Design and Implementation of a Virtual Experimental Teaching System for Deep Energy Exploitation Based on Digital Twin Technology

Abstract

The exploration and development of deep oil and gas resources are becoming the primary focus in the fossil energy sector, thereby increasing the demand for highly skilled engineers. Colleges and universities play a crucial role in cultivating talent in petroleum engineering. However, the current traditional teaching systems, particularly in experimental practices, face significant challenges, such as low efficiency, limited environments, and a disconnect between theoretical knowledge and practical application. To address these issues and enhance learners’ practical abilities and comprehension, we introduced digital twin technology into the experimental teaching of deep energy exploitation. This paper analyzes innovative pedagogical approaches, with a special emphasis on the real-time visualization of hydraulic fracturing. Supported by the National Key Laboratory of Chengdu University of Technology, our research team developed multiple digital twin platforms for both indoor and onsite hydraulic fracturing. These platforms utilize advanced algorithms and models, enabling real-time data acquisition and visualization analysis. Pilot teaching results demonstrate that the virtual experimental system based on digital twin technology encourages active learner engagement, improves their understanding of digitalization in engineering, and enhances their professional skills in deep oil and gas exploration. The digital twin-based visualization system is a valuable tool for experimental teaching in deep energy exploitation, and its application could serve as a model for other engineering disciplines.

1. Introduction

As exploration and development in shallow and intermediate formations progress, it becomes increasingly difficult to discover new oil and gas reserves. Consequently, deep resources have emerged as the primary focus for fossil energy exploration [1,2]. However, deep oil and gas in theory and technology appear to have new problems and challenges. Simultaneously, higher professional standards are being demanded for petroleum engineers [3,4]. Higher education institutions, as key platforms for cultivating talent, face the critical task of training professionals who possess both theoretical knowledge and practical skills to meet the demands of the evolving petroleum industry [5,6,7].

The traditional teaching system in Chinese universities tends to prioritize basic theoretical knowledge when educating petroleum engineering learners [8]. Educational staff develop a systematic and comprehensive curriculum based on these requirements. Through a series of courses, learners learn to apply the fundamentals of physics, mathematics, mechanics, and other disciplines to solve practical problems [9]. However, experimental teaching, a crucial component of the educational system, often follows a more conventional approach. At present, there are some bottlenecks in experimental education, including limitations in experimental facilities, equipment, and other resources. Overall, it does not fully meet learners’ needs for learning complex content. To enhance teaching effectiveness, there is an urgent need to reform the methods and systems used in experimental education. This would help achieve the goals of reducing risk, lowering costs, and increasing resourcefulness.

As digital education continues to develop, experimental teaching systems utilizing virtual simulation technology are gaining increased attention [10]. This approach has been tried in some engineering specialities [11,12,13,14]. A typical engineering simulation teaching system generally comprises a simulation experimental platform, a construction process support model, and a virtual experimental teaching platform. By integrating this new teaching method, learners have improved their understanding of practical experiments through virtual–real interaction. Although some universities have tried experimental teaching systems based on virtual simulation technology, these courses are primarily focused on surface engineering.

Unlike surface engineering, deep energy exploitation (e.g., hydraulic fracturing) typically occurs several kilometers below the surface. It involves multiple disciplines, such as fracture propagation in a multifield coupling environment and the interaction of multiple fractures. These complex phenomena require explanation through abstract and intricate theories, making theoretical learning particularly challenging [15,16,17]. A key challenge in hydraulic fracturing is the invisibility of the entire process, as feedback is typically limited to a few operational curves and high-cost monitoring data [18,19]. Although universities offer fieldwork as part of practical courses, it remains difficult for learners to grasp fully the physical processes occurring underground. To provide high-quality education, new technologies are needed. Currently, virtual simulation technology is widely used in petroleum development projects to bridge this gap. In addition, incorporating technologies developed by researchers can greatly enhance teaching. Using seismic profiles, microseismic monitoring data, and well-log interpretation results, 3D static reservoir modeling methods and hydraulic fracturing simulation models have been introduced [20,21,22,23]. These models help students better understand stratigraphic properties and the behavior of hydraulic fracture propagation. Moreover, the use of such systems not only reduces teaching costs but also minimizes carbon emissions.

