A method for constructing a self-adaptive cutting guide and control system of a coal mining machine based on digital twinning

Abstract

The application provides a coal mining machine adaptive cutting guide control system construction method based on digital twinning, and relates to the technical field of coal mining machine adaptive cutting. The method specifically comprises the following steps: designing a complex gangue-enclosed thin coal seam coal mining machine cutting state sensing method and a complex gangue-enclosed thin coal seam coal mining machine adaptive control strategy; based on the complex gangue-enclosed thin coal seam coal mining machine cutting state sensing method and the complex gangue-enclosed thin coal seam coal mining machine adaptive control strategy, a complex gangue-enclosed thin coal seam coal mining machine adaptive cutting digital twinning model is established, which is used for simulating the adaptive cutting process of the complex gangue-enclosed thin coal seam coal mining machine; a numerical model driven mechanism characteristic life prediction model and a complex gangue-enclosed thin coal seam coal mining machine adaptive cutting dynamic reliability digital twinning monitoring model are constructed, which are used for real-time evaluation of the reliability of the adaptive cutting process of the complex gangue-enclosed thin coal seam coal mining machine. The application can effectively improve the adaptive ability of the coal mining machine to the complex gangue-enclosed thin coal seam.

Description

Technical Field

[0001] This invention relates to the field of adaptive cutting technology for coal mining machines, and in particular to a method for constructing an adaptive cutting guidance and control system for coal mining machines based on digital twins. Background Technology

[0002] To promote the intelligent construction of coal mining processes, the status of coal mining machine face equipment clusters can gradually achieve transparent monitoring. However, for the mining process of thin coal seams with complex and diverse occurrence conditions, the intelligent safety decision-making throughout the entire process and the autonomous judgment of high-reliability operation throughout the entire life cycle of the coal mining machine still require online participation from operators. Moreover, due to the spatial characteristics of the geological structure of thin coal seams, the adaptive guidance and control cutting system of thin coal seam coal mining machines places higher demands on the collaborative knowledge of multiple disciplines such as mechanics, electrical, hydraulics, and control, as well as the fusion performance of multi-source heterogeneous data information. Digital twins, based on the modeling of multi-source heterogeneous data, are a new concept and system architecture for controlling physical processes. Its main feature is the adaptive capability and flexibility of the elements closely related to the physical process changing with the changes in the physical process. This is crucial for the research and application of power transmission systems for intelligent coal mining machines in complex interbedded thin coal seams. Therefore, in order to achieve intelligent, safe and efficient coal mining, the technical architecture and key technologies of intelligent mining faces based on digital twins, the construction of system frameworks and the collaborative evolution of virtual and real data have been proposed one after another. However, there is still a lack of in-depth research and effective implementation methods for the control of virtual and real collaboration of power transmission system of coal mining machine in complex thin coal seams and dynamic high-reliability operation.

[0003] The dynamic process of intelligent and efficient cutting and breaking of interbedded coal and rock by coal mining machines is characterized by non-equilibrium, nonlinearity, time-varying nature, and strong coupling. Furthermore, constructing a digital twin model of the power transmission system for intelligent cutting of interbedded coal and rock by thin coal seams involves the integration of knowledge from multiple disciplines. Currently, the main bottleneck is that the adaptive improvement of thin coal seam coal mining machines with high cutting efficiency is influenced by various heterogeneous information, motion parameters, and dynamic control strategies of the control system. The complex and diverse occurrence conditions of thin coal seams and the weak structure exhibited by the coal mining machine in response to its environment lead to significant differences in the coupling relationships among the physical model, virtual model, decision model, and interaction model. Summary of the Invention

[0004] To address the shortcomings of the existing technologies, this invention proposes a method for constructing an adaptive cutting and control system for coal mining machines based on the concept of digital twin technology. The aim is to utilize the dynamic system characteristics of the coal mining machine during the adaptive cutting process to develop an intelligent power transmission system and explore new fields and applications for intelligent coal mining machines.

