Maximum Average Error Research Articles

A New Objective Typhoon Location Algorithm Considering a Perturbation Factor Based on FY-4A Brightness Temperature Data

Abstract In this paper, a new objective typhoon positioning algorithm was proposed. The algorithm uses L1 12-channel far-infrared data of the FY-4A geostationary meteorological satellite for objective positioning, verified against best path data provided by the Tropical Cyclone Data Center of the China Meteorological Administration (CMA). By calculating the tangential and radial perturbation values of infrared brightness temperature images, the perturbation factor can be obtained. By adopting the position of the maximum perturbation factor as the center of a circle and considering a radius of no more than 20 km, the position of the minimum perturbation factor was determined; this value represents the central position of the typhoon. Tropical cyclones in 2019 and 2020 were selected for objective positioning, and the objective positioning results were verified against the corresponding time in the best path dataset. The results included centralized verification results for 29 typhoons and optimal path data in 2019. The maximum average error reached 54.67 km, with an annual average typhoon positioning error of 16.15 km. The centralization verification results for 23 typhoons and optimal path data in 2020 indicated a minimum average error of 12.71 km, a maximum average error of 18.56 km, and an annual average typhoon positioning error of 14.82 km. The positioning results for these two years suggest that the longitude obtained with the perturbation factor positioning method is satisfactory, exhibiting an error of less than 20 km. Significance Statement The purpose of this study is to help researchers make scientific discoveries and help the development of typhoon center location technology in the future. This is important because accurate positioning of typhoon center can provide effective help for typhoon path prediction and typhoon intensity determination.

Research on Three Cylinder Synchronous Control Based on Tracking Differentiator and Fuzzy PID

Aiming at the problem of three-cylinder asynchronous caused by nonlinear factors in the three-cylinder synchronous control system, a fuzzy PID synchronous control method with a tracking differentiator based on an adjacent cross-coupling control strategy is proposed. Fuzzy PID control is used to adjust P, I, and D parameters online, and the tracking differentiator is used in fuzzy control to filter the input error and error differential of fuzzy control. After the controller design is completed, the combined simulation model of the three-cylinder control system is built by AMESim and Simulink. The simulation results verify that the performance of the fuzzy PID controller with tracking differentiator is smaller than the maximum synchronization and average error of the ordinary PID, which can be better applied to the three-cylinder synchronous control system.

The Study of Sound Speed as a Function of Pressure at Different Temperatures in Biofuel Component Liquids

In the present study, an approximation is applied to study the sound speed in liquids as a function of pressure at different temperatures. The relation obtained is applied in the case of biofuel component liquids. The calculated results for each liquid were found to be in good agreement with the experimental results throughout the range of pressure and temperature. The maximum percentage error and average percentage error are not more than 5.2 and 1.9, respectively, in the entire range of pressure and temperature for all liquids. Furthermore, the internal pressure and nonlinear Bayer's parameters are also computed as a function of temperature at one atmosphere from sound speed for the first time in biofuel component liquids.

Development and validation of thermal performances in a novel thermoelectric generator model for automotive waste heat recovery systems

Integrated thermoelectric generators (TEGs) and heat exchangers (HEX) can transform heat into electrical power by converting temperature gradients between heat sources and cold sinks. Predicting TEG thermal performances with compact HEX is crucial to designing efficient TEGs for automotive waste heat recovery. However, previous investigations mainly concentrated on TEG electrical performance, neglecting the thermal performances under high Reynolds numbers (Re) of exhaust gasses. This study utilizes the model-based development (MBD) method to develop a novel 1D TEG model, focusing heat transfer coefficient α and pressure drop ΔP inside a louvered corrugated fin HEX, component temperatures, and boundary heat flux, which affect electrical power from thermoelectric modules (TEM). The performance data are measured from 36 tests under different fin pitches Fp = 1.0 - 2.0 mm, Re = 4,000–14,000, and inlet gas conditions. Two methods are used to develop the model using user-defined functions (UDFs) for friction coefficient Cf and α, accounting for actual fin geometries. In Method 1, UDF multipliers for each Fp are required to calibrate the TEG model. In Method 2, simplified UDFs for Cf and α are implemented. Method 1 shows that the model is well-calibrated with maximum average relative errors δa< 6.7% but requires model fitting for each Fp. The fast-predictive TEG model with the novel calibration Method 2 can reproduce the thermal and electrical performances with δa< 20% without model tuning, which can be used for TEG redesign and integrated into vehicle-level MBD with a conventional and electrified powertrain, leveraging a fast driving-cycle simulation.