In recent years, the rapid advancement of science and technology, particularly in computer systems, has created new opportunities for engineering education [24,25,26,27]. The digital twin and augmented reality (AR) are both significant tools in modern technology, each offering distinct advantages and serving different applications. The digital twin is able to collect data from physical entities in real time through sensors and data acquisition systems. The data can be monitored and analyzed through virtual models [28,29]. AR is primarily about overlaying virtual information onto the real world. While there is some use of sensor data, its main purpose is the enhancement of the visual experience. Unlike AR, the digital twin emphasizes the interaction between virtual models and the physical world, helping to predict and optimize complex systems. In the teaching system of deep energy exploitation, the integration of digital twin technology can significantly enhance the quality of experimental systems. It also broadens the scope of the discipline and reduces the costs associated with onsite teaching. Notably, this technology has already been successfully trialed in several petroleum engineering disciplines [30,31,32]. Some relevant techniques have been visualized. For example, a target reservoir visualization model has been developed using seismic interpretation results, geological information, and experimental data. Additionally, a real-time production data update program has been created based on monitoring data [33]. As an effective teaching tool, digital twin technology allows learners to engage more intuitively, efficiently, and actively in practical lessons. This improves their mastery and application of existing knowledge. Digital twin models are developed using standard or custom programming languages, offering highly open and inclusive features. This flexibility allows learners to incorporate innovation into the program, enabling them to apply theoretical results in practical scenarios [34,35]. Although digital twin technology has made some progress in petroleum engineering education, its application has primarily focused on data processing and integration, which can be classified as static visualization. For hydraulic fracturing, achieving dynamic visualization through digital twins remains a challenge. Currently, digital twin technology is rarely incorporated into experimental teaching systems in universities.

Taking hydraulic fracturing technology as the focus of instruction, the team from Chengdu University of Technology has developed a platform and system that uses digital twin technology to support indoor and onsite experimental teaching. This enables instructors to teach theoretical knowledge in the context of real-world engineering applications, encouraging learners to actively and positively engage in experimental learning. We have also integrated concepts such as artificial intelligence, computational programming, and numerical simulation to enrich the digital twin model of hydraulic fracturing. Learners are expected to contribute to the development of visualization techniques for deep energy exploitation. Results from the pilot teaching indicate that the digital twin-based experimental teaching system for deep energy exploitation has yielded positive outcomes. This approach not only fosters active learner participation in the learning process but also enhances their understanding of engineering digitalization.

2. Construction of Digital Twin System for Hydraulic Fracturing Visualization

Hydraulic fracturing, an effective technology for extracting oil and gas, uses high-pressure pumping units to inject highly viscous fluids into the reservoir. Fractures are then created to allow the oil or gas to escape from the fractures and flow to the surface. This process involves several physical processes, including fluid flow, rock deformation and failure, as well as the coupling between these processes. The complexity of the fracture formation in deep reservoirs, influenced by both geological and engineering factors, poses significant challenges for teaching.

On the theoretical side, courses in petroleum engineering, rock mechanics, fluid mechanics, and related fields help learners master the fundamentals. However, due to the abstraction and idealization of these subjects, learners often struggle to comprehend fully the material. Emphasizing experimental learning is essential in the educational approach to hydraulic fracturing. Given the inability to directly observe the fracturing process at depths of several thousand meters, feedback is primarily limited to operational curves or costly monitoring data. This presents a significant challenge for experimental education. Digital twin technology, through physical entity modeling, data integration, processing, and feedback mechanisms, enables real-time interaction between the physical and virtual worlds. This offers innovative solutions for teaching hydraulic fracturing visualization.

Unlike static data visualization, hydraulic fracturing visualization should accurately reflect the real-time fracture formation process within the reservoir. This is rarely achieved in current applied research within the field of petroleum engineering. Given the significant gap between the theory and practical application of hydraulic fracturing, we have developed both an indoor and a field experimental virtual system. Separate digital twin platforms were constructed for each system.

The objective of the indoor system is to enable learners to visualize operating curves and monitoring data through hydraulic fracturing simulation experiments, allowing them to apply theoretical knowledge to explain observed phenomena. The aim of the field system is to give learners a deeper understanding of the hydraulic fracturing process in reservoirs through real-world projects. With foundational knowledge gained from indoor experiments, learners can quickly adapt to the field experimental teaching. Ultimately, this approach leads to the achievement of the desired learning outcomes (Figure 1).

2.1. Digital Twin for Indoor Hydraulic Fracturing

2.1.1. Construction of an Indoor Digital Twin Platform for Hydraulic Fracturing

Supported by the geological engineering integration and equipment team of Chengdu University of Technology, we have independently developed a true triaxial hydraulic fracturing physical simulation experimental system. This system is composed of six main components:

(1)

A hydraulic fracturing rock sample loading chamber (placement of actual rock samples);

(2)

A three-directional load hydraulic servo system (simulation of deep reservoir geostress environment);

(3)

A heating system (simulation of deep reservoir temperature environment);

(4)

An acoustic emission-assisted monitoring system (simulation of onsite microseismic monitoring by real-time monitoring and recording of signals generated by rock damage during hydraulic fracturing);

(5)

A fracturing fluid pumping system (simulation of the fracturing fluid injection process, including pressure and flow control, and real-time recording of the pumping curve);

(6)

A control platform (user-friendly interfaces and data processing capabilities for centralized control and monitoring of the above systems).

