3D Model Research Articles

Whale Collections and Exhibitions at the Natural History Museum of the University of Pisa (Italy)

The Natural History Museum of the University of Pisa hosts the most important osteological collection of extant cetaceans in Italy as well as one of the most relevant all over Europe. Furthermore, it also preserves a significant palaeontological collection that includes several holotypes and otherwise unique specimens of Archaeoceti (archaic cetaceans), Mysticeti (baleen whales) and Odontoceti (toothed whales). Here, we provide a historical overview of these collections and the corresponding displays, with special attention paid to the origin, development and design of the ‘Archaeocete Hall’ and ‘Cetacean Gallery’. These comprise what may be the largest exhibition worldwide among those dedicated exclusively to cetaceans—one that includes 28 complete skeletons and one skull belonging to 27 extant species as well as fossils of nine extinct species. Our review also reveals that the museum exhibitions feature the oldest known specimen of Mesoplodon bowdoini and the type specimen of Ziphius savi, the latter being a validly described species that is currently regarded as a junior synonym of Ziphius cavirostris. Also significant is the display of several holotype specimens of fossil species such as the protocetid archaeocete Aegyptocetus tarfa, the balaenid baleen whales Balaena montalionis and Balaenula astensis, and the monodontid Casatia thermophila. The Archaeocete Hall and Cetacean Gallery are highly appreciated by visitors as well as perused by the museum’s educational team. The online archiving of 3D models of many of the MSNUP specimens on the open-access digital repository Sketchfab and their subsequent dissemination through the Wikimedia platforms has led to the creation of a major osteological resource—one that is broadly accessible to internet users worldwide.

Accelerating segmentation of fossil CT scans through Deep Learning

Recent developments in Deep Learning have opened the possibility for automated segmentation of large and highly detailed CT scan datasets of fossil material. However, previous methodologies have required large amounts of training data to reliably extract complex skeletal structures. Here we present a method for automated Deep Learning segmentation to obtain high-fidelity 3D models of fossils digitally extracted from the surrounding rock, training the model with less than 1%-2% of the total CT dataset. This workflow has the capacity to revolutionise the use of Deep Learning to significantly reduce the processing time of such data and boost the availability of segmented CT-scanned fossil material for future research outputs. Our final Unet segmentation model achieved a validation Dice similarity of 0.96.

Hybrid Ising Machine and Reinforcement Learning Approach to Folding Lattice Proteins

The hydrophobic polarity (HP) protein folding model is a simplified computation model utilized to study the folding of proteins into their three-dimensional structures. Understanding how proteins fold and misfold is fundamental to designing effective, safe drugs, allowing for targeted therapies and better patient outcomes. The amino acids of a protein can be categorized as either hydrophobic or polar, where the hydrophobic beads tend to cluster together in a properly folded protein to minimize exposure to the surrounding environment. Determining the lowest energy configurations of the HP model is a non-deterministic polynomial (NP) hard problem. One method is formulating the HP model as a quadratic unconstrained binary optimization (QUBO) problem which can be mapped to an Ising machine. In the Ising model, the lowest energy state of the system represents the correct lattice configuration of the folded protein. The Ising machine technique utilized of simulated annealing allows for future application of the work on an integrated silicon photonic chip. Alternatively, reinforcement learning designed with an energy-based reward function can optimize the HP model by adapting to the environment without the consequence of remaining in local minima like that of the Ising machine. A hybrid pipeline combining the Ising machine and reinforcement learning sequentially was designed to determine the two-dimensional lattice configurations of various protein chain lengths and sequences, combining the parallel optimization of the Ising machine on an integrated photonic platform and the adaptability of the reinforcement learning algorithm. Results found that this hybrid model improved the accuracy and ability of determining the lowest energy state in contrast to the Ising machine alone as well as improved the efficiency by reducing the number of iterations required when compared solely to reinforcement learning. Future work focuses on implementing a parallel pipeline approach as well as adapting to a three-dimensional model.