[0005] This invention proposes a method for constructing an adaptive control system for cutting complex interbedded thin coal seams using a coal mining machine based on digital twins, comprising:

[0006] Design a cutting state perception method and an adaptive control strategy for a coal mining machine with complex interbedded thin coal seams.

[0007] Based on the cutting state perception method and adaptive control strategy of the coal mining machine for complex interbedded thin coal seams, an adaptive cutting digital twin model of the coal mining machine for complex interbedded thin coal seams is established to simulate the adaptive cutting process of the coal mining machine for complex interbedded thin coal seams.

[0008] Based on the adaptive cutting digital twin model of the coal mining machine in complex thin coal seams with interbedded gangue, a digital twin model driven mechanism characteristic life prediction model and a dynamic reliability digital twin monitoring model of the adaptive cutting of the coal mining machine in complex thin coal seams with interbedded gangue are constructed to evaluate the reliability of the adaptive cutting process of the coal mining machine in complex thin coal seams with interbedded gangue in real time.

[0009] Furthermore, the specific details of the method for sensing the cutting status of a coal mining machine in complex interbedded thin coal seams are as follows:

[0010] Virtual simulation experiments based on the cutting vibration mechanism of thin coal seam mining machines were conducted to simulate the cutting process of thin coal seam mining machines under different combinations of coal-rock ratios, coal-rock structures, and coal-rock hardness. Based on the simulation results, a nonlinear correlation model between the coal-rock cutting state and the vibration signal of the thin coal seam mining machine was constructed. The nonlinear correlation model between the coal-rock cutting state and the vibration signal of the thin coal seam mining machine was represented by a one-dimensional vibration acceleration curve.

[0011] The one-dimensional vibration acceleration curve was converted into a two-dimensional time-spectrum image using an information transformation customized model. The cross-term features of the two-dimensional time-spectrum image were extracted using data analysis. The physical meaning of the two-dimensional time-spectrum image was verified by comparing the extracted cross-term features with the physical phenomena in the coal mining machine cutting process.

[0012] Using the feature information processing module, the two-dimensional time-spectrum image processed by the information transformation customized model is fused with the multi-source data information image that conforms to the structural change distribution of arbitrarily complex interbedded coal seams, and the generated multi-source data information fused image is used to build a basic data sample library for cutting state recognition.

[0013] A cut-off state classification and recognition model was constructed using a basic data sample library for cut-off state recognition.

[0014] Furthermore, the specific content of the adaptive control strategy for the complex interbedded thin coal seam mining machine is as follows:

[0015] Define the performance indicators of coal mining machines for complex interbedded thin coal seams, including: cutting area, productivity, cutting ratio energy consumption, cutting resistance, load fluctuation coefficient, and coal loading rate;

[0016] Define the motion parameters of a coal mining machine for complex interbedded thin coal seams, including: traction speed v. q and rotational speed n;

[0017] A multi-task Gaussian algorithm was used to fit the fitting surfaces and fitting errors of each performance index and the motion parameters of the coal mining machine, so as to obtain the functional equations of each performance index and the motion parameters of the coal mining machine and construct a comprehensive performance index evaluation model.

[0018] Based on the actual working conditions, the constraints of the motion parameters of the coal mining machine are defined. Based on the auxiliary design and load calculation software of the drum of the thin coal seam coal mining machine, the comprehensive performance index evaluation model is divided into several regions along the cutting direction. In each region, the piecewise fitting method is used to approximate the constraints of the motion parameters of the coal mining machine.

[0019] The optimization objective is defined as maximizing cutting area, productivity, and coal loading rate, while minimizing cutting specific energy consumption, cutting resistance, and load fluctuation coefficient. Based on the approximation ideal solution ranking theory, the combination of motion parameters of the coal mining machine under different working conditions is used as the evaluation object, and performance indicators are used as evaluation indicators. An n*m decision matrix between the evaluation object and the evaluation indicators is established and then normalized and standardized to generate the optimal combination of motion parameters of the coal mining machine under different working conditions; where n represents the number of evaluation objects and m represents the number of evaluation indicators.

