Error Ratio Research Articles

Prescribed Performance Control for PEM Fuel Cell Air Supply System Based on Fully Actuated Approach With Fixed Regulation Time

ABSTRACTThe problem of air feed system control with guaranteed prescribed performance and fixed time control is investigated for proton exchange membrane fuel cell (PEMFC). First of all, the original model of PEMFC air feed system is converted to fully actuated system using higher order fully actuated system (HOFAS) approach. A fixed time extended state observer (ESO) is employed to approximate the cathode pressure. The parameter uncertainty is compensated by an adaptive radial basis function neural network (RBFNN). Then a prescribed performance sliding mode controller is designed with fixed convergence time. The proposed control law could guarantee the prescribed transient response for oxygen excess ratio (OER) error under stack current changing and bounded uncertainty. Lyapunov approach is utilized to proof fixed‐time stability of proposed controller. Simulation results of the study tests illustrate that the presented controller is efficient under constant OER tracking control, constant OER tracking control with parameter perturbation, and variable OER tracking control, respectively.

Prediction of Grease Performance and Optimal Additive Ratio Based on the SSA-GDA-LSSVM Model

In this paper, to address the issue of compounding three additives in PTFE grease, we propose a machine learning model based on SSA-GDA-LSSVM to predict both the tribological performance and the optimal ratio of additives in PTFE grease. Gaussian data augmentation expanded the experimental data, and the Sparrow Algorithm optimized hyperparameters of the Least Squares Support Vector Machine. SHAP analysis clarified model predictions, and a Non-Dominated Sorting Genetic Algorithm identified optimal additive ratios, which were experimentally validated. The results showed that the model predicted friction coefficients and wear scar widths with R² values exceeding 0.97, and the experimental error for optimal ratios was less than 1%.

Modeling and Simulation of Wide-Frequency Characteristics of Electromagnetic Standard Voltage Transformer

To achieve the broadband applicability of standard voltage transformers in “dual high” power systems, an equivalent circuit model of the standard voltage transformer is first established. Using the complex magnetic permeability method and utilizing existing core parameters, the excitation impedance values are obtained. Next, based on the equivalent circuit model, the no-load error function of the standard voltage transformer is analyzed, and through simulation, the no-load error response curve of the standard voltage transformer in the frequency range of 20 Hz to 3000 Hz is derived. The simulation results indicate that within the 20 Hz to 700 Hz range, both the no-load ratio error and the no-load angular error meet the accuracy requirements, with the ratio error within ±0.05% and the angular error within 2′. Additionally, the derivation of the error transfer function demonstrates the correlation between the no-load error values and the number of turns and cross-sectional area of the standard voltage transformer. Simulation results, obtained by increasing and decreasing the number of turns and cross-sectional area by 10%, provide valuable insights for the error compensation and structural design of standard voltage transformers.

WiTUnet: A U-shaped architecture integrating CNN and Transformer for improved feature alignment and local information fusion

Low-dose computed tomography (LDCT) has emerged as the preferred technology for diagnostic medical imaging due to the potential health risks associated with X-ray radiation and conventional computed tomography (CT) techniques. While LDCT utilizes a lower radiation dose compared to standard CT, it results in increased image noise, which can impair the accuracy of diagnoses. To mitigate this issue, advanced deep learning-based LDCT denoising algorithms have been developed. These primarily utilize Convolutional Neural Networks (CNNs) or Transformer Networks and often employ the Unet architecture, which enhances image detail by integrating feature maps from the encoder and decoder via skip connections. However, existing methods focus excessively on the optimization of the encoder and decoder structures while overlooking potential enhancements to the Unet architecture itself. This oversight can be problematic due to significant differences in feature map characteristics between the encoder and decoder, where simple fusion strategies may hinder effective image reconstruction. In this paper, we introduce WiTUnet, a novel LDCT image denoising method that utilizes nested, dense skip pathway in place of traditional skip connections to improve feature integration. Additionally, to address the high computational demands of conventional Transformers on large images, WiTUnet incorporates a windowed Transformer structure that processes images in smaller, non-overlapping segments, significantly reducing computational load. Moreover, our approach includes a Local Image Perception Enhancement (LiPe) module within both the encoder and decoder to replace the standard multi-layer perceptron (MLP) in Transformers, thereby improving the capture and representation of local image features. Through extensive experimental comparisons, WiTUnet has demonstrated superior performance over existing methods in critical metrics such as Peak Signal-to-Noise Ratio (PSNR), Structural Similarity (SSIM), and Root Mean Square Error (RMSE), significantly enhancing noise removal and image quality. The code is available on github https://github.com/woldier/WiTUNet.