Improving a 1D Hydraulic Model to Include Bridges as Internal Boundary Conditions

The paper describes the implementation of internal boundary conditions in the 1D ORSADEM hydraulic model to simulate the effect of a hydraulic in-line structure. The proposed model introduces a simplified representation of the bridge geometry by imposing an equivalent narrowing, computed according to the opening size and characteristics, combined with the mass and energy balance at the structure. The model is then applied to a series of experimental tests concerning the propagation of shock waves through different types of bridges, representing different flow conditions, from free surface flow to overflow. The tests are also simulated with the original 1D ORSADEM model, including the standard head losses and the cross-section narrowing due to the presence of a structure. The comparison with the experimental measurements shows that the proposed model can simulate the shock wave flow through the bridges with a higher accuracy than the standard formulation. These findings highlight the possibility of properly evaluating the backwater effect at bridges even with a simple 1D model if the physical narrowing of the cross-section is modeled.

Exploring additive manufacturing in assistive technologies to transform the educational experience: empowering inclusion

This study explores additive manufacturing's potential in creating tailored assistive technologies, fostering inclusion in education. To design ergonomic solutions for diverse user needs, promoting inclusion and autonomy. A sequential approach involves assessment, 3D modeling, additive manufacturing, and aligning with the Inclusion Support and Accessibility Promotion Program (PIPA). Various assistive technologies were developed, addressing specific needs. The “Ergonomic Pen and Pencil Reamer” improves grip accuracy, and the “Folding Handle for Folding Cup” offers an ergonomic solution. The “Guideline Ruler” supports music education, with accessories like the “Template for Clefs and Musical Notes” and “Thimble with Support Base for Musical Instrument Strings” enhancing versatility. The “Ring with Front Support” aids art education, and the “Support for Scissors” promotes independence. These innovations contribute to inclusion and autonomy. The research underscores the importance of additive manufacturing in crafting personalized solutions, propelling inclusion in education and daily activities. Identified challenges, including material selection and adaptation to diverse needs, signal areas for future research. Continuous collaboration with end-users and professionals remains crucial for enhancing usability and effectiveness, reinforcing the commitment to promoting inclusion and autonomy.

Evaluation and design of a bite force measuring load cell

The measurement of bite force can be essential for assessing oral health and the functioning of the stomatognathic system. However, its use in dental offices is limited due to operational complexities and costs. This study focused on evaluating the mechanical strength of the user-friendly and low manufacturing cost load cell under development within LABBIO – Bioengineering Laboratory at UFMG, the gnathodynamometer, tailored for children. 3D modeling in SolidWorks and Finite Element Method (FEM) structural analysis were employed to ensure its ability to withstand the desired 1000 N load. Initially, the device did not meet the desired load, resulting in iterations to modify its geometry. After iterations including the replacement of the initial material with a more suitable alternative, removal of edges, and other adjustments, a final prototype meeting mechanical strength and functionality requirements was developed.

Experimental and numerical analysis of ballistic impat and material characterization of GFRP and Kevlar 29/epoxy composite laminate

Through experimental testing and numerical simulations, this study examines the effects of ballistic events on thin FRP composite laminated plates fabricated using a hand lay-up method. Experimental ballistic impact tests using a pneumatic gun are conducted on composite plates reinforced with woven glass and woven Kevlar 29 fibers. An advanced three-dimensional finite element model programed in Ansys/Autodyn v19.1 is employed to verify the experimental findings and analyze the ballistic perforation characteristics of the target. The crucial material constants needed for the constitutive material model used in the simulation are acquired through precise experimentation on samples prepared from the fabricated laminates. Significant agreement is observed between the FE simulations and experimental findings, particularly concerning the assessment of residual velocities of the projectile and damage pattern in the laminates. The results of this study show that when subjected to ballistic impact by a flat-ended cylindrical projectile, the thin woven FRP composite primarily experiences damage characterized by delamination, fiber breakage and matrix cracking. Additionally, based on current simulations, it is observed that the ballistic limit velocity of the Kevlar 29/epoxy laminate exceeds that of GFRP by 25.64% when both materials have an equal thickness of 2.8 mm.

