Established Regression Model Research Articles

Optimization of design parameters and operation conditions of solar-air source heat pump coupled system for rural buildings in cold and severe cold regions

Integrating solar energy utilization into air source heat pump heating systems can effectively cut down on energy consumption. However, the complexity of coupled systems poses a challenge to system performance optimization. In this paper, a solar-air source heat pump coupled system designed for heating in cold (Beijing) and severe cold (Changchun) regions is developed and analyzed by TRNSYS software. Response surface methodology (RSM) is employed to establish regression models for different performance indicators, and then the non-dominated sorting genetic algorithm-II (NSGA-II) is adopted to solve the multi-objective optimization of interactive performance parameters through Pareto optimal solution set. The regression models of the performance indicators, i.e., the energy consumption (P), the solar energy assurance rate (Ar) and the ratio of heating supply from solar energy storage tank to building load (Hs), are found to be closely related to the area of solar collectors (A) and the volumes of storage tanks (Vl for solar collector loop and Vs for ASHP loop). The obtained Pareto set successfully balances the optimal values of interactive performance indicators. In cold regions, P, Ar, Hs of the coupled system can be improved respectively by 31.79%, 33.00% and 40.07% compared to the single air source heat pump system. While in severe cold regions, these improvements are 16.25%, 15.35% and 28.38%, respectively. The design method and optimization procedure of this study provide a basis for the optimization of the solar and auxiliary heat source coupled heating systems.

Research on rock strength prediction model based on machine learning algorithm

The compressive strength of rocks is one of its mechanical characteristics. It has been a difficult problem to predict rock compressive strength conveniently and efficiently, and to solve the limitations of traditional rock compressive strength tests such as high cost, long time consumption, and reliability assurance. In this study, a data set containing 1774 groups of rock compressive strength test data was constructed through file retrieval, including 9 input parameters: rock type, temperature, confining pressure, dimension of specimen, shape of specimen, and experimental method. Eight supervised learning algorithms were used to learn the rock compressive strength test data, and eight rock compressive strength prediction models considering multiple factors were established to obtain a better method of predicting rock compressive strength. By selecting different features, the optimal feature combination for predicting rock compressive strength was obtained, and the optimal parameters for different models were obtained through the Sparrow Search Algorithm (SSA). Finally, four regression evaluation indicators, including mean absolute error (MAE), root mean square error (RMSE), mean absolute percentage error (MAPE), and coefficient of determination (R2), were used to evaluate the predictive performance of the established regression models. The results showed that the best-trained model had a MAPE as low as 3.61%, MAE as low as 9.19 MPa, and R2 as high as 0.995. It is noteworthy that AdaBoost was found to be the best model for predicting rock compressive strength. This study presents a significant advancement in the field by demonstrating the effectiveness of machine learning algorithms in this context, which have not been extensively applied to rock compressive strength predictions. The findings suggest that these models can offer substantial improvements over traditional methods, not only in accuracy but also in operational efficiency. This research is important for geotechnical engineering, as accurate rock strength predictions are critical for the design and stability assessments of construction projects, ultimately contributing to safer and more cost-effective engineering solutions.

Prediction of electrochemical properties of La–Y–Ni-based hydrogen storage alloys based on machine learning

La–Y–Ni-based hydrogen storage alloys, renowned for their high hydrogen capacity and chemical stability, show significant developmental potential. However, their development is heavily reliant on extensive and costly experimental work. This study established regression models to predict three key properties—electrochemical capacity (0.2C), cyclic stability (80%Age), and high-rate discharge performance (1C)—using Random Forest (RF), Extreme Gradient Boosting (XGB), and Ridge regression (Ridge) algorithms. The RF algorithm outperformed the others, with test set R2 values exceeding 0.8 for all properties. Using SHapley Additive exPlanations (SHAP) for model interpretation, this study quantitatively analyzed the optimal models. Under the constraint of the total mass ratio of elements on the A side and B side being fixed at 100, this study analyzed the optimal Mn + Al intervals for replacing Ni on the B side and the optimal Y and Y + Ce intervals for replacing La on the A side.