As shown in Figure 2, the acoustic emission monitoring system and the fracturing fluid pumping system are critical components of the overall system, providing key feedback from hydraulic fracturing experiments. These two systems serve as important data sources for the development of digital twin models used for fracture visualization.

Based on this system, we have integrated three key modules of data connection, data processing, and data feedback using digital twin technology. This integration enables real-time data acquisition and visual analysis of the fracturing process, providing an accurate representation of the entire indoor hydraulic fracturing operation. The main features of the digital twin technology are as follows: (1) the data connection module collects data from each monitoring system; (2) the data processing module employs advanced algorithms and models to analyze the data; and (3) the data feedback module adjusts experimental parameters in real time based on the analysis results, ensuring the accuracy and reliability of the simulation.

In this way, digital twin technology makes it possible to simulate and monitor the actual operation. The real-time visual analysis provided allows instructors to explain hydraulic fracturing concepts more vividly, helping learners understand and grasp the process more intuitively. This system offers an innovative and efficient platform for indoor experimental teaching. Compared to traditional text-based methods, the novel visualization technology effectively enhances learners’ interest and attention and improves their practical ability in fracturing testing.

2.1.2. Digital Twin-Based Visualization of Indoor Hydraulic Fracturing

By integrating digital twin technology with the simulation experiment system, we have developed a visualization technique for indoor hydraulic fracturing experiments. The methodology employed in the fracturing experiment is as follows: (1) Place the rock sample (400 mm in size) inside the loading chamber and install 16 acoustic emission monitoring probes along with data transmission lines. (2) Use the control system platform to apply temperature and stress to the rock sample, simulating the actual reservoir environment. (3) Inject fracturing fluid into the rock sample using the pumping device while simultaneously activating the acoustic emission monitoring system. (4) Record the pumping curve and acoustic emission monitoring data in real time using the data acquisition equipment.

Interpreting acoustic emission data to identify fractures presents a significant challenge, and this stage is critical for the real-time visualization of hydraulic fracturing experiments. In response to this challenge, our team has developed a fracture fitting technique based on acoustic emission data (Figure 3):

(1)

Removal of noise points using the DBSCAN algorithm of density space-based data clustering

The algorithm of density-based spatial clustering of applications with noise (DBSCAN) divides regions with sufficient density into clusters and discovers arbitrarily shaped clusters in a spatial database with noise. It defines a cluster as the largest set of densely connected points.

In hydraulic fracturing, fracture propagation results in both effective and ineffective responses. Ineffective responses typically occur due to natural fracture slippage in regions far from the well, where the resultant event points appear isolated and discrete. Both effective and ineffective responses are produced by the same mechanism, and noise reduction cannot be directly achieved through waveform or frequency analysis. To address this, we introduce the DBSCAN algorithm. This algorithm estimates the density of each data point by analyzing its neighborhood, allowing for the calculation and analysis of spatial density within arbitrarily distributed sets of spatial points. The primary goal is to remove acoustic emission noise points, thereby enhancing the accuracy of fracture identification.

(2)

Fitting fracture surfaces using the RANSAC algorithm

The algorithm of random sample consensus (RANSAC) uses an iterative approach to estimate the parameters of a mathematical model from a set of observations that contain outliers. Compared to the least squares algorithm, it incorporates the idea of eliminating outlier data and, therefore, gives faster and more accurate identification results for data samples with some erroneous data.

For effective microseismic events, accurate extraction of fracture information and fitting of the fracture surface is critical. The RANSAC algorithm is adopted to fit the fracture surface. Firstly, a randomly selected subset of the spatial point set is selected based on its ability to satisfy the fit equation. Then, an estimation of the fracture parameters is derived through experimental means. The entire set of data points is evaluated for compliance with the criteria. This process is repeated until the result meets the specified threshold metrics, ensuring an accurate representation of the fracture surface.

(3)

Fracture shape recognition by alpha shape

Fracture orientation and shape parameters are also essential for accurately fitting the final fracture surface. When representing the fracture surface morphology, the number and distribution of vertices determine how well the fracture shape corresponds to the microseismic event points. The alpha-shape algorithm provides an efficient and straightforward approach for quickly extracting boundary points. For a planar point cloud with an arbitrary shape, the use of a rolling disk along the outer boundary helps generate the desired fracture shape.