[0020] A mechanical-electrical-hydraulic-control coupling model is built based on the optimal combination of motion parameters of the coal mining machine under different working conditions. The control sequence is designed based on the built mechanical-electrical-hydraulic-control coupling model. The control sequence is as follows: the control strategy of the coal mining machine is: the drum speed control model takes priority over the traction speed control model, the traction speed control model takes priority over the drum speed control model, and the coordinated control model.

[0021] A model is established to correlate the control strategy with the coal seam structure and motion parameters based on the designed control sequence, which is used to automatically select the optimal control strategy based on the coal seam structure and motion parameters.

[0022] A deep reinforcement learning method is used to optimize the model that correlates the control strategy with the coal seam structure and motion parameters, resulting in an adaptive control model. This model is used to generate an adaptive cutting optimal control strategy based on changes in actual working conditions.

[0023] Furthermore, the adaptive cutting digital twin model of the complex interbedded thin coal seam mining machine includes: a physical perception layer, a virtual simulation layer, a twin decision layer, and a data interaction layer;

[0024] The physical sensing layer is used to construct the physical structure system of the coal mining machine, and to intelligently monitor the physical structure system of the coal mining machine by designing multiple physical sensors, so as to obtain the working status information of the coal mining machine in complex interbedded thin coal seams in real time.

[0025] The virtual simulation layer is used to construct a high-precision three-dimensional twin model of a dynamic, complex, thin coal seam with interbedded gangue. By combining the real-time working status information of the coal mining machine in the complex, thin coal seam with interbedded gangue, the adaptive control model is used to generate an adaptive cutting optimal control strategy, which is used to adaptively adjust and control the action of the physical structure system of the coal mining machine in the physical perception layer.

[0026] The twin decision layer is used to optimize the cutting process of the coal mining machine in complex interbedded thin coal seams by using the cutting state classification and recognition model to identify the cutting state of the coal mining machine in real time based on the working state information obtained by the physical perception layer and the dynamic complex interbedded thin coal seam high-precision three-dimensional twin model constructed by the virtual simulation layer.

[0027] The data interaction layer is used to transmit and interact data between the physical perception layer, the virtual simulation layer, and the twin decision layer, so as to realize the dynamic adjustment of adaptive cutting of the coal mining machine in complex interbedded thin coal seams.

[0028] Furthermore, the physical sensing layer is constructed as follows: the physical sensing layer comprises three parts, namely: a coal mining machine physical structure system, a multi-physical sensor design, and a data acquisition system; wherein the coal mining machine physical structure system is a coal mining machine adaptive cutting control integrated test bench designed based on similarity theory and cutting experiments;

[0029] The design method of the multi-physical sensor is as follows: By analyzing the load-bearing characteristics of key components of the coal mining machine's cutting section through dynamic analysis, the sensing force points are determined, and the layout of the multi-physical sensor is designed based on these force points. The sensors are arranged according to this layout, and based on the method for sensing the cutting state of a coal mining machine in complex interbedded thin coal seams, a spiral drum cutting vibration sensing system, a square head vibration sensing system, and a rocker arm vibration sensing system are constructed. These systems are used to acquire the attitude parameter information of the coal mining machine in complex interbedded thin coal seams in real time. Furthermore, the multi-source data information collected in real time by each physical sensor during the coal mining machine's cutting process is used to identify the cutting state of the thin coal seam coal mining machine, thereby achieving dynamic monitoring of the coal mining machine's cutting process in complex thin coal seams. The attitude parameter information includes: traction speed, rotational speed, and the height of the rocker arm adjustment.

[0030] The attitude parameter information of the complex interbedded thin coal seam mining machine and the multi-source data information collected by multiple physical sensors are used; wherein the attitude parameter information is used as the working status information of the complex interbedded thin coal seam mining machine.