Numerical Investigation of Flow Characteristics for Gas–Liquid Two–Phase Flow in Coiled Tubing

Coiled tubing (CT) is widely used for horizontal well fracturing, squeeze cementing, and sand and solid washing in the oil and gas industry. During CT operation, a gas–liquid two-phase flow state appears in the tubing. Due to the secondary flow, this state produces a more extensive flow-friction pressure loss, which limits its application. It is crucial to understand the gas–liquid flow behavior in a spiral tube for frictional pressure drop predictions in the CT technique. In this study, we numerically investigated the velocity distribution and phase distribution of a gas–liquid flow in CT. A comparison of experimental data and simulated results show that the maximum average error is 2.14%, verifying the accuracy of the numerical model. The gas and liquid velocities decrease first and then rise along the axial direction due to the effect of gravity. Due to the difference in the gas and liquid viscosity, i.e., the flow resistance of the gas and liquid is different, the gas–liquid slip velocity ratio is always greater than 1. The liquid velocity exhibits a D-shaped step distribution at different cross-sections of spiral tubing. The secondary-flow intensity, caused by radial velocity, increases along the tubing. Due to the secondary-flow effect, the zone of the maximum cross-section velocity is off-center and closer to the outside of the tube. However, under the combined action of centrifugal force and the density difference between gas and liquid, the variation in the gas void fraction along the tubing is relatively stable. These research results are helpful in understanding the complex flow behavior of gas–liquid two-phase flow in CT.

Compactness prediction of asphalt concrete using Ground-Penetrating Radar: A comparative study

The compactness is an important indicator of quality control during the construction of the asphalt concrete pavement. The construction quality of asphalt concrete can be continuously, quickly, and non-destructively tested by establishing the correlation between dielectric constant and compactness through ground-penetrating radar. In this study, five compactness prediction models were established based on the relationship between volume parameters and the dielectric constant of each component of asphalt concrete. 18 samples of Sup20 asphalt concrete (Superior Performing Asphalt Pavement with the nominal aggregate size of 19 mm) were used to verify the established compactness prediction models. The results showed that the average relative error, maximum error, and root mean square error of the CP_Rayleigh model (compactness prediction model of Rayleigh.) were minimum, and thus the CP_Rayleigh model was selected for the following-up research in this study. The shapes of scatterers (aggregates and pores) were not spherical, and a shape factor u of 0.033 was introduced to modify the CP_Rayleigh model. The prediction accuracy of compactness can be effectively improved by the modified CP_Rayleigh model. Combined with field engineering verification, it was shown that the modified CP_Rayleigh model can be used to predict the compactness of asphalt concrete in the field.

An Analytical Solution to Predict the Distribution of Streamwise Flow Velocity in an Ecological River with Submerged Vegetation

Aquatic submerged vegetation is widespread in rivers. The transverse distribution of flow velocity in rivers is altered because of the vegetation. Based on the vegetation coverage, the cross-section of the ecological channels can be divided into the non-vegetated area and the vegetated area. In the vegetated area, we defined two depth-averaged velocities, which included the water depth-averaged velocity, and the vegetation height-averaged velocity. In this study, we optimized the ratio of these two depth-averaged velocities, and used this velocity ratio in the Navier–Stokes equation to predict the lateral distribution of longitudinal velocity in the open channel that was partially covered by submerged vegetation. Based on the Navier–Stokes equations, the term “vegetation resistance” was introduced in the vegetated area. The equations for the transverse eddy viscosity coefficient ξ, friction coefficient f, drag force coefficient Cd, and porosity α were used for both the non-vegetated area and the vegetated area, and the range of the depth-averaged secondary flow coefficient was investigated. An analytical solution for predicting the transverse distribution of the water depth-averaged streamwise velocity was obtained in channels that were partially covered by submerged vegetation, which was experimentally verified in previous studies. Additionally, the improved ratio proposed here was compared to previous ratios from other studies. Our findings showed that the ratio in this study could perform velocity prediction more effectively in the partially covered vegetated channel, with a maximum average relative error of 4.77%. The improved ratio model reduced the number of parameters, which introduced the diameter of the vegetation, the amount of vegetation per unit area, and the flow depth. This theoretical ratio lays the foundation for analyzing the flow structure of submerged vegetation.

Automatic Measurement of External Thread at the End of Sucker Rod Based on Machine Vision.