A novel hybrid model for predicting the bearing capacity of piles

Due to the uncertainty of soil condition and pile design characteristics, it is always a challenge for geotechnical engineers to accurately determine the bearing capacity of piles. The main objective of this study is to propose a hybrid model coupling least squares support vector machine (LSSVM) with an improved particle swarm optimization (IPSO) algorithm for the prediction of bearing capacity of piles. The improved PSO algorithm was used to optimize the LSSVM hyperparameters. The performance of the IPSO-LSSVM model was compared with seven artificial intelligence models, namely adaptive neuro-fuzzy inference system (ANFIS), M5 model tree (M5MT), multivariate adaptive regression splines (MARS), gene expression programming (GEP), random forest (RF), regression tree (RT) and a stacked ensemble model. Six statistical indices (e.g., coefficient of determination (R2), mean absolute error (MAE), root mean squared error (RMSE), relative root mean squared error (RRMSE), BIAS and discrepancy ratio (DR)) were used to evaluate the performance of the models. The R2, MAE, RMSE, RRMSE and BIAS values of the IPSO-LSSVM model were 1, 4.27 kN, 6.164 kN, 0.005 and 0, respectively, for the training datasets and 0.9977, 22 kN, 36.03 kN, 0.0275 and –11, respectively, for the testing datasets. Compared with the ANFIS, MARS, GEP, M5MT, RF, RT and the stacked ensemble models, the proposed IPSO-LSSVM model shows high accuracy and robustness on the test datasets. In addition, the sensitivity, uncertainty, reliability and resilience of the IPSO-LSSVM model were also analyzed in this study. First published online 22 October 2024

Noise reduction method for mine wind speed sensor data based on CEEMDAN-wavelet threshold

The mine wind speed sensor is an important intelligent sensing equipment in the mine intelligent ventilation system that can provide accurate and key wind speed parameters for the intelligent ventilation system. The turbulent pulsation characteristics of the airflow in the underground tunnel are a major factor for the inaccurate measurement of mine wind speed. Therefore, according to the random non-stationary characteristics of a turbulent pulsation signal, a denoising method based on adaptive complete ensemble empirical mode decomposition (CEEMDAN) combined with the wavelet threshold is proposed for suppressing the turbulent pulsation noise in the wind speed signal. First, the CEEMDAN algorithm is used for decomposing the wind speed signal into a series of IMF components. Second, the continuous mean square error criterion is used for determining the high-frequency IMF components with more noise. The wavelet threshold denoising method is used for denoising the high-frequency IMF components with more noise. Finally, the denoised IMF components and remaining low-frequency IMF components are reconstructed for obtaining the denoised signal. The results of the denoising analysis of measured turbulent pulsation signals, comparative analysis of denoising of simulated turbulent pulsation signals by different joint denoising methods, and denoising analysis of actual mine wind speed sensor data indicate that the joint denoising method proposed in this study has a higher signal-to-noise ratio and lower root mean square error of the wind speed signal after denoising. Compared with the EMD-wavelet threshold and EEMD-wavelet threshold denoising methods, the denoising method proposed in this study is better and has higher denoising accuracy, which provides a new method for processing actual mine wind speed sensor data.

Research on Target Image Classification in Low-Light Night Vision.

In extremely dark conditions, low-light imaging may offer spectators a rich visual experience, which is important for both military and civic applications. However, the images taken in ultra-micro light environments usually have inherent defects such as extremely low brightness and contrast, a high noise level, and serious loss of scene details and colors, which leads to great challenges in the research of low-light image and object detection and classification. The low-light night vision image used as the study object in this work has an excessively dim overall picture and very little information about the screen's features. Three algorithms, HE, AHE, and CLAHE, were used to enhance and highlight the image. The effectiveness of these image enhancement methods is evaluated using metrics such as the peak signal-to-noise ratio and mean square error, and CLAHE was selected after comparison. The target image includes vehicles, people, license plates, and objects. The gray-level co-occurrence matrix (GLCM) was used to extract the texture features of the enhanced images, and the extracted image texture features were used as input to construct a backpropagation (BP) neural network classification model. Then, low-light image classification models were developed based on VGG16 and ResNet50 convolutional neural networks combined with low-light image enhancement algorithms. The experimental results show that the overall classification accuracy of the VGG16 convolutional neural network model is 92.1%. Compared with the BP and ResNet50 neural network models, the classification accuracy was increased by 4.5% and 2.3%, respectively, demonstrating its effectiveness in classifying low-light night vision targets.