[0031] The data acquisition system is used to monitor and record the operating parameters of the coal mining machine adaptive cutting control integrated test bench and the working status data collected in real time by multiple physical sensors.

[0032] Furthermore, the method for constructing the virtual simulation layer is as follows: coal and rock samples are collected from typical working faces and their physical and mechanical properties are tested. The physical property parameters of the coal and rock mass based on discrete element method are determined and a discrete element model of the coal and rock mass is constructed. Based on the application programming interface (API) technology, the discrete element model of the coal and rock mass is corrected by compiling a replacement API for multiple particle type clusters. A high-precision three-dimensional twin model of dynamic complex interbedded thin coal seam with real-time correction and replacement of particle sets is constructed.

[0033] Based on the attitude parameter information of the coal mining machine in the complex interbedded thin coal seam acquired in real time, the cutting action of the coal mining machine is simulated and the dynamic complex interbedded thin coal seam particle set is corrected and replaced in real time. Based on the adaptive control strategy of the coal mining machine in the complex interbedded thin coal seam, an adaptive control model is constructed using multi-source data information collected by multiple physical sensors. The adaptive control model is used to generate the optimal adaptive cutting control strategy of the coal mining machine. The adjustment action of the coal mining machine is virtually executed according to the generated optimal adaptive cutting control strategy, and the expected working state information is generated.

[0034] The virtually executed adjustment actions are fed back to the physical perception layer to guide the physical structure system of the coal mining machine to perform actions, obtain the working status information after the action, evaluate the effectiveness of the adaptive cutting optimal control strategy by comparing the expected working status information and the working status information after the action, adjust the adaptive cutting optimal control strategy according to the evaluation results, and then virtually execute the adjustment action of the coal mining machine again to form a closed-loop control process.

[0035] A rigid-flexible coupled virtual prototype model of a coal mining machine was established to cut coal and rock. The hydraulic system and electrical control system of the coal mining machine were integrated into the virtual simulation layer. Interface technology was used to realize the interconnection between the models, determine the input and output of each model, and perform joint simulation to simulate the cutting process of a coal mining machine in a complex thin coal seam with interbedded gangue.

[0036] Furthermore, the construction method of the twin decision layer is as follows: using the discrete element model of coal and rock mass constructed in the virtual simulation layer as the basis, the test model of coal mining machine cutting coal seam is constructed by changing the position, thickness and rock properties of the interbedded rock layer in the discrete element model of coal and rock mass in the virtual simulation layer. The test model is then used to analyze the stress law, vibration characteristics and cutting tooth state of the spiral drum cutting coal and rock in the coal mining machine. A process model of coal and rock crushing, collapse, flow and interface pressure formation during the coal mining machine system cutting interbedded coal and rock is also constructed for coal and rock crushing, collapse, flow and interface pressure formation during the coal mining machine cutting process.

[0037] Based on the model of the formation process of coal and rock crushing, collapse, flow and coupling interface pressure during the cutting of interbedded coal and rock by the coal mining machine system, a cutting state classification and identification model is constructed using the cutting state perception method of coal mining machine in complex interbedded thin coal seams, which is used to monitor the cutting state of thin coal seam coal mining machine.

[0038] During the cutting process of a coal mining machine in a complex thin coal seam with interbedded rock, the characteristics of the coal and rock flow velocity field are obtained by extracting the running speed and displacement parameters of coal and rock particles. The dynamic load distribution of the coal mining machine is simulated by optimizing the running speed and displacement parameters of coal and rock particles.

[0039] Based on the characteristics of the coal and rock flow velocity field and the dynamic load distribution of the coal mining machine, a virtual simulation layer is used to simulate the cutting process of a coal mining machine in a complex thin coal seam with interbedded rock. The cutting state classification and identification model is used to identify the cutting state of the coal mining machine in real time. Based on the identification results and the working state information obtained from the physical perception layer, the cutting parameters and cutting path are dynamically adjusted to optimize the cutting process of the coal mining machine in a complex thin coal seam with interbedded rock.