Mathematical statistical methods for stroke prognosis prediction and their clinical application research

Abstract. Stroke is a serious illness, with a global disability rate of over 50% and a mortality rate of up to 30%, making research on stroke prognosis prediction of significant societal importance. This paper comprehensively analyzes the application of mathematical statistical methods in stroke prognosis prediction, aiming to explore how these methods can enhance the accuracy of prognosis predictions, thereby providing patients with personalized treatment plans and improving their long-term rehabilitation process. Initially, the article introduces the severity of stroke and the importance of prognosis prediction, outlining the diversified development trends in current stroke prognosis prediction research. Subsequently, the article detailedly summarizes 11 statistical methods commonly used in stroke prognosis prediction, dividing these methods into three categories: methods suitable for analysis at the initial stage of treatment, methods suitable for data processing during the mid-study phase, and methods for integrating all data to establish regression models. Through specific case studies, this paper demonstrates the application of these statistical methods in actual research, including the use of descriptive statistics in MRI image analysis, the application of T-tests and ANOVA in comparing different treatment effects, and the importance of regression analysis in establishing prognosis models, including linear regression, logistic regression, and multiple regression analysis when considering multiple independent variables. This research not only provides a precise method for predicting the prognosis of stroke patients but also offers theoretical support for medical teams to formulate personalized treatment plans, enabling researchers to more accurately predict the prognosis of stroke patients, providing more personalized and effective treatment options. This contributes to reducing the risks during the patients rehabilitation process and improving the quality of life.

Analysis and optimization of abrasive waterjet dressing parameters for surface texturing of diamond grinding wheels

With the development of ultra-precision machining technology, the grinding performance of diamond grinding wheels has put forward higher requirements. In this study, a new method of dressing diamond grinding wheels using abrasive waterjet (AWJ) technology is proposed to address the issues of workpiece damage and wheel clogging that occur when grinding difficult-to-machine materials with conventional diamond grinding wheels. The main process parameters were determined based on the theoretical model for dressing diamond grinding wheels using AWJ. The response surface methodology (RSM) and the backpropagation artificial neural network (BP-ANN) were employed to establish regression models between process parameters and microgroove characteristics. A comparison was performed to evaluate the prediction performance of both RSM and BP-ANN. The results indicated that both approaches BP-ANN and RSM are powerful tools for predicting microgroove characteristics. This work provides theoretical guidance for texturing diamond wheels surface.

Real-Time Dialysis Dose: Ionic Dialysis Versus Classical Urea Kinetic Modeling Indices.