The algorithms described above are being successfully applied to build a digital twin technology for the real-time visualization of hydraulic fracturing experiments. Figure 4 illustrates our visualization of the display process. In the left-hand figures, the curves represent the pressure, while the “O/X” symbols denote rock damage events, all of which were recorded in real time. Using digital twin technology, these damage events were utilized to model and visualize the formation of fractures within the rock, as depicted in the right-hand figures. Compared to traditional teaching methods, the experimental practice teaching approach based on digital twin technology significantly expands instructional resources and optimizes the effectiveness of the teaching process. Providing effective support is crucial to helping learners achieve their learning outcomes. This approach plays a vital role in ensuring learners reach their desired educational goals.

2.2. Digital Twin for Onsite Hydraulic Fracturing

2.2.1. Construction of an Onsite Digital Twin Platform for Hydraulic Fracturing

Unlike indoor experiments, the effectiveness of onsite hydraulic fracturing is primarily evaluated through pumping profiles. Although monitoring tools like microseismic and fiber optics can be introduced, their high cost makes them impractical. After extensive research, our team has developed a digital twin platform that employs numerical simulation and machine learning as its core technologies, enabling real-time visualization of onsite fracturing.

The platform incorporates two key technologies: (1) a numerical simulation method for hydraulic fracturing that optimizes both computational efficiency and simulation accuracy, and (2) machine learning algorithms for analyzing fracturing curves that strike a balance between training efficiency and prediction accuracy. The development of these techniques requires not only a deep understanding of hydraulic fracturing theory but also expertise in numerical computation, machine learning, and data programming. To achieve the anticipated outcome of onsite experimental instruction, we have developed an integrated process platform employing digital twin technology. The instructor simply inputs the necessary data, while the platform handles the real-time visualization of the fracture formation process. During the experimental teaching process, the instructor can explain how changes in the pumping pressure curve relate to fracture propagation. Compared to traditional methods that rely solely on pump pressure curve information, this visualization-based approach is more efficient and intuitive, allowing learners to understand better how theoretical knowledge applies to real-world engineering scenarios. This method significantly enhances learner comprehension and optimizes the effectiveness of onsite experimental teaching.

2.2.2. Digital Twin-Based Visualization of Onsite Hydraulic Fracturing

Onsite visualization of fracturing includes the following key issues (Figure 5):

(1)

Construct an accurate geological model based on reservoir geological information, including natural fractures, rock properties, and in-situ stress.

(2)

Develop a numerical model of hydraulic fracturing and input the relevant operational parameters.

(3)

Perform multiple fracturing simulations and build a database of pumping curves.

(4)

Develop a machine learning algorithm to fit onsite pumping curves in real time, while constraining key variables.

(5)

Input the determined variable parameters and obtain real-time fracture morphology through fracture simulation.

In this process, steps 1–3 are completed prior to the onsite experimental sessions, while steps 4–5 are carried out during the onsite experiments. Our team has successfully developed the corresponding techniques and systems to support this workflow.

For the efficient and accurate numerical system of hydraulic fracturing, we employ the displacement discontinuity numerical method. This method calculates the displacement and stress distributions on the fracture surface with high computational efficiency, making it well-suited for simulating multi-fracture propagation at an engineering scale. The system supports key functions, including proppant transport, fracture height expansion, the interaction between hydraulic and natural fractures, fracturing fluid leakage, and changes in geostress.

For the efficient and accurate machine learning algorithms (Figure 6), we utilize an intelligent optimization algorithm with regularized parameters, where the optimization function serves as the model, and the objective function value represents the actual response. By combining this with the JAYA algorithm, the parameters are optimally trained. This approach significantly reduces the computational time of the algorithm and inversion studies while also improving inversion accuracy.

Building on these key technologies, we have successfully developed a real-time visualization platform for onsite hydraulic fracturing. The visualization process is illustrated in Figure 7. Compared to traditional methods, the onsite experimental teaching system based on digital twin technology significantly improves the integration of theory and practice. This approach is crucial in helping learners apply their learning to real-world engineering practice.

3. Design of Hydraulic Fracturing Experimental Teaching Based on Digital Twin Technology

3.1. Design of a Pilot Teaching Subgroup for an Experimental Course on Hydraulic Fracturing

In engineering education, effectively supporting learners in achieving the desired learning outcomes is crucial. Using the digital twin platforms we developed for both indoor and onsite hydraulic fracturing, the teaching team designed two types of programs: indoor experiments and onsite experiments. These courses were offered during the summer holidays, and pilot classes were conducted to test the effectiveness of the new experimental teaching systems.