Aiming at the low efficiency of manual measurement of threads and the lack of practicability in machine vision measurement before, online size measurement of threads at the end of sucker rods based on machine vision was studied. A robotic arm is used to carry an optical device to achieve high-quality image acquisition of threads. Based on the prior knowledge of the thread profile angle, the directional edge detection operator is customized to achieve the accurate detection of the left and right edges of the thread. Noise filtering, sorting, and left and right edge-matching algorithms based on connected domains are developed to eliminate the interference effects of electrostatic dust and oil pollution in online measurement, and the dimension of thread profile angles, pitches, major diameters, and minor diameters can be precisely calculated. The experimental results show that the screw thread parameter measurement time is about 0.13 s; the maximum and minimum average errors of the thread angles are 0.011° and 0.632°, respectively; and the total average deviation is less than 0.08°. For the screw thread pitch, major diameter, minor diameter, and pitch diameter parameter measurement, the deviation of the measurement results between the proposed method and the universal tool microscope (UTM) method is less than 10 μm. It fully proves the effectiveness and accuracy of the method in this paper and, at the same time, shows that the method has good real-time performance and high application significance, which lays a good foundation for the subsequent online thread measurement.

Computational procedure for simulating stall process of reciprocating compressors

This article presents in detail a computational procedure for simulating the stall process of reciprocating compressor. The focus of the work is the calculation of the compressor stall limits from the available torque curves of electrical motor. The procedure is composed by an algorithm to compute the compressor stall using a developed compressor simulation model. The compressor model considers the simulation of flow and mechanical interactions in the compressor components and was implemented in the GT-Power platform, a system simulation software. The compressor simulation results for steady state operation were compared with experimental data, indicating that the simulation model represented very well the overall compressor behavior, with a maximum relative error of 1.0% for the compressor global performance. The results for the compressor stall obtained by the proposed computational procedure agreed with the experimental stall tests, with a maximum average relative error of about 6.7%. The computed stall curve had similar trends in relation to experimental results. Performed parametric analyses considering statistical rank correlations indicated that the electric motor tension and motor temperature were the most influencing parameters on compressor stall. The obtained results show that the developed procedure could be an efficient tool for compressor design.

Deep-neural-network-based angles-only relative orbit determination for space non-cooperative target

Angles-only relative orbit determination for space situational awareness is subjected to a well-known range observability problem, that is, the range between two spacecraft is weakly observable or unobservable. However, in previous studies, the solutions to this problem using nonlinear relative motion dynamics have been limited by the computational complexity and long arc of observation. To this end, this study develops a simple and fast algorithm to address the angles-only relative orbit determination problem by exploiting a deep neural network (DNN), which has a significantly strong capability of capturing nonlinearity. Emphasis is placed on the construction of a nonlinear mapping model from the line-of-sight angles to the relative orbit state by training the designed DNN. First, a training dataset generator including nonlinear relative motion dynamics and a line-of-sight measurement model is established to generate the training data for the DNN. Second, the DNN frame, including the network structure, data processing, and network training algorithm for angles-only relative orbit determination, is designed. Subsequently, a digital simulation system comprising error models, reference missions and trajectories, and computation models for error estimation is established. Thus, the anti-noise and generalization performance of the nonlinear mapping model on GEO-type orbits are verified using digital simulations. The results indicate that the proposed algorithm is effective, whereby the estimation accuracy for the relative position is generally better than that for the relative velocity. In the case of a co-elliptical orbit, the estimated errors of distance and velocity in each direction are less than 9.7% and 48.2%, respectively, whereas the maximum average errors are approximately 1.1% and 4.2%, respectively, where only three sets of angle measurements with an interval of 600 s are available. However, in the case of a non-coplanar orbit, the interval between angles can be decreased to 50 s, when the corresponding estimated error of distance in each direction is less than 9.9%, and the maximum average error is approximately 1.1%. Additionally, the sensitivity of the proposed algorithm to the arc length, number, and interval of the measurements is analyzed.