Introduction Worldwide, over 850 million people need renal replacement therapy, and over 80% of them are on hemodialysis. The term "dialysis adequacy" is most often associated with the achievement of minimally acceptable indices - single pool Kt/V (spKt/V) and urea reduction ratio (URR%) and largely does not take into account other clinical indicators in patients. In addition, it should be taken into account that in the conditions of clinical practice, their measurement is carried out according to standards controlled by the regulatory structures and is carried out with a frequency between one and three months, and not during each dialysis procedure. Accordingly, based on the obtained value, we assume that it refers to the urea clearance of the entire dialysis prescription. A new methodology for measuring dialysis adequacy through ionic dialysis allows dose estimation in real time without the need for additional blood testsby recording the difference in sodium ion conductivity at the dialyzer inlet and outlet. The latter corresponds to the dialysis dose delivered and can be measured at each dialysis session, enabling timely therapeutic intervention at minimal cost to the community. The aim of this original article is to present ion dialysis to the community as a reliable and inexpensive alternative to classical urea kinetic modeling (UKM), evaluating the comparability of the two methods and the possible sources of error in the result. Material and methods This was a retrospective study of 32 patients undergoing hemodialysis at the Clinic of Nephrology and Dialysis at the University Hospital St. Marina (Varna, Bulgaria) for the period between January and December 2020.For each of the patients, four measurements (every three months) of the delivered dialysis dose were performed by ion dialysis and classical urea kinetic modeling in order to assess the comparability and reliability of the two methods. Results The analysis of the results proved a high correlation between the validated indicators of dialysis adequacy (spKt/V; URR) and those registered with online monitoring (online clearance monitor) by ionic dialysis - online Kt/V (onKt/V)while at the same time reporting a significant difference in the two methods - primarily based on the anthropometric formulas used to estimate the volume. The established regression models confirmed the high predictive value of ionic dialysis in relation to the actually delivered dose. Discussion Despite the high bond strength, our recorded values ​​for onKt/V are 8% lower compared to results using UKM. Therefore, onKt/V as an assessment metric has the ability to underestimate the dialysis dose received. Various factors as a possible source of error and its cause have been reported by some authors. Several studies have reported differences of 2-5% in instantaneous conductance measurements, which they associate mainly with differences in the diffusion coefficients of urea and sodium, as well as different effects of dialysis membrane charge or inadequate ultrafiltration correction. Conclusions Our study confirms that online clearance monitoring (OCM) is a practical non-invasive tool for daily use that complements the classic performance of OCM by helping to deliver an adequate dialysis dose with increased patient benefit and minimal cost of financial resources.

Can Digital Transformation Drive Green Innovation in China’s Construction Industry under a Dual-Carbon Vision?

Against the backdrop of increasing global environmental pollution and energy consumption, green innovation is necessary to achieve green transformation. As an industry with a huge demand for resources and energy consumption, the construction industry shoulders the mission of the times to promote green innovation to enhance the ability of sustainable development. Digital technology provides new opportunities for green innovation in the construction industry. However, the impacts and mechanisms of digital transformation driving green innovation have not been thoroughly studied. In this paper, 121 listed companies in China’s construction industry are selected as a sample from 2011 to 2021, and a total of 1331 annual observations are obtained, and the impact and mechanism of digital transformation on construction enterprises’ green innovation are empirically analyzed by establishing regression models. The study indicates that digital transformation can facilitate green innovation in construction companies by enhancing corporate risk-taking and improving corporate governance. Compared with non-state-owned enterprises, state-owned enterprises have more endogenous incentives for green transformation based on multiple pressures, which to some extent weakens the driving role of digital transformation. The driving effect of enterprises’ digital transformation is more significant when the intensity of regional environmental regulation is high. This paper examines how the digitization of construction enterprises can lead to new greening ideas from the perspective of green innovation. It provides an important theoretical basis and decision-making reference to support the construction industry in its digital transformation and realize the goal of “dual carbon”.

Effect of diet, vitamin D3 and other factors on genital prolapse recurrensce

<p>This study addresses the pressing need for further investigation into risk factors contributing to genital prolapse recurrence, with a focus on factors like younger age and body mass index (BMI) that have been confirmed in the literature. Conducted as a cross-sectional study involving 300 post-operative cases of genital prolapse patients, the case groups comprising 210 individuals received regular gynecological check-ups every six months during the initial five years post-surgery, followed by annual visits, wherein lifestyle, diet, and laboratory values were monitored and adjusted. In contrast, the control group (90 patients) did not undergo post-operative gynecological follow-up. The results indicate that weight lifting, heavy physical work, menopause, constipation, insufficient protein intake, diseases associated with prolonged cough, and BMI are strongly associated with the likelihood of recurrent genital prolapse, while a weight reduction of 5 kg or more exhibits a protective effect (p<0.001). The established regression model proves statistically significant, explaining 84.1% of genital prolapse recurrence factors, with a sensitivity of 84.8% and specificity of 98.8%. These findings emphasize the importance of postoperative lifestyle monitoring, nutritional guidance, and immune support to reduce the risk of genital prolapse recurrence.</p>

