularly on a fine scale, owing to the complex error sources due to the emission, transport, and removal of pollutants. In this study, we proposed a fusion of site-available machine learning prediction, which is from our regional feature selection-based machine learning model (RFSML v1.0), and a CTM prediction. Compared to the normal pure machine learning model, the fusion system provides a gridded prediction with relatively high accuracy. The prediction fusion was conducted using the Bayesian-theory-based ensemble Kalman filter (EnKF). Background error covariance was an essential part in the assimilation process. Ensemble CTM predictions driven by the perturbed emission inventories were initially used for representing their spatial covariance statistics, which could resolve the main part of the CTM error. In addition, a covariance inflation algorithm was designed to amplify the ensemble perturbations to account for other model errors next to the uncertainty in emission inputs. Model evaluation tests were conducted based on independent measurements. Our EnKF-based prediction fusion presented superior performance compared to the pure CTM. Moreover, covariance inflation further enhanced the fused prediction, particularly in cases of severe underestimation.

Influence of Solar Wind Variations on the Shapes of Venus’ Plasma Boundaries Based on Venus Express Observations

The interaction between the solar wind and Venus creates an induced magnetosphere. The regions of the induced magnetosphere are separated by plasma boundaries, where their shapes and sizes are influenced by variations in the surrounding environment. Investigations of the boundaries and their variability can help us understand the solar wind’s effect on Venus and unmagnetized planets in general. In this study, the bow shock and ion composition boundary locations are investigated using, for the first time, the full data set of plasma and magnetic field measurements by Venus Express taken during 2006–2014. The locations of the boundaries are examined with respect to upstream conditions and the solar cycle. We find, in agreement with previous studies using Pioneer Venus Orbiter measurements, that the bow shock location is mostly sensitive to the solar cycle and the dynamic pressure and that it exhibits asymmetries in the terminator plane, depending on the direction of the interplanetary magnetic field. The asymmetries are mainly attributed to the asymmetry in the pickup ion distribution and thus mass loading in the magnetosheath. The ion composition boundary on the dayside is found to decrease in altitude with increasing dynamic pressure during solar maximum (2006–2011), but shows no clear trends during solar minimum (2011–2014) conditions.

A Fast Matching Method for the SAR Images with Large Viewing Angles Based on Inertial Navigation Information and Neighborhood Structure Consensus

In the field of multi-view SAR target location, the greater the difference in viewing angles, the higher the target location accuracy. However, this makes it difficult to match the same target between the SAR images with different viewing angles. The performance of traditional SAR image-matching algorithms will deteriorate or even fail to match the images correctly when the viewing angle is gradually increased. To solve this problem, a fast SAR matching method for the SAR images with large viewing angles based on inertial navigation information and neighborhood structure consensus (ININSC) is proposed in this paper. In this algorithm, the key targets are detected in the SAR images by using the maximum connected domain algorithm and the K-means clustering algorithm, and the connected domain centroid of the target is taken as a feature point. The approximate position of the key targets after the viewing angle change is found through inertial navigation information, and then accurate and fast matching is achieved by using the consensus of the neighborhood topological structure of the key targets. The measured data sets demonstrate that compared with traditional SAR image-matching algorithms, the proposed ININSC algorithm solves such a problem that SAR images cannot be accurately matched under the differences at large viewing angles and has better robustness and timeliness.

Global dose distributions of neutrons and gamma-rays on the Moon

Dose assessment on the lunar surface is important for future long-term crewed activity. In addition to the major radiation of energetic charged particles from galactic cosmic rays (GCRs), neutrons and gamma-rays are generated by nuclear interactions of space radiation with the Moon’s surface materials, as well as natural radioactive nuclides. We obtained neutron and gamma-ray ambient dose distributions on the Moon using Geant4 Monte Carlo simulations combined with the Kaguya gamma-ray spectrometer measurement dataset from February 10 to May 28, 2009. The neutron and gamma-ray dose rates varied in the ranges of 58.7–71.5 mSv/year and 3.33–3.76 mSv/year, respectively, depending on the lunar geological features. The lunar neutron dose was high in the basalt-rich mare, where the iron- and titanium-rich regions are present, due to their large average atomic mass. As expected, the lunar gamma-ray dose map was similar to the distribution of natural radioactive elements (238U, 232Th, and 40K), although the GCR-induced secondary gamma-ray dose was significant at ~ 3.4 mSv/year. The lunar secondary dose contribution resulted in an additional dose of 12–15% to the primary GCR particles. Global dose distributions on the lunar surface will help identify better locations for long-term stays and suggest radiation protection strategies for future crewed missions.