Real-Time Measurement of Moisture Content of Paddy Rice Based on Microstrip Microwave Sensor Assisted by Machine Learning Strategies

Moisture content is extremely imoprtant to the processes of storage, packaging, and transportation of grains. In this study, a portable moisture measuring device was developed based on microwave microstrip sensors. The device is composed of three parts: a microwave circuit module, a real-time measurement module, and software to display the results. This work proposes an improvement measure by optimizing the thickness of paddy rice samples (8–13 cm) and adding the ambient temperatures and the moisture contents (13.66–27.02% w.b.) at a 3.00 GHz frequency. A random forest, decision tree, k-nearest neighbor, and support vector machine were applied to predict the moisture content in the paddy rice. Microwave characteristics, phase shift, and temperature compensation were selected as the input variables to the prediction models, which have achieved high accuracy. Among those prediction models, the random forest model yielded the best performance with highest accuracy and stability (R2 = 0.99, RMSE = 0.28, MAE = 0.26). The device showed a relatively stable performance (the maximum average absolute error was 0.55%, the minimum absolute error was 0.17%, the mean standard deviation was 0.18%, the maximum standard deviation was 0.41%, and the minimum standard deviation was 0.08%) within the moisture content range of 13–30%. The instrument has the advantages of real-time, simple structure, convenient operation, low cost, and portability. This work is expected to provide an important reference for the real-time in situ measurement of agricultural products, and to be of great significance for the development of intelligent agricultural equipment.

An improved mathematical model bidirectional coupling of heat-water and mechanics during vacuum pre-cooling

Vacuum pre-cooling is a multi-physics process coupling heat transfer, water transport, and shrinkage, the stresses caused by water evaporation and temperature variation in food interact with its deformation. However, the developed models usually assume that the shrinkage of food is insignificant so that it is ignored, which is inconsistent with the actual vacuum pre-cooling process and makes the simulation results inaccurate. Taking strawberry as an example, a multiphase porous media model with heat-water and mechanics (HWM) bi-directional coupling was developed for food vacuum pre-cooling. The accuracy of the HWM model was verified by comparing the heat-water (HW) model without considering shrinkage and the experiments. The results show that the development of the evaporation front predicted by the HWM model is more sensitive to the vacuum chamber pressure change compared with the HW model. The maximum average relative errors corresponding to the temperature and shrinkage of the HWM model are 0.27% and 6.95%, respectively, which are better than the HW model. In addition, the edge gradient effect that occurs in the middle stage of vacuum pre-cooling was observed. When the vacuum chamber pressure approached the lowest point, the evaporation region rapidly developed toward the core. The increase of evaporation driving force and the decrease of boundary moisture caused drastic temperature and shrinkage changes on the product surface. The model developed in this paper considering shrinkage provides technical support to optimize the vacuum pre-cooling process and improve product quality.

Machine learning based swift online capacity prediction of lithium-ion battery through whole cycle life

The machine learning (ML) based methods show great promise for the online battery capacity prediction. However, the feasibility and simplicity of indispensable feature extraction are both challenging to existing ML based methods, which restrict the implementation in electric vehicle usages. To reduce the computational burden and simultaneously enhance the accuracy of prediction results, this paper proposes a ML based approach for swift capacity prediction leveraging fractional charging voltage segments. A total number of 21 voltage feature segments (VFSs) are intercepted with different voltage ranges to analyze the influence of voltage intervals on the capacity prediction. Meanwhile, three advanced ML models including random forest regression (RFR), relevance vector machine (RVM) and Gaussian process regression (GPR) are meticulously designed to build a mapping function between the intercepted VFSs and battery capacity. The battery capacity prediction can be attained subsequently based on the established matched relationship using the VFS captured in real-time. The experimental and contrastive results show that the best model can accurately predict battery capacity through whole cycle life with the maximum average relative error of only 1.95%. This work emphasizes the application potential of combining straightforward and reliable feature with ML algorithms for online battery capacity prediction.

Automatic registration of urban high-resolution remote sensing images based on characteristic spatial objects

Automatic registration of high-resolution remote sensing images (HRRSIs) has always been a severe challenge due to the local deformation caused by different shooting angles and illumination conditions. A new method of characteristic spatial objects (CSOs) extraction and matching is proposed to deal with this difficulty. Firstly, the Mask R-CNN model is utilized to extract the CSOs and their positioning points on the images automatically. Then, an encoding method is provided to encode each object with its nearest adjacent 28 objects according to the object category, relative distance, and relative direction. Furthermore, a code matching algorithm is applied to search the most similar object pairs. Finally, the object pairs need to be filtered by position matching to construct the final control points for automatic image registration. The experimental results demonstrate that the registration success rate of the proposed method reaches 88.6% within a maximum average error of 15 pixels, which is 28.6% higher than that of conventional optimization method based on local feature points. It is reasonable to believe that it has made a beneficial contribution to the automatic registration of HRRSIs more accurately and efficiently.