Research on accurate analysis of coal quality using NIRS-XRF fusion spectroscopy in complex coal type scenarios

Coking coal in coal chemical enterprises presents a challenge due to its diverse types, wide sources, and variable quality, influenced by varying degrees of metamorphism and physicochemical properties. These differences not only impact coal quality but also hinder accurate analysis. Rapid and precise coal quality detection amidst diverse coal types is crucial for maintaining stable production and ensuring coke quality. This study employs near-infrared spectroscopy (NIRS) and X-ray fluorescence spectroscopy (XRF) fusion spectroscopy analysis, along with Principal Component Analysis (PCA) and t-Distributed Stochastic Neighbor Embedding (t-SNE), for data dimensionality reduction and visualization to devise a coal sample classification strategy. Subsequently, Support Vector Machine (SVM) is employed for automatic coal sample classification based on this strategy. Finally, Partial Least Squares Regression (PLSR) is used to establish regression models and evaluate their performance in predicting coal quality. Results show that classified regression model achieves R2 values of 0.9987, 0.9955, and 0.9997 for ash content, volatile matter, and sulfur content, with corresponding root mean square error for prediction (RMSEP) of 0.31 %, 1.34 %, and 0.05 %, and the mean absolute relative error for prediction (MARDP) of 2.48 %, 3.58 %, and 3.57 %, respectively. Compared to the unclassified model, there is a significant enhancement in prediction accuracy. The classification and modeling method proposed herein effectively improve the accuracy of coal quality analysis in complex coal type scenarios, crucial for industries like coal chemical engineering to enhance production efficiency and optimize coal resource utilization.

Performance evaluation and multi-objective optimization of a tubular indirect evaporative cooler integrated with moisture-conducting fibers

Indirect evaporative cooling (IEC) technology is an energy-efficient approach for regulating the indoor thermal environment of buildings. The conventional tubular indirect evaporative cooler (TIEC) may have a relatively low cooling efficiency due to poor wettability issues. The application of moisture-conducting fibers provides a feasible way to solve the above problem. However, the integration of moisture-conducting fibers with TIEC is still in the exploratory stage. This study proposed a novel moisture-conducting fiber-assisted TIEC and conducted a multi-objective optimization. An experimental facility and theoretical model of the proposed moisture-conducting fiber-assisted TIEC were developed. Based on the numerical model validated by experiments and response surface methodology (RSM), the regression models for performance prediction of the cooler were established. Eight input parameters including inlet air parameters, operating parameters and geometric parameters were selected, and four performance evaluation indicators were chosen as output responses. The parameter sensitivity of the regression models was analyzed. The multi-objective optimization was performed by considering the influence of different relative weights assigned to the output responses. Furthermore, the performance of the optimized cooler applied in different climate zones was predicted. The results showed that the product air temperature drop could achieve 8.8–11.3 °C after cooling by the cooler. The established regression models can predict the performance of the moisture-conducting fiber-assisted TIEC conveniently and effectively, which is expected to guide the design and optimization of engineering practices.

Assessment of post-pandemic economic recovery in Beijing, China, using night-time light data

ABSTRACT The coronavirus disease 2019 (COVID-19) pandemic posed a significant impact on human society, resulting in substantial changes in night-time light (NTL). This study developed a novel NTL index to estimate the quarterly Gross Domestic Product (GDP) at the county level in Beijing. This study aimed to investigate the economic recovery status in Beijing by establishing regression models of economic development before and after the pandemic. The estimated quarterly GDP at the county level was found to have a strong correlation with actual quarterly GDP statistics, with an average coefficient of determination (R 2) of 0.790. The regression models of the economic development of Beijing county-level districts before and after the pandemic were established after seasonal decomposition of the quarterly GDP. Based on the economic recovery framework for the pandemic in Beijing, most districts in the city had exhibited low economic resilience in terms of resistance and recoverability before and after the pandemic. Although the economy of Beijing is recovering gradually, it has not yet restored to pre-pandemic normal levels. These findings imply that NTL data can help assess the impact of emergency events on the disturbance of economic activity.