This teaching initiative was undertaken by our team, independent of the college. It also required both the time and willingness of the learners to participate. We recruited a total of only 20 third-year undergraduate students from various engineering disciplines, including civil engineering, mechanical engineering, petroleum engineering, and geological engineering. The undergraduate courses for these four majors at Chengdu University of Technology were relatively similar. Fortunately, the GPA of the participants ranged from 3 to 3.5. Our primary consideration during the grouping process was to ensure that each group included students from different majors. Ultimately, the students were divided into four groups of five.

The twenty-day program included hydraulic fracturing theory, indoor experimental teaching and onsite experimental teaching. For the four groups of learners, we applied different teaching methods. By comparing learning effects and learner feedback, we were able to confirm the feasibility and advantages of the digital twin-based experimental teaching approach. The learning arrangements for these four groups are shown in

Table 1. Courses of the four working groups.

Group Number	Theory Course	Indoor Experimental Course	Onsite Experimental Course
			Traditional Mode	Newly Mode
Group 1	(1) Course in rock mechanics (2) Hydraulic fracturing technical course	/	/	/
Group 2	(1) Course in rock mechanics (2) Hydraulic fracturing technical course	(1) Rock mechanics laboratory course (2) MATLAB programming	/	/
Group 3	(1) Course in rock mechanics (2) Hydraulic fracturing technical course	(1) Rock mechanics laboratory course (2) MATLAB programming	Hydraulic fracturing tutorial	/
Group 4	(1) Course in rock mechanics (2) Hydraulic fracturing technical course	(1) Rock mechanics laboratory course (2) MATLAB programming	/	(1) Fracturing simulation (2) Hydraulic fracturing tutorial

:

3.2. Process of Pilot Teaching for an Experimental Course on Hydraulic Fracturing

Since theoretical knowledge is fundamental to practical application, we arranged five theory lectures for all four groups of students. Lecture 1 introduced hydraulic fracturing, while Lectures 2–3 covered rock mechanics. Lectures 4–5 focused on hydraulic fracturing technology, primarily addressing the basic concepts and process technology. However, it was challenging to fully familiarize the students with hydraulic fracturing in just two lectures.

For the indoor experimental teaching, we arranged four lectures. The first lecture introduced the basic rock mechanics experiment. The second lecture covered the hydraulic fracturing experiment, including the true tri-axial hydraulic fracturing system and the simulation of hydraulic fracturing in the field. The third lecture introduced the digital twin technology developed for indoor hydraulic fracturing. The fourth lecture guided the learners through completing a hydraulic fracturing experiment. Through indoor experiments and practical teaching, learners gained an intuitive understanding of fracturing basics, with particular emphasis on the process of fracture propagation in the rock. During fracturing, pumping curves and acoustic emission data were recorded in real time. The extension of the fractures within the rock sample was observed using the digital twin visualization technique (Figure 4). At the end of the experiment, the learners observed the actual morphology of the fractures produced. After completing the indoor experimental teaching, learners consulted relevant literature and applied theoretical knowledge to explain the experimental phenomena. Finally, the understanding of hydraulic fracturing technology was improved. The laboratory report was completed.

The third group of learners proceeded directly to the fracturing control room for the in-situ experimental teaching. Using the pumping curve, the instructor introduced the process of fracture construction in the field. The possible occurrence of underground fracture propagation through the changes in each parameter curve was explained. While the pump curves contain valuable information, the traditional mode of experimental teaching was often monotonous, leading to low learner engagement and participation.

Since the digital twin of the onsite fracturing experiment was developed using numerical simulation, it was essential to provide the fourth group of students with an introduction to this technique. Before taking them to the onsite fracturing control room, we arranged two lectures to explain the hydraulic fracturing simulation technique. The learners were instructed to complete an onsite fracturing simulation and discussed the fracture propagation by adjusting different parameters. They then completed a report on the numerical simulation of onsite reservoir fracturing. By combining the five reports, we could get a variety of working conditions. This would help us to deepen our understanding of the subsequent experimental learning in the field. In the fracturing control room, we connected the digital twin platform developed for onsite hydraulic fracturing to the fracturing operating system. Compared to the traditional mode, this approach integrated fracture propagation visualization, allowing us to explain the pumping pressure curve in a more graphical and intuitive manner. For example, when artificial fracture propagation was restricted, the pressure curve rose (Figure 7), and when the artificial fracture propagated along a natural fracture, the pressure curve dropped (Figure 7). As the learners had completed the report on the numerical simulation of the onsite fracturing, they could try to make their own proposals based on the on-site fracturing situation. Predictions of the extent of the fracture could also be made. This added a lot of interactivities to the onsite experimental teaching.