[0040] Furthermore, the data interaction layer is constructed as follows: for the multiple physical sensors set in the physical perception layer, each physical sensor is treated as a sensor node. A cluster head node is selected, and a clustering strategy is adopted to divide all sensor nodes into different clusters. A network topology adapted to the underground scenario is designed. A routing protocol is formulated based on the network topology and clustering strategy. A multi-objective algorithm is used to optimize the routing protocol. The network topology is optimized by clustering, optimizing cluster heads, and constructing a fitness function model. The optimized network topology is used as the communication structure of the data interaction layer. The routing protocol is loaded by reading the external file of the import model of the real-time information of coal and rock cutting status.

[0041] Furthermore, the method for constructing the numerical model-driven mechanism-based lifetime prediction model is as follows:

[0042] The wear of cutting teeth and helical blades of coal mining machines under different interbedded coal seams was determined by multi-domain collaborative coupling simulation method. The wear distribution characteristics of cutting teeth and helical blades of coal mining machines under different interbedded coal seams were analyzed based on the measured wear.

[0043] Based on the wear distribution characteristics of coal mining machine cutting teeth and spiral blades under different interbedded coal seams, the single-factor method was used to analyze the influence of coal and rock compressive strength, drum structure parameters, and coal mining machine kinematic parameters on spiral drum wear.

[0044] Operational data of complex interbedded thin coal seams were collected as historical big data, including cutting parameters, cutting damage, cutting impact signal frequency, and cutting time. Based on the results of single-factor analysis, a big data model was constructed for cutting parameters, cutting damage, cutting impact signal frequency, and cutting time. The big data model was then used to fit the historical big data to generate the initial life prediction curve of the complex interbedded thin coal seam mining machine.

[0045] Acceleration signals and operating data of coal mining machines in complex interbedded thin coal seams are collected in real time, and the current life prediction curve of coal mining machines in complex interbedded thin coal seams is generated using big data models.

[0046] The method based on numerical modeling is used to mine the life decay characteristics of coal mining machines. Based on the generated initial life prediction curve and the current life prediction curve, the big data model is adjusted using meta-learning theory and the fitted historical life prediction curve is corrected in real time by combining the life decay characteristics.

[0047] Furthermore, the method for constructing the adaptive cutting dynamic reliability digital twin monitoring model of the coal mining machine in complex interbedded thin coal seams is as follows:

[0048] The cutting process of the coal mining machine under different coal seam conditions is simulated by using a rigid-flexible coupled virtual prototype model of the coal mining machine in the virtual simulation layer. The structural characteristics of the coal mining machine in real-time service are obtained based on the simulation results. The structural characteristics are used to identify the areas of interest of key components of the coal mining machine cutting section, and the state functions of the maximum stress and maximum amplitude of the key components are established.

[0049] Based on the state functions of the maximum stress and maximum amplitude of key components, the stress reliability, frequency reliability and amplitude reliability of key components are evaluated by stress-intensity interference theory, resonance failure theory and reliability sensitivity theory respectively. Then, the reliability of the design variables of key components of the coal mining machine cutting section is evaluated by sensitivity analysis and used as the reliability sensitivity index of the coal mining machine adaptive cutting system.

[0050] Based on the Copula function, the failure mode correlation and key component correlation are established by using historical failure data and simulation results of key components of the coal mining machine cutting section. The reliability of key components and the reliability of the coal mining machine adaptive cutting system are obtained, and the reliability of key components and the reliability of the coal mining machine adaptive cutting system are used as the reliability indicators of the coal mining machine adaptive cutting system.

[0051] Based on the simulation results, the equivalent stress index of the key components of the coal mining machine's cutting section and the dynamic characteristic index of the coal mining machine were obtained.