Remote Sensing Monitoring of Grassland Locust Density Based on Machine Learning.

The main aim of this study was to utilize remote sensing data to establish regression models through machine learning to predict locust density in the upcoming year. First, a dataset for monitoring grassland locust density was constructed based on meteorological data and multi-source remote sensing data in the study area. Subsequently, an SVR (support vector regression) model, BP neural network regression model, random forest regression model, BP neural network regression model with the PCA (principal component analysis), and deep belief network regression model were built on the dataset. The experimental results show that the random forest regression model had the best prediction performance among the five models. Specifically, the model achieved a coefficient of determination (R2) of 0.9685 and a root mean square error (RMSE) of 1.0144 on the test set, which were the optimal values achieved among all the models tested. Finally, the locust density in the study area for 2023 was predicted and, by comparing the predicted results with actual measured data, it was found that the prediction accuracy was high. This is of great significance for local grassland ecological management, disaster warning, scientific decision-making support, scientific research progress, and sustainable agricultural development.

Mixture ratio optimization of basalt fiber recycled aggregate concrete based on RSM

In this paper, the effects of W/B ratio, recycled aggregate (RA) content and basalt fiber (BF) dosage on the compressive strength, splitting tensile strength and slump of concrete were investigated and a regression model was developed for analysis and prediction, and the regression model was validated using ANOVA and confidence analysis. Statistical analysis revealed that W/B ratio, RA content and BF dosage had significant effects on all properties of concrete. The established regression model fits the experimental data well and can accurately predict the properties of concrete. It was also found that there was a significant interaction between W/B ratio and BF dosage, as well as RA content and BF dosage, which suggests that the combination of these factors should be considered in the design of concrete mixture ratio to achieve optimum performance. The predicted values of compressive strength, split tensile strength and slump of concrete under optimum proportioning were in good agreement with the experimental values. Therefore, optimizing the concrete proportion by response surface methodology can improve the testing efficiency and obtain basalt fiber recycled aggregate concrete that meets the design requirements and has better overall performance.

Application of the Surface Regression Technique for Enhancing the Input Factors and Responses for Processing Coconut Oil under Vertical Compression.

This study optimized the input processing factors, namely compression force, pressing speed, heating temperature, and heating time, for extracting oil from desiccated coconut medium using a vertical compression process by applying a maximum load of 100 kN. The samples' pressing height of 100 mm was measured using a vessel chamber of diameter 60 mm with a plunger. The Box-Behnken design was used to generate the factors' combinations of 27 experimental runs with each input factor set at three levels. The response surface regression technique was used to determine the optimum input factors of the calculated responses: oil yield (%), oil expression efficiency (%), and energy (J). The optimum factors' levels were the compression force 65 kN, pressing speed 5 mm min-1, heating temperature 80 °C, and heating time 52.5 min. The predicted values of the responses were 48.48%, 78.35%, and 749.58 J. These values were validated based on additional experiments producing 48.18 ± 0.45%, 77.86 ± 0.72%, and 731.36 ± 8.04 J. The percentage error values between the experimental and the predicted values ranged from 0.82 ± 0.65 to 2.43 ± 1.07%, confirming the suitability of the established regression models for estimating the responses.

A framework for computer-aided high performance titanium alloy design based on machine learning