Throughout the teaching process, we carried out questionnaires with four groups of learners. These surveys were essential for capturing the learners’ learning progress and their cognitive processes in relation to the course content. By communicating with the learners, we were able to assess whether the new experimental teaching systems were effective in helping them achieve the desired learning outcomes. This feedback enabled us to implement a continuous improvement process for future teaching.

At the end of the course, we developed both theoretical and experimental evaluations. The theoretical assessment focused on fundamental concepts and tested learners’ mastery of theoretical knowledge. The experimental assessment evaluated their ability to apply this knowledge through a typical onsite fracturing case. For example, how to identify the fracture propagation through the pumping curve and how to optimize the design parameters to achieve a better fracturing effect. Learners’ learning outcomes were determined according to the assessment results. Finally, the advanced experimental teaching system based on digital twin technology was demonstrated.

4. Evaluation and Analysis of Hydraulic Fracturing Experimental Teaching Based on Digital Twin Technology

4.1. Evaluation of Questionnaires Throughout the Course

4.1.1. Theory of Hydraulic Fracturing Training

To assess the learners’ level of knowledge about hydraulic fracturing, we surveyed four groups of learners before the class. The results showed that 80% of the students had never heard of hydraulic fracturing. Although 20% of the students (primarily those studying petroleum engineering) had heard of the technology, their knowledge was limited to understanding that it is an effective method for increasing oil and gas production.

At the end of the theory lesson, we chose not to conduct an immediate assessment. Instead, we engaged in an interactive session with the students, focusing on three key points: (1) Whether they were still interested in continuing the course on hydraulic fracturing; (2) whether they felt they were able to learn it properly; and (3) suggestions for improving the course. Encouragingly, 80% of learners remained interested in the course. However, only 40% felt confident in their ability to grasp the material fully (

Table 2. Questionnaire survey after theory training.

Group Number	Are You Interested in the Course?		Can You Learn the Course Well?
	Interesting	Confuse	Confident	Lack of Confidence
Group 1	4	1	2	3
Group 2	3	2	1	4
Group 3	5	0	2	3
Group 4	4	1	3	2

).

The learners also provided valuable suggestions for improving the course. For example, they expressed a need for more lessons, increased onsite teaching, and the inclusion of more animations in the curriculum. Overall, we concluded that hydraulic fracturing was a difficult course. Most learners did not perform well after a brief period of theoretical study. They also recognized that for engineering courses, experimental learning is just as important as theoretical instruction.

4.1.2. Indoor Experimental Teaching

Before the indoor experimental lesson, we had a discussion with the learners (except for Group 1). Given the difficulty of learning hydraulic fracturing theory, 45% of the learners hoped the indoor experiment would complement theoretical learning. In addition, 40% hoped the indoor experiment would bridge the gap between theoretical learning and practical engineering. The remaining learners had little expectation for experimental learning.

After completing the indoor experimental course, we continued face-to-face communication with the learners. The topic was what they had gained from the experimental learning process. The main feedback from the learners was as follows: (1) Most learners felt that the experimental learning had deepened their understanding of the theoretical concepts; (2) some learners found the experimental visualization very interesting and beyond their expectations and expressed interest in applying it to other subjects (e.g., civil engineering students); (3) a few learners felt that their weak theoretical knowledge limited their understanding of the experimental results.

According to the questionnaire survey after the indoor experimental teaching, the number of learners in Groups 2–4 who expressed interest in the course or felt confident in learning the course well had not increased significantly (

Table 3. Questionnaire survey after indoor experimental teaching.

Group Number	Are You Interested in the Course?		Can You Learn the Course Well?
	Interesting	Confuse	Confident	No Confidence
Group 1	/	/	/	/
Group 2	3	2	2	3
Group 3	5	0	2	3
Group 4	4	1	4	1

). To further improve the teaching effect of experimental practice, we plan to introduce experimental tutorials in the future. These tutorials will help students become more familiar with the experimental equipment and procedures. Additionally, more advanced digital twin technology will be incorporated to improve teaching outcomes. Overall, learners recognized the importance of the indoor experimental course, as it helped them consolidate and deepen their theoretical knowledge and better prepare for fieldwork.