State of Health and Lifetime Prediction of Lithium-ion Batteries Using Self-learning Incremental Models

Lithium-ion batteries are key energy storage elements in the context of environmental-aware energy systems representing a crucial technology to achieve the goal of zero carbon emission. Therefore, its conditions must be monitored to guarantee the safe and reliable operation of the systems that use these components. Furthermore, lithium-ion batteries’ prognostics and health management policies must cope with the nonlinear and time-varying nature of the complex electrochemical dynamics of battery degradation. This paper proposes an incremental-learning-based algorithm to estimate the State of Health (SoH) and the Remaining Useful Life (RUL) of lithium-ion batteries based on measurement data streams. For this purpose, a two-layer framework is proposed based on incremental modeling of the SoH. In the first layer, a set of representative features are extracted from voltage and current data of partial charging and discharging cycles; these features are then used to train the proposed model in a recursive procedure to estimate the battery’s SoH. The second layer uses the capacity data for incremental learning of an Autoregressive (AR) model for the SoH, which will be used to propagate the battery’s degradation through time to make the RUL prediction. The proposed method was applied to two datasets for experimental evaluation, one from CALCE and another from NASA. The proposed framework was able to estimate the SoH of 8 different lithium-ion cells with an average percentage error below 1.5% for all scenarios, while the lifetime model predicted the cell’s RUL with a maximum average error of 25%.

Combining reinforcement learning and conventional control to improve automatic guided vehicles tracking of complex trajectories

AbstractWith the rapid growth of logistics transportation in the framework of Industry 4.0, automated guided vehicle (AGV) technologies have developed speedily. These systems present two coupled control problems: the control of the longitudinal velocity, essential to ensure the application requirements such as throughput and tag time, and the trajectory tracking control, necessary to ensure the proper accuracy in loading and unloading manoeuvres. When the paths are very short or have abrupt changes, the kinematic constraints play a restrictive role, and the tracking control becomes more challenging. In this case, advanced control strategies such as those based on intelligent techniques, including machine learning (ML) can be useful. Hence, in this work, we present an intelligent hybrid control scheme that combines reinforcement learning‐based control (RLC) with conventional PI regulators to face both control problems simultaneously. On the one hand, PIs are used to control the speed of each wheel. On the other hand, the input reference of these regulators is calculated by the RLC in order to reduce the guiding error of the path tracking and to maintain the longitudinal speed. The latter is compared with a PID path following controller. The PID regulators have been tuned by genetic algorithms. The RLC allows the vehicle to learn how to improve the trajectory tracking in an adaptive way and thus, the AGV can face disturbances or unknown physical system parameters that may change due to friction and degradation of AGV mechanical components. Extensive simulation experiments of the proposed intelligent control strategy on a hybrid tricycle and differential AGV model, that considers the kinematics and the dynamics of the vehicle, prove the efficiency of the approach when following different demanding trajectories. The performance of the RL tracking controller in comparison with the optimized PID gives errors around 70% smaller, and the average maximum error is also 48% lower.

Experimental and modeling investigations of hydrate phase equilibria in natural clayey-silty sediments

Rapid and accurate acquisition of hydrate phase equilibria in sediments is crucial for gas hydrate resource assessment and exploitation and hydrate-based technological applications. However, current models have significant limitations and deficiencies in predicting hydrate phase equilibria in sediments, especially clayey-silty sediments prevalent in nature. To this end, a unified hydrate phase equilibrium model was developed, which considers the combined inhibitions of surface adsorption, capillary effect, and salt solution. It is derived that water content and salinity can be used as direct input parameters to the model to estimate the hydrate dissociation temperature depression. Using the stepwise heating method, the hydrate dissociation conditions in various natural sediments were measured. It suggests that montmorillonite has the greatest dissociation temperature depression, followed by SA sediments and kaolin, while pore hydrates in silts exhibit almost the same thermodynamic equilibrium conditions as bulk hydrates unless the water content is below 15%. Special attention was also paid to the additional inhibition of surface adsorption on hydrate phase equilibria in sediments with salt solutions. This inhibition was found to be quite pronounced in clay-rich sediments, especially expansive clays, but essentially negligible in silty or sandy sediments. The maximum average error of 3.29% was observed between model predictions and measured data. Additionally, the amounts of clathrated water (CW), non-equilibrium unclathrated water (NEUW), and equilibrium unclathrated water (EUW) in sediments after the end of hydrate formation were determined using the model. It reveals that CW has the highest proportion in silts, reaching 68.9%, while only 0.231 in montmorillonite; in contrast, EUW is dominant in montmorillonite; NEUW seems to be independent of lithology but shows a strong dependence on the water content. It implies that silty sediments are more conductive to hydrate-based gas storage and CO2 sequestration than clayey sediments. This study provides some practical insights and implications for hydrate phase behaviors in natural sediments and can potentially be employed as a tool for phase equilibrium prediction with universality and accuracy.