Improving the performance of artificial neural networks trained on synthetic data in gas spectroscopy – a study on two sensing approaches

Abstract Artificial neural networks (ANNs) are used in quantitative infrared gas spectroscopy to predict concentrations on multi-component absorption spectra. Training of ANNs requires vast amounts of labelled training data which may be elaborate and time consuming to obtain. Additional data can be gained by the utilization of synthetically generated spectra, but at the cost of systematic deviations to measured data. Here, we present two approaches to train ANNs with a combination of comparatively small, measured data sets and synthetically generated data. For the first approach a neural network is trained hybridly with synthetically generated infrared absorption spectra of mixtures of N2O and CO and measured zero-gas spectra, taken with a mid-infrared dual comb spectrometer. This improves the mean absolute error (MAE) of the network predictions from 0.46 to 0.01 ppmV and 0.24 to 0.01 ppmV for the concentration predictions of N2O and CO respectively for zero-gas measurements which was previously observed for training with purely synthetic data. At the same time a similar performance on spectra from gas mixtures of 0–100 ppmV N2O and 0 to 60 ppmV CO was achieved. For the second approach an ANN pre-trained on synthetic infrared spectra of mixtures of acetone and ethanol is retrained on a small dataset consisting of 26 spectra taken with a mid-infrared photoacoustic spectrometer. In this case the MAE for the concentration predictions of ethanol and acetone are improved by 45 % and 20 % in comparison to purely synthetic training. This shows the capability of using synthetically generated data to train ANNs in combination with small amounts of measured data to further improve neural networks for gas sensing and the transferability between different sensing approaches.

Effects of the fundamental axes of variation in structural diversity on the forest canopy temperature in an urban area

The spatial distribution and heterogeneity of forest canopy elements reveal the fundamental dimensions of plant structure variations. Forests characterized by greater structural complexity and diversity intercept solar radiation more effectively, directly influencing the thermal environment and energy balance of the canopy. However, the axes of variation in the distribution and heterogeneity of the canopy remain largely unknown, which limits our understanding of how structural diversity responds to canopy temperature variability. Here, we derived a set of structural diversity metrics from a dataset of canopy structure measurements obtained using unmanned aerial vehicle-light detection and ranging across major forest communities in an urban area in 2021 and 2022. We also explored the key axes of structural diversity variability and tested their predictive power for canopy temperature. The results showed that: (1) most of the variability within structural diversity (83.6 % and 81.8 %) was captured by the three key axes in 2021 and 2022. The first axis was primarily driven by structural heterogeneity, representing the heterogeneity of vegetation distribution within the canopy. The second axis was primarily influenced by the interaction between height and cover/openness, indicating the vertical structure and horizontal distribution pattern of the canopy. The third axis represented the horizontal coverage and density of the canopy. (2) In both 2021 and 2022, the second axis was identified as the most influential predictor of canopy temperature, as evidenced by R2 values of 0.46 and 0.28, respectively. The model incorporating all three axes of structural diversity achieved the highest accuracy in predicting the canopy temperature for 2021 (R2 = 0.68, AIC = 81.35, ΔAIC = 0, and RMSE = 0.89). Prior research on canopy temperature prediction has overlooked the true potential of principal component axes derived from structural diversity. The findings present a novel approach for selecting structural diversity indicators for future investigation.

An Assessment of Global Dimming and Brightening during 1984–2018 Using the FORTH Radiative Transfer Model and ISCCP Satellite and MERRA-2 Reanalysis Data

In this study, an assessment of the FORTH radiative transfer model (RTM) surface solar radiation (SSR) as well as its interdecadal changes (Δ(SSR)), namely global dimming and brightening (GDB), is performed during the 35-year period of 1984–2018. Furthermore, a thorough evaluation of SSR and (Δ(SSR)) is conducted against high-quality reference surface measurements from 1193 Global Energy Balance Archive (GEBA) and 66 Baseline Surface Radiation Network (BSRN) stations. For the first time, the FORTH-RTM Δ(SSR) was evaluated over an extended period of 35 years and with a spatial resolution of 0.5° × 0.625°. The RTM uses state-of-the-art input products such as MERRA-2 and ISCCP-H and computes 35-year-long monthly SSR and GDB, which are compared to a comprehensive dataset of reference measurements from GEBA and BSRN. Overall, the FORTH-RTM deseasonalized SSR anomalies correlate satisfactorily with either GEBA (R equal to 0.72) or BSRN (R equal to 0.80). The percentage of agreement between the sign of computed GEBA and FORTH-RTM Δ(SSR) is equal to 63.5% and the corresponding percentage for FORTH-RTM and BSRN is 54.5%. The obtained results indicate that a considerable and statistically significant increase in SSR (Brightening) took place over Europe, Mexico, Brazil, Argentina, Central and NW African areas, and some parts of the tropical oceans from the early 1980s to the late 2010s. On the other hand, during the same 35-year period, a strong and statistically significant decrease in SSR (Dimming) occurred over the western Tropical Pacific, India, Australia, Southern East China, Northern South America, and some parts of oceans. A statistically significant dimming at the 95% confidence level, equal to −0.063 Wm−2 year−1 (or −2.22 Wm−2) from 1984 to 2018 is found over the entire globe, which was more prevalent over oceanic than over continental regions (−0.07 Wm−2 year−1 and −0.03 Wm−2 year−1, statistically significant dimming at the 95% confidence level, respectively) in both hemispheres. Yet, this overall 35-year dimming arose from alternating decadal-scale changes, consisting of dimming during 1984–1989, brightening in the 1990s, turning into dimming over 2000–2009, and brightening during 2010–2018.