Three-Dimensional Spectral Element Method Implementation for Evaluating Rooted Soil Behavior in Slope Stability Analysis

Bioengineering techniques are being increasingly adopted as a sustainable solution to soil slope instability. Despite their recognized benefits, however, the mechanistic contribution of vegetation to slope stability remains inadequately understood due to the intricate nature of soil–root interactions and the complexity of root architectures. Most existing research predominantly offers qualitative assessments of vegetation effectiveness. This study aims to numerically substantiate the role of vegetation as a bioengineering technique for soil slope stabilization. Various plant species used commonly for soil stabilization were identified, and undisturbed soil samples were collected to quantify the shear strength parameters of the soils from both barren and vegetated slopes, along with root tensile strengths. A comprehensive topographic survey was conducted to capture the precise topography of the study area. Utilizing these primary data, in this work we develop 3D models for five representative plants within each species category and employ the Spectral Element Method (SEM), an advanced higher-order formulation of the finite element method (FEM), within the 3D domain to evaluate the factor of safety for the soil slopes. The SEM offers superior accuracy and stability in numerical computations. The results obtained through the SEM were corroborated through the FEM modeling and were found to be consistent with other established methodologies. This innovative approach of 3D SEM aims to quantitatively assess the impact of vegetation on soil slope stability and provide a more rigorous understanding of bioengineering applications in geotechnical engineering.

Comparison of compartmental analytical Blood-Oxygen-Level-Dependent functional Magnetic Resonance Imaging models against Monte Carlo simulations performed over cortical micro-angiograms.

Blood oxygen level-dependent functional magnetic resonance imaging (BOLD fMRI) arises from a physiological and physical cascade of events taking place at the level of the cortical microvasculature which constitutes a medium with complex geometry. Several analytical models of the BOLD contrast have been developed, but these have not been compared directly against detailed bottom-up modeling methods. Using a 3D modeling method based on experimentally measured images of mice microvasculature and Monte Carlo simulations, we quantified the accuracy of two analytical models to predict the amplitude of the BOLD response from 1.5 to 7T, for different echo time (TE) and for both gradient echo and spin echo acquisition protocols. We also showed that accounting for the tridimensional structure of the microvasculature results in more accurate prediction of the BOLD amplitude, even if the values for SO2 were averaged across individual vascular compartments. A secondary finding is that modeling the venous compartment as two individual compartments results in more accurate prediction of the BOLD amplitude compared with standard homogenous venous modeling, arising from the bimodal distribution of venous SO2 across the microvasculature in our data.

Enhancing medical anatomy education with the integration of virtual reality into traditional lab settings.

In medical education, traditional anatomy labs have relied heavily on the hands-on dissection of cadavers to teach the complex spatial relationships within the human body. However, the advent of virtual reality (VR) technology offers the potential for significantly enhancing this traditional approach by providing immersive, interactive 3D visualizations that can overcome some of the limitations of physical specimens. This study explores the integration of VR into a traditional gross anatomy lab to enrich the learning experience for medical students. Methods included the deployment of a VR application developed to complement the dissection process, featuring detailed 3D models of human anatomy that students could manipulate and explore digitally. Approximately 60 s-year medical students participated in the lab, where they engaged with both traditional dissection and the VR application. Results indicated that the VR integration not only increased engagement and satisfaction but also improved the students' ability to understand anatomical structures and their spatial relationships. Moreover, feedback from students suggested more efficient learning and retention than with traditional methods alone. We conclude that VR technology can significantly enhance medical anatomy education by providing an adjunct to traditional dissection, potentially replacing certain aspects of physical specimens with digital simulations that offer repeatable, detailed exploration without the associated logistical and ethical constraints.

Computer-Simulated Virtual Image Datasets to Train Machine Learning Models for Non-Invasive Fish Detection in Recirculating Aquaculture.

Artificial Intelligence (AI) and Machine Learning (ML) can assist producers to better manage recirculating aquaculture systems (RASs). ML is a data-intensive process, and model performance primarily depends on the quality of training data. Relatively higher fish density and water turbidity in intensive RAS culture produce major challenges in acquiring high-quality underwater image data. Additionally, the manual image annotation involved in model training can be subjective, time-consuming, and labor-intensive. Therefore, the presented study aimed to simulate fish schooling behavior for RAS conditions and investigate the feasibility of using computer-simulated virtual images to train a robust fish detection model. Additionally, to expedite the model training and automate the virtual image annotation, a process flow was developed. The 'virtual model' performances were compared with models trained on real-world images and combinations of real and virtual images. The results of the study indicate that the virtual model trained solely with computer-simulated images could not perform satisfactorily (mAP = 62.8%, F1 score = 0.61) to detect fish in a real RAS environment; however, replacing a small number of the virtual images with real images in the training dataset significantly improved the model's performance. The M6 mixed model trained with 630 virtual and 70 real images (virtual-to-real image ratio: 90:10) achieved mAP and F1 scores of 91.8% and 0.87, respectively. Furthermore, the training time cost for the M6 model was seven times shorter than that for the 'real model'. Overall, the virtual simulation approach exhibited great promise in rapidly training a reliable fish detection model for RAS operations.