[0052] A comprehensive performance evaluation model for dynamic reliability of a coal mining machine during adaptive cutting of complex interbedded coal and rock was constructed using equivalent stress indices of key components in the cutting section, state functions of maximum stress and maximum amplitude of key components, stress reliability, frequency reliability and amplitude reliability of key components, system reliability sensitivity index, system reliability index, and dynamic characteristic index of the coal mining machine. Based on the theory of genetic algorithm, genetic encoding, fitness analysis and multi-generation evolution were performed on the comprehensive performance evaluation model for dynamic reliability of the coal mining machine during adaptive cutting of complex interbedded coal and rock to obtain the inherent frequencies of key components of the coal mining machine, which were used to optimize key components of the cutting section and the adaptive cutting system of the coal mining machine.

[0053] Historical cutting data, cutting parameters, and reliability evaluation data of the coal mining machine are acquired and fused through multi-domain collaborative control technology. The fused data is then used to train a particle swarm optimization long short-term memory (PSO-LSTM) model to obtain a predictive model for evaluating the reliability of key components and the adaptive cutting system of the coal mining machine.

[0054] A digital twin monitoring model for the dynamic reliability of adaptive cutting of coal mining machines in complex interbedded thin coal seams was built using this prediction model. This model is used to optimize key components and the adaptive cutting system of the coal mining machine based on the evaluation results of the prediction model, so as to realize real-time evaluation and monitoring of the dynamic reliability of key components and the adaptive cutting of the coal mining machine during the adaptive cutting process of coal mining machines in complex interbedded thin coal seams.

[0055] The beneficial effects of adopting the above technical solution are as follows:

[0056] This invention is based on the concept of digital twin technology. It employs a combination of on-site testing and sampling, virtual prototyping technology, multi-domain modeling and collaborative simulation, and experimental analysis to establish a multi-domain deep fusion experimental system for multi-source heterogeneous data during the adaptive control and cutting process of a coal mining machine in complex interbedded thin coal seams. It constructs a coupling theory and method for a mechatronic-hydraulic-control integrated virtual prototype between a discrete element bonded contact model of coal and rock mass and the power transmission system of the coal mining machine. Using bidirectional coupled coal and rock cutting simulation data streams as the main thread, and employing accurate characterization of the coal mining machine's cutting state information and reasonable data processing rules, it combines big data analysis to design a highly accurate coal mining machine cutting state perception model with strong adaptability to thin coal seams. Through coal mining machine cutting state perception and decision-making control methods, it designs an adaptive cutting guidance and control system that enhances the synergy of the data twin of the coal mining machine in complex interbedded thin coal seams by improving the adaptability of the coal mining machine to such conditions.

[0057] The method of this invention establishes an adaptive control system for cutting complex interbedded thin coal seams in coal mining machines, which integrates physical models, virtual models, decision models, and interaction models. This is also an important measure to ensure the optimal comprehensive performance of thin coal seam coal mining machines under complex geological conditions and to protect the safety and reliability of coal mining machines under changes in external loads. Attached Figure Description

[0058] Figure 1 This is a schematic diagram of the intelligent, efficient, and reliable truncation and control digital twin system with structural evolution in this embodiment;

[0059] Figure 2 This is a flowchart illustrating the cutting status perception of a coal mining machine in complex interbedded thin coal seams in this embodiment.

[0060] Figure 3This is a flowchart illustrating the optimization design of the adaptive control strategy for a coal mining machine in a complex thin coal seam with interbedded rock in this embodiment.

[0061] Figure 4 This is a schematic diagram of the digital space for adaptive cutting of a coal mining machine in complex interbedded thin coal seams in this embodiment;

[0062] Figure 5 This is a schematic diagram of the mechanistic lifetime prediction model driven by the numerical model in this embodiment;

[0063] Figure 6 This is a schematic diagram of the adaptive cutting dynamic reliability digital twin monitoring model of the coal mining machine in complex interbedded thin coal seams in this embodiment. Detailed Implementation

[0064] To facilitate understanding of this application, specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and embodiments. The following embodiments are illustrative of the invention but are not intended to limit its scope. Rather, these embodiments are provided to provide a more thorough and complete understanding of the disclosure of this application.