Dynamical Embedding of Single-Channel Electroencephalogram for Artifact Subspace Reconstruction.

This study introduces a novel framework to apply the artifact subspace reconstruction (ASR) algorithm on single-channel electroencephalogram (EEG) data. ASR is known for its ability to remove artifacts like eye-blinks and movement but traditionally relies on multiple channels. Embedded ASR (E-ASR) addresses this by incorporating a dynamical embedding approach. In this method, an embedded matrix is created from single-channel EEG data using delay vectors, followed by ASR application and reconstruction of the cleaned signal. Data from four subjects with eyes open were collected using Fp1 and Fp2 electrodes via the CameraEEG android app. The E-ASR algorithm was evaluated using metrics like relative root mean square error (RRMSE), correlation coefficient (CC), and average power ratio. The number of eye-blinks with and without the E-ASR approach was also estimated. E-ASR achieved an RRMSE of 43.87% and had a CC of 0.91 on semi-simulated data and effectively reduced artifacts in real EEG data, with eye-blink counts validated against ground truth video data. This framework shows potential for smartphone-based EEG applications in natural environments with minimal electrodes.

The impact of camera-monitor system viewing angles on drivers’ distance perception: A simulated driving study

Camera-monitor systems (CMS) are increasingly used in driving. CMS separates the driver's sight line from the camera view, due to the lack of mirror reflection, only changing the camera's visual axis angle may affect the driver's rear view perception. While previous research has explored camera height and field of view, the effects of horizontal and vertical viewing angles alone remain unclear. This study aimed to evaluate how the horizontal viewing angle and vertical viewing angle of CMS camera affect distance estimation and car-following tasks. By changing the horizontal and vertical viewing angle, different self-vehicle references and horizon positions were formed in the image. Two experiments were conducted with the CMS around the steering wheel (Experiment 1) and at the bottom of the A-pillar (Experiment 2). Independent variables were the horizontal viewing angle (reference scale: 1/4, 1/3, 1/2) and vertical viewing angle (horizon position: 1/2, 1/3). Dependent variables included distance estimation error ratio and following distance. Experiment 1 demonstrated a significant interaction effect: a smaller reference scale and higher horizon position reduced distance underestimation. Additionally, a smaller reference scale for the participants' self-vehicle resulted in shorter following distances. In Experiment 2, the distance estimation outcomes on the left display aligned with those of Experiment 1; however, the influence of the viewing angle was diminished on the right display. The study suggests CMS design should balance vehicle reference inclusion with environmental cues, enhancing distance perception and driving safety. The consistency between CMS design and driver familiarity also needs to be considered.

Modelling the evolving critical state of crushable soils and parameter optimization with an improved genetic algorithm (AIS-RCGA)

The critical state line (CSL) plays an essential role in the constitutive modelling of granular soils. It serves as a reference line for the measurement of the state parameter. The critical state of crushable soils cannot accurately be determined from experimental data because of the ever-changing soil properties due to particle breakage. In this study, two new parameters eB and eb are introduced, which account for the final position and the evolution of CSL of crushable soils during shearing, respectively. To identify the optimal CSL-related parameters from experimental data, a hybrid genetic algorithm combining artificial immune system (AIS) and real-coded genetic algorithm (RCGA), namely AIS-RCGA is adopted. The fitness is defined to minimize the prediction error of both void ratio (or, pore water pressure) and deviatoric stress. We have refined the tuning methods for several hyperparameters of AIS-RCGA and proposed a novel method to assess the similarity of individuals within AIS-RCGA. Results show that the new method is more efficient in finding the global optimal for our problem. With optimized model parameters, the new constitutive model can accurately predict the response of crushable soil, outperforming other constitutive model reported in the literature.