Virtual Tours as Effective Complement to Building Information Models in Computer-Aided Facility Management Using Internet of Things

This study investigates the integration of Building Information Models (BIMs) and Virtual Tour (VT) environments in the Architecture, Engineering and Construction (AEC) industry, focusing on Computer-Aided Facility Management (CAFM), Computerized Maintenance Management Systems (CMMSs), and data Life-Cycle Assessment (LCA). The interconnected nature of tasks throughout a building’s life cycle increasingly demands a seamless integration of real-time monitoring, 3D models, and building data technologies. While there are numerous examples of effective links between IoT and BIMs, as well as IoT and VTs, a research gap exists concerning VT-BIM integration. This article presents a technical solution that connects BIMs and IoT data using VTs to enhance workflow efficiency and information transfer. The VT is developed upon a pilot based on the Controlled Environments for Living Lab Studies (CELLS), a unique facility designed for flexible monitoring and remote-control processes that incorporate BIMs and IoT technologies. The findings offer valuable insights into the potential of VTs to complement and connect to BIMs from a life-cycle perspective, improving the usability of digital twins for beginner users and contributing to the advancement of the AEC and CAFM industries. Our technical solution helps complete the connectivity of BIMs-VT-IoT, providing an intuitive interface (VT) for rapid data visualisation and access to dashboards, models and building databases. The practical field of application is facility management, enhancing monitoring and asset management tasks. This includes (a) sensor data monitoring, (b) remote control of connected equipment, and (c) centralised access to asset-space information bridging BIM and visual (photographic/video) data.

Refined Prior Guided Category-Level 6D Pose Estimation and Its Application on Robotic Grasping

Estimating the 6D pose and size of objects is crucial in the task of visual grasping for robotic arms. Most current algorithms still require the 3D CAD model of the target object to match with the detected points, and they are unable to predict the object’s size, which significantly limits the generalizability of these methods. In this paper, we introduce category priors and extract high-dimensional abstract features from both the observed point cloud and the prior to predict the deformation matrix of the reconstructed point cloud and the dense correspondence between the reconstructed and observed point clouds. Furthermore, we propose a staged geometric correction and dense correspondence refinement mechanism to enhance the accuracy of regression. In addition, a novel lightweight attention module is introduced to further integrate the extracted features and identify potential correlations between the observed point cloud and the category prior. Ultimately, the object’s translation, rotation, and size are obtained by mapping the reconstructed point cloud to a normalized canonical coordinate system. Through extensive experiments, we demonstrate that our algorithm outperforms existing methods in terms of performance and accuracy on commonly used benchmarks for this type of problem. Additionally, we implement the algorithm in robotic arm-grasping simulations, further validating its effectiveness.

Calculation of AC loss using 2D homogenization method for HTS synchronous condenser rotor and validation

Compared with low-temperature superconductors, second-generation high-temperature superconducting materials offer higher current densities and temperature margins, which provide a boost to the development of superconducting condenser towards large capacity and light weight. In this paper, a 3D-to-2D dimensionality reduction and simplification method is proposed to address the problem of massive computation of 3D finite element model during the optimization of superconducting condenser rotor performance parameters. This method splits the pole based on the geometric characteristics of the rotor pole, establishes a 2D simplified model of each part of the pole after splitting, and uses the simplified model to calculate the key physical quantities such as the magnetic field, AC loss, temperature change, etc., during the excitation process of the superconducting condenser rotor. This method effectively reduces the computational complexity of the model and improves the efficiency of the calculation and iterative optimization of the magnet pole parameters. In order to verify the correctness of the electromagnetic calculation results of the simplified model, the rotor magnet model was assembled to carry out experiments in liquid nitrogen bath, and the experimental results were in good agreement with the calculation results.

The Bony Nasal Cavity and Paranasal Sinuses of Big Felids and Domestic Cat: A Study Using Anatomical Techniques, Computed Tomographic Images Reconstructed in Maximum-Intensity Projection, Volume Rendering and 3D Printing Models.