[0065] This implementation method takes an anisotropic, complex, thin coal seam mining machine with interbedded gangue as the research object. It constructs an adaptive cutting guidance and control mechanism and a reliability evaluation digital twin model modeling method to consider the synergistic coupling characteristics of multiple fields including mechanics, electronics, hydraulics, and control. It conducts cross-boundary optimization design of the mining machine's cutting state perception under the influence of multi-source data information, different geometric and kinematic parameters, multi-generational evolutionary structure, health status performance, and reliability data models. Ultimately, it forms a set of intelligent, efficient, and reliable cutting control digital twin systems with structural evolution, such as... Figure 1 As shown.

[0066] This embodiment presents a method for constructing an adaptive cutting and control system for a coal mining machine based on digital twins, such as... Figure 2 As shown, the method includes the following steps:

[0067] Design a cutting state perception method and an adaptive control strategy for a coal mining machine with complex interbedded thin coal seams.

[0068] The specific details of the method for sensing the cutting status of a coal mining machine in complex interbedded thin coal seams are as follows:

[0069] Virtual simulation experiments based on the cutting vibration mechanism of thin coal seam mining machines were conducted to simulate the cutting process of thin coal seam mining machines under different combinations of coal-rock ratios, coal-rock structures, and coal-rock hardness. Based on the simulation results, a nonlinear correlation model between the coal-rock cutting state and the vibration signal of the thin coal seam mining machine was constructed. The nonlinear correlation model between the coal-rock cutting state and the vibration signal of the thin coal seam mining machine was represented by a one-dimensional vibration acceleration curve.

[0070] In this embodiment, such as Figure 2 As shown, a virtual simulation experiment based on the cutting vibration mechanism of a thin coal seam mining machine is used to analyze the influence of different coal-rock ratios, coal-rock structures, and coal-rock hardness on the time-domain information of the vibration signal experienced by the cutting part of the mining machine during the cutting process. Specifically, combinations of different coal-rock ratios, coal-rock structures, and coal-rock hardness are set. A one-dimensional curve of vibration acceleration is obtained by conducting a virtual simulation experiment based on the cutting vibration mechanism of a thin coal seam mining machine, and the time-domain information is extracted. Based on the extracted time-domain information, a nonlinear correlation model reflecting the coal-rock cutting state of the thin coal seam mining machine and its corresponding vibration signal is constructed.

[0071] A customized model for information transformation was used to convert the one-dimensional vibration acceleration curve into a two-dimensional time-spectrum image. Data analysis was then used to extract the cross-term features of the two-dimensional time-spectrum image. The physical meaning of the two-dimensional time-spectrum image was verified by comparing the extracted cross-term features with the physical phenomena in the coal mining machine cutting process.

[0072] In this embodiment, such as Figure 2 As shown, based on the STFT stationary transform theory with time-domain to frequency-domain characteristics, the one-dimensional vibration acceleration curve is transformed into a two-dimensional time-spectrum image through three steps: segmenting the non-stationary vibration signal, transforming multiple stationary signals, and fitting the signal using a window-covering method. Data analysis is used to analyze the cross-term features in the transformed image. By suppressing interference terms in the cross-term features and verifying them using the retained cross-term features, the transformed two-dimensional time-spectrum image accurately reflects the physical process of coal and rock cutting by the thin coal seam mining machine, preventing the loss of identification feature information, verifying the physical meaning represented by the image itself, and realizing the customized model structure design for information conversion. The analysis results show that the STFT transform effectively preserves the time-frequency resolution of the rocker arm shell vibration information, highlights the characteristic information of energy clusters, and clearly distinguishes the distribution patterns of energy characteristics under different working conditions.

[0073] Using the feature information processing module, the two-dimensional time-spectrum image processed by the information transformation customized model is fused with the multi-source data information image that conforms to the structural variation distribution of arbitrarily complex interbedded coal seams, and the generated multi-source data information fused image is used to build a basic data sample library for cutting state recognition.