Validation of a full-shift benzene exposure empirical model developed for work on offshore petroleum installations on the Norwegian continental shelf

Workers on offshore petroleum installations might be exposed to benzene, a carcinogenic agent. Recently, a full-shift benzene exposure model was developed based on personal measurements. This study aimed to validate this exposure model by using datasets not included in the model. The exposure model was validated against an internal dataset of measurements from offshore installations owned by the same company that provided data for the model, and an external dataset from installations owned by another company. We used Tobit regression to estimate GM (geometric mean) benzene exposure overall and for individual job groups. Bias, relative bias, precision, and correlation were estimated to evaluate the agreement between measured exposures and the levels predicted by the model. Overall, the model overestimated exposure when compared to the predicted exposure level to the internal dataset with a factor of 1.7, a relative bias of 73%, a precision of 0.6, a correlation coefficient of 0.72 (p = 0.019), while the Lin’s Concordance Correlation Coefficient (CCC) was 0.53. The model underestimated exposure when compared to the external dataset with a factor of about 2, with a relative bias of −45%, a precision of 1.2, a correlation coefficient of 0.31 (p = 0.544), and a Lin’s CCC of 0.25. The exposure model overestimated benzene exposure in the internal validation dataset, while the precision and the correlation between the measured and predicted exposure levels were high. Differences in measurement strategies could be one of the reasons for the discrepancy. The exposure model agreed less with the external dataset.

Open EDFA gain spectrum dataset and its applications in data-driven EDFA gain modeling

Optical networks satisfy high bandwidth and low latency requirements for telecommunication networks and data center interconnection. To improve network resource utilization, machine learning (ML) is used to accurately model optical amplifiers such as erbium-doped fiber amplifiers (EDFAs), which impact end-to-end system performance such as quality of transmission. However, a comprehensive measurement dataset is required for ML to accurately predict an EDFA’s wavelength-dependent gain. We present an open dataset consisting of 202,752 gain spectrum measurements collected from 16 commercial-grade reconfigurable optical add–drop multiplexer (ROADM) booster and pre-amplifier EDFAs under varying gain settings and diverse channel-loading configurations over 2,785 hours in total, with a total dataset size of 3.1 GB. With this EDFA dataset, we implemented component-level deep-neural-network-based EDFA models and use transfer learning (TL) to transfer the EDFA model among 16 ROADM EDFAs, which achieve less than 0.18/0.24 dB mean absolute error for booster/pre-amplifier gain prediction using only 0.5% of the full target training set. We also showed that TL reduces the EDFA data collection requirements on a new gain setting or a different type of EDFA on the same ROADM.

A Novel Response Factor-Based Method for In Situ Measurement of Wall Thermal Resistance

The heat flow meter method (HFM) is one of the most-used methods for the in situ measurement of wall thermal resistance. However, the standard HFM method has some issues: it is challenging to balance simplicity and accuracy in data analysis and the measurement period needs to be shorter. In this paper, a new dynamic data analysis method for the in situ measurement of wall thermal resistance is introduced, which is based on a truncated form of the infinite response factors for a wall heat conduction process and has a theoretically deducted convergence criteria for the automatic termination of an in situ measurement. The efficacy of the proposed method is validated by a theoretical analysis and by experiments from one simulation dataset and one measurement dataset. Preliminary experimental results show that the proposed method can reduce the measurement time by about one-third on average while maintaining the same accuracy as the standard average method. Due to the advantages of a clear physical meaning, a simple principle, and a short measurement period, the proposed method contributes to the quick and accurate estimation of the wall thermal resistance in buildings.

The bioelectrical impedance analysis (BIA) international database: aims, scope, and call for data.