This study aims to develop three-dimensional printing models of the bony nasal cavity and paranasal sinuses of big and domestic cats using reconstructed computed tomographic images. This work included an exhaustive study of the osseous nasal anatomy of the domestic cat carried out through dissections, bone trepanations and sectional anatomy. With the use of OsiriX viewer, the DICOM images were postprocessed to obtaining maximum-intensity projection and volume-rendering reconstructions, which allowed for the visualization of the nasal cavity structures and the paranasal sinuses, providing an improvement in the future anatomical studies and diagnosis of pathologies. DICOM images were also processed with AMIRA software to obtain three-dimensional images using semiautomatic segmentation application. These images were then exported using 3D Slicer software for three-dimensional printing. Molds were printed with the Stratasys 3D printer. In human medicine, three-dimensional printing is already of great importance in the clinical field; however, it has not yet been implemented in veterinary medicine and is a technique that will, in the future, in addition to facilitating the anatomical study and diagnosis of diseases, allow for the development of implants that will improve the treatment of pathologies and the survival of big felids.

Six-Degree-of-Freedom Pose Estimation Method for Multi-Source Feature Points Based on Fully Convolutional Neural Network

An object’s six-degree-of-freedom (6DoF) pose information has great importance in various fields. Existing methods of pose estimation usually detect two-dimensional (2D)-three-dimensional (3D) feature point pairs, and directly estimates the pose information through Perspective-n-Point (PnP) algorithms. However, this approach ignores the spatial association between pixels, making it difficult to obtain high-precision results. In order to apply pose estimation based on deep learning methods to real-world scenarios, we hope to design a method that is robust enough in more complex scenarios. Therefore, we introduce a method for 3D object pose estimation from color images based on farthest point sampling (FPS) and object 3D bounding box. This method detects the 2D projection of 3D feature points through a convolutional neural network, matches it with the 3D model of the object, and then uses the PnP algorithm to restore the feature point pair to the object pose. Due to the global nature of the bounding box, this approach can be considered effective even in partially occluded or complex environments. In addition, we propose a heatmap suppression method based on weighted coordinates to further improve the prediction accuracy of feature points and the accuracy of the PnP algorithm in solving the pose position. Compared with other algorithms, this method has higher accuracy and better robustness. Our method yielded 93.8% of the ADD(-s) metrics on the Linemod dataset and 47.7% of the ADD(-s) metrics on the Occlusion Linemod dataset. These results show that our method is more effective than existing methods in pose estimation of large objects.

Digital analysis of the human maxilla to enable semistandardized template tool reconstructions with free fibula transplants

ObjectivesThis study analyzed the human maxilla to support the development of mean-value-based cutting guide systems for maxillary reconstruction, bridging the gap between freehand techniques and virtual surgical planning (VSP).Materials and methodsThis retrospective cohort study used routine CT scans. DICOM data enabled 3D modelling and the maxilla was divided into four regions: paranasal (R1), facial maxillary sinus wall (R2), zygomatic bone (R3) and alveolar process (R4). Surface comparisons were made with a reference skull. Statistical analyses assessed anatomical variations, focusing on mean distance (Dmean), area of valid distance (AVD), integrated distance (ID) and integrated absolute distance (IAD). The study addressed hemimaxillectomy defects for two-segmental reconstructions using seven defined bilateral points to determine segmental distances and angles.ResultsData from 50 patients showed R2 as the most homogeneous and R4 as the most heterogeneous region. Significant age and gender differences were found in R3 and R4, with younger patients and females having more outliers. Cluster analysis indicated that males had R1 and R3 positioned anterior to the reference skull. The mean angle for segmental reconstruction was 131.24° ± 1.29°, with anterior segment length of 30.71 ± 0.57 mm and posterior length of 28.15 ± 0.86 mm.ConclusionsAnatomical analysis supported the development of semistandardized segmental resection approaches. Although gender and anatomical differences were noted, they did not significantly impact the feasibility of mean-value-based cutting-guide systems.Clinical relevanceThis study provides essential anatomical data for creating cost-effective and efficient reconstruction options for maxillary defects, potentially improving surgical outcomes and expanding reconstructive possibilities beyond current techniques.