Assessing the influence of climate change on multiple climate indices in Nepal using CMIP6 global climate models

Global climate models (GCMs) serve as essential tools for projecting future climate trends, but their coarse resolution limits localized impact assessments in sectors like hydrology, agriculture, and biodiversity. Observation data with a spatial resolution of a few kilometers are crucial for downscaling and bias-correcting GCMs at finer resolutions. However, Nepal's extreme topography and organizational challenges have led to uneven distribution of meteorological stations and inconsistent data quality. Moreover, CMIP6-based climate extremes projections for the entire country are currently unavailable. To tackle these challenges, we developed a comprehensive national database for Nepal, offering high-resolution historical and projected precipitation and temperature data analyzed through 25 climate extreme indices from the Expert Team on Climate Change Detection and Indices (ETCCDI). Initially, observation grid data were prepared at a daily timescale with a spatial resolution of 0.05° × 0.05° for baseline period (1981–2010) using the Asian Precipitation High-Resolved Observational Data Integration Toward Evaluation (APHRODITE), the fifth generation of the European Centre for Medium-Range Weather Forecasts Reanalysis (ERA5), and available good quality observed climate data. This data was then utilized to downscale and bias-correct 18 CMIP6 GCMs for 2015–2100 under four SSPs (1–2.6, 2–4.5, 3–7.0, 5–8.5). Quantile mapping was employed for the bias correction of the CMIP6 GCMs. The performance of the multimodal ensemble (MME) indicated better Nash-Sutcliffe Efficiency (NSE), root mean square error ratio (RSR), and Percent Bias (PBIAS) of climate extreme indices for the historical period. A comparative analysis was conducted across Nepal's major geographic regions to account for spatial variability in regional climate systems. The finer-resolution dataset can be crucial to deepen our understanding of climate impacts, and climate change, and eventually informing the policy-making in Nepal. Moreover, the methodology can be effectively replicated in data-scarce developing nations to promote climate research and adaptation efforts.

Adaptive spatial down-sampling method based on object occupancy distribution for video coding for machines

As the performance of machine vision continues to improve, it is being used in various industrial fields to analyze and generate massive amounts of video data. Although the demand for and consumption of video data by machines has increased significantly, video coding for machines needs to be improved. It is therefore necessary to consider a new codec that differs from conventional codecs based on the human visual system (HVS). Spatial down-sampling plays a critical role in video coding for machines because it reduces the volume of the video data to be processed while maintaining the shape of the data’s features that are important for the machine to reference when processing the video. An effective method of determining the intensity of spatial down-sampling as an efficient coding tool for machines is still in the early stages. Here, we propose a method of determining an optimal scale factor for spatial down-sampling by collecting and analyzing information on the number of objects and the ratio of the area occupied by the object within a picture. We compare the data reduction ratio to the machine accuracy error ratio (DRAER) to evaluate the performance of the proposed method. By applying the proposed method, the DRAER was found to be a maximum of 21.40 dB and a minimum of 11.94 dB. This shows that video coding gain for the machines could be achieved through the proposed method while maintaining the accuracy of machine vision tasks.

Implementation of Optical Coherent Communication System between Two Computers

This work represents implementation and investigation of optical coherent communication system between two computers. A single mode optical fiber is selected as transmission medium. The data are sent via the RS-232 standard interface with a bit rate of 9.6 kbps from personal computer (PC1) by line receive to convert the data from electrical levels (-12/+12 V) into TTL level (0/5 V). The modulation of this data was accomplished by internal modulation using laser diode type (HFCT-5208M) 1310 nm wavelength. The optical D-coupler was used to combine the optical signal that come from laser source with optical signal of laser local oscillator (OTS-304XI) at 1310/1550 nm wavelength to obtain coherent (homodyne and heterodyne) detection respectively. A PIN photodetector (HFCT-5208M) is used. Calculations of Signal to Noise Ratio (SNR) and Bit Error Rate (BER) for coherent detection were measured at different length of the optical fiber. Result show that high SNR and low BER for heterodyne detection than for homodyne detection.

Design and Stress Analysis of Bevel Line Gears with Vertical Flank Suitable for Micro Machining

The line gear (LG) exhibits promising applications in micromachinery owing to its simple structure and minimal teeth count. This paper aims to advance the meshing theory of LG and introduce a conical LG tooth configuration specifically tailored for micro-machining that is capable of driving with high performance. Based on the meshing principle of LG, a conical line gear pair with vertical flank (VFLG) of the driving gear was designed. Subsequently, adhering to the curvature non-interference principle that no local interference occurs in the meshing process and the specified range of fitting error in one direction, the design parameters for LG were determined. A comparison of the contact stress between two sets of different contact line types revealed the distinct advantage of VFLG with parameters identical to those of traditional LG. The experimental results conclusively demonstrate a transmission ratio error of 0.004, affirming the feasibility of the design. The curvature non-interference design formula proposed in this paper refines the LG design theory, and the novel LG design method presented holds significant implications for subsequent micromachining.