4.1.3. Onsite Experimental Teaching

We carried out the onsite experimental teaching with two groups of learners, with the third group using the traditional platform and the fourth group using the new platform. Both groups of learners were given a precourse questionnaire. Both groups completed a precourse questionnaire, expressing their hopes that the onsite experiments would help them gain a deeper understanding of the theory and explain certain phenomena encountered in real engineering. Most learners showed a higher interest after the onsite experimental lessons.

After completing the onsite experimental course, we communicated with the two groups. The third group provided different feedback compared to the fourth group. Learners in the traditional mode group reported little benefit from the field. Beyond gaining a basic understanding of the process, their knowledge of fracturing was limited to indoor experimental learning. Due to their limited expertise, they found the traditional pump curves difficult to interpret and the sessions unengaging. As a result, the number of learners in Group 3 who felt confident in mastering the course remained unchanged (

Table 4. Questionnaire survey after traditional onsite experimental teaching.

Group Number	Are You Interested in the Course?		Can You Learn the Course Well?
	Interesting	Confuse	Confident	No Confidence
Group 1	/	/	/	/
Group 2	/	/	/	/
Group 3	5	0	2	3
Group 4	/	/	/	/

).

In contrast to the traditional mode group, the learners in the new mode group showed greater initiative and motivation, aided by the visualization platform. They felt that the visualization technology helped bridge the gap between their understanding and the complexities of onsite fracturing. Finally, all the learners in Group 4 expressed interest in the course, with only one learner lacking confidence in learning the course well (

Table 5. Questionnaire survey after new onsite experimental teaching.

Group Number	Are You Interested in the Course?		Can You Learn the Course Well?
	Interesting	Confuse	Confident	No Confidence
Group 1	/	/	/	/
Group 2	/	/	/	/
Group 3	/	/	/	/
Group 4	5	0	4	1

).

Overall, introducing a new mode using digital twins in onsite experimental teaching can significantly improve learners’ interest and confidence, particularly for those who felt confused after the theoretical lessons. Several benefits have been demonstrated, including more engaging course content, a wider variety of teaching methods, and reduced learning difficulty. This approach is crucial in helping learners achieve excellent outcomes.

4.2. Assessment of Theoretical and Experimental Teaching

In the theory test, we found that the results of the four groups did not differ significantly. Because experiential teaching may help learners better grasp knowledge related to actual hydraulic fracturing, learners in Groups 2–4 may have been better equipped to answer open-ended questions than those in Group 1. In the experimental assessment, the four groups showed varying levels of performance. For example, in response to the question, “How to increase the stimulation extent of a reservoir” (

Table 6. Assessment of experimental teaching with “How to increase the stimulation extent of reservoir”.

Group Number	Response from Learners	Evaluation from Instructors
Group 1	Focus mainly on the single fracture.	The ideas were more theoretical. The proposed solutions were detached from engineering practice.
Group 2	Extend to multiple fractures and attempt an operational perspective.	The ideas required further transformation into field parameters.
Group 3	Optimize the process from the engineering point of view.	The ideas were mostly the result of conventional thinking and were not sufficiently novel.
Group 4	Optimize the process not only from a technical point but also in terms of fracture propagation.	Though the ideas were a little immature, they represented a match between theory and practice.

), the fourth group, who used digital twin visualization, demonstrated a deeper understanding of fracture propagation. Their responses were more detailed and exhibited a higher level of expertise compared to the other three groups. Overall, the development of digital twin-based hydraulic fracturing visualization technology played a crucial role in enhancing onsite experimental teaching.

4.3. Feedback on Teaching

At the end of the course, we asked the learners for feedback, and two key themes emerged. The first theme was excitement about using visualization techniques in the experimental courses. They expressed a desire to learn more and even pursue further education. Some of their comments included: “The digital twins are a great technique that will help me succeed in the courses”, “I want to explore more about visualization techniques”, and “I hope other courses will also adopt visualization platforms”. The second theme focused on more constructive suggestions. Some of the comments included: “Digital twins are interesting, but I think there should be more courses”, “With the current visualization technology, we are users and I’d like the opportunity to develop it”, “I think the course is quite intense and I would recommend to slow down the pace so that we can digest the learning of the course in a better way”. We concluded that the digital twin visualization technique helps boost learner motivation to tackle complex subjects. It is our hope that learners will push themselves to become more independent, creative, and lifelong learners.

5. Discussion

5.1. Advantages of Digital Twins in Experimental Teaching

The course represented by hydraulic fracturing is highly specialized and integrated. We introduce the digital twin technology for experimental teaching. Then, the learners’ ability to master knowledge and information gathering can be improved to a certain extent. It is hoped to achieve effective learning results.