Titanium alloy exhibits exceptional performance and a wide range of applications, with the high performance serving as the foundation for the development. However, traditional material design methods encounter numerous calculations and experimental trial-and-error processes, leading to increased costs and decreased efficiency in material design. The data-driven model presents an intriguing alternative to traditional material design methods by offering a novel approach to expedite the materials design process. In this study, a framework for computer-aided design high performance titanium alloys based on machine learning is proposed, which constructs an intelligent search space encompassing various combinations of 18 elements to facilitate alloy design. Firstly, a proprietary dataset was constructed for titanium alloy materials using feature design and a combination of unsupervised and supervised feature engineering methods. Secondly, six machine learning algorithms were employed to establish regression models, and the hyperparameters of each algorithm were optimized to improve model performance. Thirdly, the model was screened using five regression algorithm evaluation methods. The results demonstrated that the selected optimized model achieved an R2 value of 0.95 on the verification set and 0.93 on the test set, yielding satisfactory outcomes. Finally, a comprehensive model framework along with an intelligent search methodology for designing high-strength titanium alloys has been established. It is believed that this method is also applicable to other properties of titanium alloys and the optimization of other materials.

摆式主被动悬架喷雾机喷杆稳定性试验研究

When a sprayer is operating in the field, the uneven ground excitation causes the spray boom to move irregularly, significantly affecting the spray distribution uniformity and reducing the effectiveness of pesticide application. Installing a suspension between the vehicle and the boom is a crucial method to improve the boom stability. In this paper, experimental research on the stability of a boom with an active and passive pendulum suspension was carried out. The results of the transient response test of the passive suspension demonstrate that an increase in the suspension rotation damping coefficient reduces the overshoot of the system but slows down the response speed. Conversely, an increase in the suspension rotation stiffness coefficient speeds up the response speed. The results of the dynamic response test of the active suspension indicate that a smaller adjustment threshold of the control system for the boom inclination angle results in higher control accuracy. However, when the threshold is less than 1cm, the boom becomes challenging to balance. The results of the combination experiments based on the response surface method reveal that the rotation stiffness coefficient, rotation damping coefficient, unit forward speed, and their interactions significantly impact the adjustment time of the boom and the variation coefficient of the boom inclination angle. Through contribution rate analysis, the influence order of each factor on the adjustment time and variation coefficient was obtained. Additionally, the analysis of variance results show that the established regression model fits the actual situation well, and has reference significance for the design and application of the suspension.

Effect of Liver Segments and Hepatic Fibrosis Grade on Repeatability, Reliability, and Diagnostic Efficiency of Intravoxel Incoherent Motion

Intravoxel incoherent motion diffusion-weighted imaging (IVIM-DWI) is considered a potential marker of hepatic fibrosis (HF). To explore the influencing factors of repeatability and reliability in IVIM-DWI parameters of ROI-based liver segments in participants with HF and healthy volunteers (HV) and to assess the diagnostic efficiency of these parameters in HF. Participants with early HF (EHF, n=59) or advanced HF (AHF, n=38) and HV (n=48) were recruited. Two examiners measured IVIM data using mono-, bi-exponential and stretched exponential models. The results and influencing factors of repeatability and reliability of IVIM-DWI, and the diagnostic efficiency were analyzed. The repeatability of D* (CV: 26.62-41.47%) and DDC (CV: 18.01-34.40%) was poor, the repeatability of ADC (CV: 4.95-9.76%), D (CV: 7.09-15.52%), f (CV: 9.35-17.15%), and α (CV: 7.48-13.81%) was better; ordered logistic regression showed statistically significant results of IVIM-derived parameters; the reliability showed no obvious trend, and ordered logistic regression showed statistically significant results of IVIMderived parameters, groups, and partial hepatic segments (all p<0.001). IVIM-derived parameters with relatively good repeatability (CV<20%) and reliability (ICC>0.4) were used to establish regression models for differential diagnosis. The AUC of regression models was 0.744-0.783 (EHF vs. AHF), but no statistically significant parameters were found in the HV vs EHF comparison. IVIM-derived parameters were the most important factors affecting the repeatability and reliability, while staging of HF and hepatic segments may be the influencing factors of reliability. IVIM-derived parameters showed medium diagnostic efficiency in distinguishing between EHF and AHF.