Uncertainty study of two-phase flow distribution characteristics measurement based on bubble trajectory tracking method

Bubbles undergo complex motions such as nonlinear trajectories in two-phase flow, where the position, velocity magnitude, and direction of bubble centers continuously change as they evolve under the influence of forces. The complexity of bubble trajectory increases the difficulty in measuring the flow field distribution characteristics, which has been understudied in the previous. The bubble trajectory tracking model based on the Monte Carlo method is constructed to study the uncertainty of two-phase flow distribution characteristics measurement by the conductivity probe. The Monte Carlo method is adopted to randomly generate the bubble velocity at the initial moment and the bubble force characteristics are embedded to simulate the three-dimensional motion of the bubble. The influence of various factors on the distribution of effective bubble numbers and velocity relative error at the local point is quantitatively analyzed. It is found that the lateral spacing of the sensors has the greatest influence on the measurement results of small-sized bubbles, the accuracy is greatly reduced when the space is larger than the radius of the measured bubbles. Moreover, the measurement accuracy of the probe for medium-sized bubbles is high, and the effect of bubble lateral velocity and aspect ratio on the effective bubble ratio and velocity relative error is negligible. In addition, the maximum measurement range of the probe at local points is investigated which is found to be 4 mm in the lateral direction and 2 mm in the longitudinal direction. Finally, uniformly distributed bubbles are generated in a 20 mm × 10 mm transverse plane and the distribution of the effective bubble ratio of the probe is found to be consistent at different measurement points, which verifies the reliability of the probe measurement.

Aerodynamic design of two-stage centrifugal compressor for vehicle hydrogen fuel cell

Abstract The on-board hydrogen fuel cell is a power generation device that converts the chemical energy of hydrogen and oxygen directly into electrical energy. The centrifugal compressor mainly provides oxygen for the cathode of the fuel cell system. With the rapid development of the hydrogen fuel cell industry, the key components of the air compressor field are also developing rapidly. In the technical route, most companies launched two-stage compression centrifugal air compressors in response to the increasing power of the product development trend. The newly developed hydrogen fuel cell system power has reached above 120kW to meet the needs of heavy trucks and other applications. Both the flow rate and pressure ratio of the air compressor are required to increase accordingly. Based on the reduction of parasitic power, the optimal efficiency at the working point of the two-stage centrifugal air compressor is taken as the design goal, from one-dimensional aerodynamic calculation to two-dimensional flow channel design. Finally, the feasibility and accuracy of the aerodynamic design are verified based on key parameters such as pressure ratio, flow rate, and efficiency of the final stage through CFD flow field simulation and test data comparison. The results show that the flow rate is consistent with the calculated flow field, and the maximum error of pressure ratio and efficiency is less than 2%, which meets the target requirements. The research results can provide a research and development basis for designing high-performance centrifugal air compressors for fuel cell vehicles.

Research on the calibration method of winding deformation tester

Abstract The focus of this paper is the calibration method and process for winding deformation tester, encompassing its working principle, metrological characteristics, calibration conditions and measurement methods. The calibration conditions are divided into environmental conditions and measurement standards and other equipments. The measurement methods are divided into indication error of scanning frequency and indication error of amplitude ratio. The key steps in the calibration process are elaborated, including the selection of standard equipments, error analysis, data processing. The method’s feasibility is verified through experiments.