Engineered Liver-derived Decellularized Extracellular Matrix-based Three-Dimensional Tumor Constructs for Enhanced Drug Screening Efficiency

Abstract The decellularized extracellular matrix (dECM) has emerged as an effective medium for replicating the in vivo-like conditions of the tumor microenvironment (TME), thus enhancing the screening accuracy of chemotherapeutic agents. However, recent dECM-based tumor models have exhibited challenges such as uncontrollable morphology and diminished cell viability, hindering the precise evaluation of chemotherapeutic efficacy. Herein, we utilized a tailor-made microfluidic approach to encapsulate dECM from porcine liver in highly poly(lactic-co-glycolic acid) (PLGA) porous microspheres (dECM-PLGA PMs) to engineer a three-dimensional (3D) tumor model. These dECM-PLGA PMs-based microtumors exhibited significant promotion of hepatoma carcinoma cells (HepG2) proliferation compared to PLGA PMs alone, since the infusion of extracellular matrix (ECM) microfibers and biomolecular constituents within the PMs. Proteomic analysis of the dECM further revealed the potential effects of these bioactive fragments embedded in the PMs. Notably, dECM-PLGA PMs-based microtissues effectively replicated the drug resistance traits of tumors, showing pronounced disparities in half-maximal inhibitory concentration (IC50) values, which could correspond with certain aspects of the TME. Collectively, these dECM-PLGA PMs substantially surmounted the prevalent challenges of unregulated microstructure and suboptimal cell viability in conventional 3D tumor models. They also offer a sustainable and scalable platform for drug testing, holding promise for future pharmaceutical evaluations.

Vanillin, ferulic acid and their 1:1 combination inhibit lipid accumulation in 3T3 L1 adipocytes and 3D spheroids

BackgroundObesity is poised to be a major healthcare crisis worldwide. Genetic predisposition, inadequate activity, changing lifestyle and dietary patterns are cited as major causes for obesity. Even as a number of anti-obesity medications hit the market, there is still an ongoing quest to explore natural compounds, which are perceived as safer alternatives, for their anti-obesity activity. This study explores the anti-obesity potential of dietary polyphenols vanillin, ferulic acid and their combination using 3T3 L1 adipocytes and their 3D spheroids. MethodsStudies were conducted on differentiated 3T3 L1 adipocytes and their 3D spheroids. Assays conducted on 3T3 L1 adipocytes include Oil red O, fluorescent Nile Red staining and triglyceride quantification to assess effect on lipid droplet accumulation. 2 NBDG was used to assess glucose uptake following drug treatment. 3D spheroid cultures were generated and triglyceride content was quantified. Effect of drug treatment on gene expression was analysed using qRT-PCR. Results of monolayer culture were compared with 3D spheroid models. ResultsVanillin, ferulic acid and their combination lower intracellular triglyceride content and lipid droplet accumulation, inhibiting glucose uptake and conversion to triglycerides in 3T3 L1 adipocytes and their 3D spheroids. Compounds and their combination downregulated mRNA expression of C/EBP α and PPAR ɣ, FAS, ACC1, GLUT4, LPL, aP2. Vanillin treatment upregulated leptin mRNA expression. ConclusionVanillin, ferulic acid and their combination lower lipid accumulation and glucose uptake in 3T3 L1 adipocytes and 3D spheroids.

Numerical simulation of thermal performance of H2O-EGS and CO2-EGS based on thermal-hydraulic-mechanical coupling method

The use of hydrothermal geothermal methods in Enhanced Geothermal Systems (EGS) presents challenges like reduced thermal storage life and high external energy consumption. Due to its stable heat production time and lower external energy demand, CO2 has the potential to be substituted for H2O. The research zone for this study was chosen to be located in the HDR reservoir in the Gonghe Basin of Qinghai. A three-dimensional discrete fracture model based on a thermal-hydraulic-mechanical coupling method is established, where numerical simulations are conducted using COMSOL software. The discussion focuses on the comparison of heat production effects between H2O-EGS and CO2-EGS in different injection and extraction scenarios are discussed. The results indicate that by lowering the injection temperature and increasing the injection rate, the EGS net heat production rate can be increased, but it also accelerates the heat breakthrough time and shortens the reservoir life. Although CO2-EGS has a lower heat extraction rate in the early stage of thermal recovery than H2O-EGS, it has a longer stable heat production time and a more energy-efficient heat production process. Therefore, compared to H2O-EGS, CO2-EGS has more economic and social benefits.