Design and analysis of three-dimensional chiral metamaterials for enhanced torsional compliance

Three-dimensional (3D) chiral structures are a subset of mechanical metamaterials known for their distinctive deformation characteristics, including the presence of auxetic and compression–torsion coupling behaviors. In this study, a novel 3D chiral metamaterial with high torsional compliance was designed. A gammadion-shaped chiral cell comprising inclined ligaments was incorporated into the chiral structure for enhanced torsional compliance. Finite element analyses (FEAs) were performed to demonstrate that the proposed chiral structure outperformed other chiral structures in terms of torsional compliance. Furthermore, FEAs were performed to investigate the influence of the slenderness ratio and ligament diameter of the chiral structure, on attaining a maximum torsional compliance of 18.6° N−1. These 3D chiral structures were fabricated using additive manufacturing techniques, and experimental validations were performed to observe and confirm the compression–torsion coupling behaviors. To further expand the utility of these 3D chiral structures, statistical analyses were conducted to establish regression models for effective density and torsional compliance, as functions of the slenderness ratio and ligament diameter. These regression models can enhance the applicability of the proposed chiral structures in the development of versatile functional components that require compression–torsion coupling behaviors.

The Study of Various Regression Models Establishment to Identify Farmland Soil Moisture Content at Different Depths Using Unmanned Aerial Vehicle Multispectral Data: A Case in North China Plain

Soil moisture content is one of the most important soil indices for agriculture production. With the increasing food requirement and limited irrigation water sources, it is of great significance to accurately and quickly measure the soil moisture content for precision irrigation, especially in deficient agricultural areas, such as North China Plain. To achieve this goal, more attention was paid to the application of unmanned aerial vehicle multispectral reflectance technology. However, it was urgent to enhance the regression models between spectral data and soil realistic moisture content, and there were limited studies about the regression research on deep soil layers. Thus, the farmland of winter wheat–summer maize double cropping at Yongnian District, Hebei, North China, was selected as the study area. A six-band multispectral camera mounted on a low-altitude unmanned aerial vehicle (UAV) was used to obtain the field spectral reflectance with bands from 470~810 nm, and meanwhile, soil moisture content at different depths (10, 20, 30, 40, 50, and 60 cm) was measured after maize sowing. Unary linear regression (ULR), multivariate linear regression (MLR), ridge regression (RR), and an artificial neural network (ANN) were employed to establish regression models. The results demonstrated that the sensitive bands of spectral reflectance were 690 nm, 470 nm, and 810 nm. Those models all established significant regression at the depths of 0–20 cm and 40–60 cm, particularly at 10, 50, and 60 cm soil layers. However, for a depth of 20–40 cm, the prediction accuracy was generally lower. Among MLR, RR, and BP models, the MLR exhibited the highest identification accuracy, which was most recommended for the application. The findings of this study provide technical guidance and effective regression for the multispectral reflectance on the farmland of North China Plain, especially for deep soil layer moisture prediction.

Research on microstructure evolution and surface quality of WEDM for magnetic refrigerant rare-earth gadolinium

Rare-earth gadolinium (Gd) is preferable for manufacturing regenerators of the core components of room-temperature magnetic refrigeration owing to its unique magnetocaloric and mechanical properties. However, the surface quality of the regenerator plays a crucial role in the heat transfer effect and service life of magnetocaloric systems during wire electrical discharge machining (WEDM) when fabricating rare-earth Gd array microstructure regenerators. In this study, different process parameters were used to conduct a process experiment of the WEDM of rare-earth Gd. First, the evolution of the surface microstructure and its causes were analyzed using a single-factor experiment, while a corrosion test was conducted on the samples. The analysis showed that the pulse-on time and open voltage considerably affected the surface quality of the processed samples, while the samples with better surfaces exhibited good corrosion resistance. Additionally, a Taguchi experiment was designed, and a regression analysis used to establish regression models between the process parameters (pulse-on time, pulse-off time, peak current, open voltage, and water pressure) and both surface roughness (SR) and material removal rate (MRR). The results showed that the average prediction errors of SR and MRR were only 5.34% and 5.48%, respectively.