Rapid and non-destructive determination of oil content in fuzzy cottonseeds via near-infrared reflectance spectroscopy

In China, enterprises specializing in cottonseed processing predominantly acquire fuzzy cottonseeds. However, the expeditious determination of oil content within these fuzzy cottonseeds remains a major challenge. This particular challenge considerably influences the economic efficiency of enterprises engaged in cottonseed processing. To address this concern, this research acquired near-infrared spectral (NIRS) information from fuzzy cottonseeds and applied the Standard Normal Variate (SNV) pretreatment to eliminate data noise. Subsequently, the Uninformative Variable Elimination (UVE) was employed to select feature wavelength points reflecting fuzzy cottonseed oil content. To further accurately extract the feature wavelengths of fuzzy cottonseeds, Competitive Adaptive Reweighted Sampling (CARS) and Random Frog (RF) were utilized for a second round of feature wavelength extraction. Finally, prediction models for fuzzy cottonseed oil content were established using partial least squares regression (PLSR), Support Vector Regression (SVR), and Least Squares Support Vector Machines (LSSVM). Results revealed that the LSSVM model constructed using the feature selected by the UVE-CARS had the best performance in predicting the oil content of fuzzy cottonseeds. The values of Coefficient of Determination (R2), Root Mean Square Error (RMSE), Residual Error Ratio (RER), and Residual Predictive Deviation (RPD) were 0.8605, 0.0075, 12.8116, and 3.8154, respectively. The LSSVM model built on the feature wavelength points selected by the UVE-RF demonstrated the second-best prediction performance with an R2 of 0.8329, an RMSE of 0.0091, an RER of 10.4861, and an RPD of 3.1228. The validation experiments corroborated these findings, it was found that the R2 was 0.8122, the RMSE was 0.0063, The RER was 7.6190, and the RPD was 2.9036. This underscores the effectiveness of the method employed in this study for detecting the oil content in fuzzy cottonseeds. The approach presented in this study serves as a crucial reference for the swift and accurate assessment of oil content in fuzzy cottonseeds, providing an impactful technical solution for cottonseed processing enterprises to enhance their economic outcomes.

Improving the steering performance of microrobots by using a magnetically actuated technique

Mobile microrobots show the significance of potential medical procedures and challenge areas within the human body. An applied electromagnetic field is used as a precise magnetic actuation system for controllable steering mobility, and positioning is used. In this work, an adaptable microrobot is developed with an effective technique to enhance steering. An electromagnetic system is employed to produce an external magnetic field that guides a microrobot to improve its steerability. In particular, a driving system of eight coils creates a magnetic field that actuates the microrobot and enables upward and downward motion. Accordingly, three types of microrobot designs are proposed. Consequently, an efficient variation of steering in terms of deformation angle is achieved. The corresponding performance is determined mathematically in terms of the main system parameters, such as external magnetic field and microrobot geometry. In addition, finite element model is used for comparison and validation. Simulation results show that for a relatively high value of magnet field, namely, 15 mT and 90°, the deformation angles were 36.98° for the first design, 39.11° for the second design, and 34.48° for the third design in upward deformation. By contrast, the downward deformation ranges were –40.42°, –39.46°, and –37.04° for the first, second, and third designs, respectively, at –90° The error between the upward and downward steering of the second design had the lowest value of 0.35° The second design of the microrobots has the least error between the upward and downward deformation ratio at 0.89 %. In comparison, the first design has a 9.3 % error, and the third design has a 7.42 % error, with a relatively high value of magnetic field density (15 mT). Simulations have shown that the proposed second design is suitable for navigating narrow artery branches with the least error ratio and more precise control. This work verifies that the model could accurately control the steering of microrobots.

Study on microwave ablation temperature prediction model based on grayscale ultrasound texture and machine learning.

Temperature prediction is crucial in the clinical ablation treatment of liver cancer, as it can be used to estimate the coagulation zone of microwave ablation. Experiments were conducted on 83 fresh ex vivo porcine liver tissues at two ablation powers of 15 W and 20 W. Ultrasound grayscale images and temperature data from multiple sampling points were collected. The machine learning method of random forests was used to train the selected texture features, obtaining temperature prediction models for sampling points and the entire ultrasound imaging area. The accuracy of the algorithm was assessed by measuring the area of the hyperechoic area in the porcine liver tissue cross-section and ultrasound grayscale images. The model exhibited a high degree of accuracy in temperature prediction and the identification of coagulation zone. Within the test sets for the 15 W and 20 W power groups, the average absolute error for temperature prediction was 1.14°C and 4.73°C, respectively. Notably, the model's accuracy in measuring the area of coagulation was higher than that of traditional ultrasonic grey-scale imaging, with error ratios of 0.402 and 0.182 for the respective power groups. Additionally, the model can filter out texture features with a high correlation to temperature, providing a certain degree of interpretability. The temperature prediction model proposed in this study can be applied to temperature monitoring and coagulation zone range assessment in microwave ablation.