Bioelectrical impedance analysis (BIA) is a technique widely used for estimating body composition and health-related parameters. The technology is relatively simple, quick, and non-invasive, and is currently used globally in diverse settings, including private clinicians' offices, sports and health clubs, and hospitals, and across a spectrum of age, body weight, and disease states. BIA parameters can be used to estimate body composition (fat, fat-free mass, total-body water and its compartments). Moreover, raw measurements including resistance, reactance, phase angle, and impedance vector length can also be used to track health-related markers, including hydration and malnutrition, and disease-prognostic, athletic and general health status. Body composition shows profound variability in association with age, sex, race and ethnicity, geographic ancestry, lifestyle, and health status. To advance understanding of this variability, we propose to develop a large and diverse multi-country dataset of BIA raw measures and derived body components. The aim of this paper is to describe the 'BIA International Database' project and encourage researchers to join the consortium. The Exercise and Health Laboratory of the Faculty of Human Kinetics, University of Lisbon has agreed to host the database using an online portal. At present, the database contains 277,922 measures from individuals ranging from 11 months to 102 years, along with additional data on these participants. The BIA International Database represents a key resource for research on body composition.

Timely estimation of leaf chlorophyll fluorescence parameters under varying light regimes by coupling light drivers to leaf traits.

Unveiling informative chlorophyll a fluorescence (ChlF) parameters and leaf morphological/biochemical traits under varying light conditions is important in ecological studies but has less been investigated. In this study, the trait-ChlF relationship and regressive estimation of ChlF parameters from leaf traits under varying light conditions were investigated using a dataset of synchronous measurements of ChlF parameters and leaf morphological/biochemical traits in Mangifera indica L. The results showed that the relationships between ChlF parameters and leaf traits varied across light intensities, as indicated by different slopes and intercepts, highlighting the limitations of using leaf traits alone to capture the dynamics of ChlF parameters. Light drivers, on the other hand, showed a better predictive ability for light-dependent ChlF parameters compared to leaf traits, with light intensity having a large effect on light-dependent ChlF parameters. Furthermore, the responses of ФF and NPQ to light drivers differed between leaf types, with light intensity having an effect on ФF in shaded leaves, whereas it had a primary effect on NPQ in sunlit leaves. These results facilitate and deepen our understanding of how the light environment affects leaf structure and function and, therefore, provide the theoretical basis for understanding plant ecological strategies in response to the light environment.

Volatility Parametrization of Low-Volatile Components of Ambient Organic Aerosols Based on Molecular Formulas.

Evaluating the volatility of organic compounds based solely on their molecular formulas would avoid tough demands in deriving molecular structures. Here, we deployed an iodide-adduct Long Time-of-Flight Chemical Ionization Mass Spectrometry (LToF-CIMS) combined with a Filter Inlet for Gases and AEROsols (FIGAERO) to investigate molecular formulas and thermograms of organic compounds on ambient particulate samples collected in the summer of 2021 in a suburban site of Shanghai and to estimate saturation vapor pressures of low- and semivolatile components of ambient organic aerosols. Then, a hierarchical cluster analysis and a subsequent classification of obtained clusters by similarity calculation were applied to the measured data set of molecular formulas and saturation vapor pressures of organic aerosols with at least a 2/3 appearance frequency, together with a similar data set collected at a rural site in the Beijing-Tianjin-Hebei region during the winter of 2018 (Ren et al., 2018), to classify all compounds into multiple groups. For each group of compounds, parametrizations between volatility and elemental composition were derived, and then relationships between each group of parameters and the mean O:C were established to achieve a volatility-molecular formula parametrization with the O:C as a key input. Statistical comparison of estimated volatilities of low-volatile organic compounds shows a much better performance of our parametrization than previous molecular formula-based ones.

Predicting Missing Value Data on IEC TC10 Datasets for Dissolved Gas Analysis using Tertius Algorithm

IEC TC10 is the most widely used Dissolved Gas Analysis (DGA) measurement dataset nowadays. Many DGA-based studies have been carried out using conventional methods and methods based on Artificial Intelligence Techniques (AITs). DGA is a diagnostic test performed on power transformers to detect and diagnose potential faults. The test involves analyzing the gases that are dissolved in the transformer oil, which can provide important information about the condition of the transformer. DGA is a widely used technique for transformer monitoring and maintenance in the power industry. However, this dataset is not perfect. There are still many problems in this dataset, one of which is the problem of missing value data. This problem will be significant if not appropriately handled. More reliable data from DGA measurement results is an in-dispensable reference in diagnosing faults in power transformers. This study focuses on dealing with the problem of missing value data using the Tertius algorithm, then testing the results using the J48 and Random Forest algorithms. The results obtained are pretty significant. Of the total 56 missing data, 36 could be predicted perfectly. And received the results of measuring accuracy using the J48 method of 62.73% and the Random Forest method of 70.71%. This result shows that the approach we applied is relatively good for handling missing values in IEC TC10 datasets.