The initial findings showed that the learners were, in general, intimidated by the practical applications and the digital technologies in their previous learning. Many were skeptical of their own abilities. After completing the various stages of the task, the majority of students significantly improved both their self-confidence and their confidence in future career development. It is particularly noteworthy that none of the learners recruited had any basic knowledge of programming and numerical simulation prior to the experimental teaching. In the more specialized experimental teaching, digital twin technology proved effective in helping learners overcome cognitive challenges—something that was difficult to achieve with traditional teaching methods.

The results of the project case showed that after active participation in the experimental course with the new mode, learners had more initiative and were able to experiment with a variety of analysis methods. For example, the optimization process was investigated using fracture extension visualization and pumping curves. While some of the suggestions may be inappropriate, our new technology achieves the goal of encouraging learners to think. This is a quality that is often lacking in our learners. Compared to the traditional mode, the introduction of digital twin technology can engage learners in active learning and improve their overall quality.

Digital twin technology provides an intuitive and interactive learning platform that helps learners understand complex technical problems. While it also simplifies the process for instructors to impart knowledge, it places greater demands on them to stay updated with new technological developments. This teaching approach plays a crucial role in the training of skilled engineers.

5.2. Limitations of the Research and Directions for Improvement

Since participation in this study was voluntary, we cannot rule out the possibility that some students approached it with a superficial attitude. Additionally, the study was conducted with a relatively small group of students from the same school, who had varying professional backgrounds and learning abilities. Therefore, the results of the pilot lessons cannot be considered generalizable. Furthermore, due to the short duration of the course and the numerous assignments across different stages, the assessment results may not fully reflect the overall abilities of some learners.

A digital twin-based hydraulic fracturing experimental course places higher demands on the university. The university needs to acquire large-scale equipment, provide an open numerical simulation platform, and recruit instructors with experience in engineering digitalization. In addition, instructors need to improve their own learning, including new teaching modes and new knowledge.

To better promote digital twin-based experimental courses, we will continue to improve in the following areas:

Optimization of teaching resources: Using digital twin technology to provide richer teaching resources, such as virtual experiments and interactive courses. The aim is to increase learner interest and participation in learning. We are also exploring the application of this platform to laboratory courses in various disciplines, including physics, chemistry, and mechanical engineering. This will help learners better understand complex concepts through immersive virtual environments.

Personalized teaching of learners: Providing personalized content and tutoring based on the learner’s knowledge level and learning needs. The aim is to cultivate more highly skilled professionals for the oil and gas industry. This includes designing experiments appropriate for different levels of learners, such as basic experiments for undergraduates and advanced simulations for graduate students. This approach will establish a more human-centered technology, emphasizing collaboration between people and technology [36].

Professional training of instructors: Training instructors in new teaching methods, including organizing experimental sessions, teaching, and assessment, is essential to ensure that the visualization-based experimental courses run smoothly. The goal is to equip instructors with the necessary skills to implement this new teaching mode effectively.

Interuniversity cooperation and support: Through seminars or academic exchanges, we introduce the platform to other universities and encourage institutions to promote its use. Additionally, we seek to collaborate with other universities and research institutions to establish virtual laboratories based on digital twin technology, fostering joint research and teaching efforts.

Expansion to onsite engineering departments: We plan to train field engineers to promote the application of the platform in real-world projects. This initiative aims to achieve more efficient and economical oil and gas production.

6. Conclusions

In this paper, the experimental teaching of deep energy exploitation based on the digital twin technology is discussed. We focus on the development and application of a real-time visualization teaching system for hydraulic fracturing. The results of the research show that as follows:

(1)

Using hydraulic fracturing technology as the teaching focus, we have developed a platform that leverages digital twin technology to support both indoor and onsite experimental teaching. The indoor component emphasizes theoretical analysis, while the onsite component focuses on real-world simulation. This approach enables the visualization of fracture propagation during hydraulic fracturing.

(2)

The introduction of digital twin technology has significantly enhanced the teaching of hydraulic fracturing courses. By integrating virtual and real-world interactions, learners can intuitively grasp complex theoretical concepts, reducing their hesitation to engage in practical engineering applications.

(3)

Based on the analysis of questionnaire data, digital twin technology has transformed learners into active participants in the learning process. Teaching evaluation indicates that, under the new experimental teaching mode, learners demonstrate a higher level of integration between theory and practice. Their engineering application and innovation skills have also significantly improved.

(4)

The results of the pilot demonstrate that the digital twin technology effectively helps learners overcome anxieties about complex engineering courses. It plays a crucial role in boosting their self-confidence for their future careers. However, due to the limitations of the research, we need to improve the experimental course, including refining the course design and enhancing professional training.