An Online SOC-SOTD Joint Estimation Algorithm for Pouch Li-Ion Batteries Based on Spatio-Temporal Coupling Correction Method

An SOC (state of charge)-SOTD (state of temperature distribution) joint estimation algorithm is established for a pouch lithium-ion battery. This method integrates the first-order RC model, distributed heat generation model, thermal resistance network model, and spatio-temporal coupling correction based on spatial restoration algorithm and dual Kalman filter (DKF). Unlike the traditional 1-D SOT estimation model, the proposed algorithm can accurately estimate the temperature distribution in the battery online by restoring the battery surface temperature distribution, calculating the battery internal temperature distribution, and jointly correcting the SOC and average internal temperature. Then, the proposed method is used for the SOC-SOTD estimation of a 25-Ah pouch battery at the ambient temperatures of 10, 20, and 30 °C under the worldwide light-duty test cycle and new european driving cycle driving cycles, and the estimated values are verified by experiment and 3-D simulation. According to the verification results, calculation errors of SOC are no more than 2.12%, and the maximum average calculation errors are 0.16 °C for the surface temperature and 0.24 °C for the internal temperature. However, the DKF is needed for the SOC-SOTD joint estimation because the single KF can only correct SOC or average internal temperature but cannot handle the inconsistency in temperature distribution.

Numerical Analysis of Fatigue Life and strength of AA5052 Aluminum Alloy Reinforced with ZrO2, TiO2 and Al2O3 Nanoparticles

In this study, the finite element method using ANSYS workbench 16.1 has been successfully used to predict the fatigue life, fatigue strength, and the factors of safety for the as cast AA5052 as arrow matrix and its composites: AA5052/7 wt% ZrO2, AA5052/7 wt% TiO2 and AA5052/7 wt% Al2O3. The Finite Element Analysis (FEA) model was building according to dimensions of the experimental fatigue specimen. The total number of elements was 504 elements with a total number of nodes of 2572 nodes. The numerical fatigue test was processed under static structural analysis, and it has been analyzed using fatigue tool on ANSYS, Goodman theory was used for the prediction of life. The FEM using ANSYS.16.1 workbench simulation showed a good agreement with the experimental results for all the stress life curves and the highest difference in fatigue life was 17% and the lowest was 1.4%, for ZrO2 composite, while the maximum overall average error was 2.031% for AA5052 and the lowest was 0.378%, for Al2O3 composite. The maximum difference about 4.14 % between the experimental and numerical fatigue strength at 107 cycles for the row matrix and less than for the composites, also the minimum factor of safety for the AA5052 alloy is 0.8327 and for AA5052/7 wt% ZrO2 composite is 1.0709 while for both AA5052/7wt%TiO2 and AA5052/7 wt% Al2O3 composites are 1.0707 at specific design life.

Large-Scale Surface Water Mapping Based on Landsat and Sentinel-1 Images

Surface water is a highly dynamical object on the earth’s surface. At present, satellite remote sensing is the most effective way to accurately depict the temporal and spatial variation characteristics of surface water on a large scale. In this study, a region-adaptive random forest algorithm is designed on the Google Earth Engine (GEE) for automatic surface water mapping by using data from multi-sensors such as Landsat 7 ETM+, Landsat 8 OLI, and Sentinel-1 SAR images as source data, and China as a case study region. The visual comparison of the mapping results with the original images under different landform areas shows that the extracted water body boundary is consistent with the water range in the image. The cross-validation with the JRC GSW validation samples shows a very high precision that the average producer’s accuracy and average user’s accuracy of water is 0.933 and 0.998, respectively. The average overall accuracy and average kappa is 0.966 and 0.931, respectively. The independent verification results of lakes with different areas also prove the high accuracy for our method, with a maximum average error of 3.299%. These results show that the method is an ideal way for large-scale surface water mapping with a high spatial–temporal resolution.