[0074] In this embodiment, based on the information transformation customization model, a multi-source data information image fusion design method is constructed that conforms to the structural variation distribution of arbitrarily complex interbedded coal seams. The feature information processing module is modeled, and the basic data sample library for cutting state recognition is built to provide accurate original samples for the augmentation of the recognition system data.

[0075] The specific content of using the feature information processing module to fuse the two-dimensional time-spectrum image processed by the information transformation customized model with the acquired multi-source data information image that conforms to the structural variation distribution of arbitrarily complex interbedded coal seams, and using the generated multi-source data information fused image to build a basic data sample library for cutting state recognition, is as follows:

[0076] The feature information processing module includes a feature detail fusion model and a feature detail augmentation model.

[0077] The two-dimensional time-spectrum image processed by the information transformation customized model and the multi-source data information image that conforms to the structural variation distribution of arbitrarily complex interbedded coal seams are input into the feature detail fusion model. The nonlinear and multi-level morphological wavelet fusion algorithm is used to decompose the two-dimensional time-spectrum image and all multi-source data information images respectively and extract image features. Based on the extracted image features, the images are fused to generate a multi-source data information fused image and the coal and rock cutting state of the multi-source data information fused image is marked. The marked multi-source data information fused image is used as the basic data sample.

[0078] The basic data samples are input into the feature detail augmentation model. Based on the generative adversarial network (GAN) model, the model structure parameters of the generator and discriminator are set. The generator generates new data samples based on the basic data samples, and the discriminator distinguishes the new data samples generated by the generator from the basic data samples. A gradient penalty mathematical model is defined to train the generator and discriminator, resulting in a well-trained generator and discriminator.

[0079] The gradient penalty mathematical model is expressed as follows:

[0080]

[0081] Where GP represents the gradient penalty term; γ represents the penalty coefficient; This represents the expectation of the sampling point a' on the distribution P; (a' is a sampling point distributed in the synthetic data sample); This is the gradient representation of the discriminator.

[0082] In this embodiment, such as Figure 2 As shown, appropriate wavelet bases and decomposition levels are selected to fuse the two-dimensional time-spectrum image processed by the customized information transformation model with the multi-source data information image that conforms to the structural variation distribution of arbitrarily complex interbedded coal seams. For the feature detail augmentation model, a gradient penalty mathematical model satisfying the distribution of detail features is defined based on image data robustness theory. The Lipschitz function is used to constrain coherence and verify the imbalance of weight allocation. Wasserstein distance is used as a criterion for evaluating sample type to provide evaluation data for predicting the stability performance of the system.

[0083] The basic data samples are input into the trained generator and discriminator to obtain new basic data samples. The basic data samples and the new basic data samples are used to build a basic data sample library for truncated state recognition.

[0084] A cut-off state classification and recognition model was constructed using a basic data sample library for cut-off state recognition.

[0085] The method for constructing the truncated state classification and recognition model is as follows: A convolutional neural network that meets the requirements of the truncated state classification task is selected, and a classification layer is added to the end of the convolutional neural network to construct the truncated state classification and recognition model. This model is used to identify the truncated state category of the input sample data by extracting features from the input sample data. Initial convolutional and pooling parameters are set in the truncated state classification and recognition model. Training sets with different amounts of data are selected from the truncated state recognition basic data sample library, and the same truncated state classification and recognition model is trained using different datasets. The particle swarm optimization algorithm is used to optimize the truncated state classification and recognition model, resulting in an optimized truncated state classification and recognition model. Simultaneously, the classification accuracy and mean root mean square error of the truncated state classification and recognition model trained on different datasets are recorded. By analyzing the linear relationship between the amount of data and the classification accuracy and mean root mean square error of the truncated state classification and recognition model, the performance of the truncated state classification and recognition model under different data conditions is evaluated, thereby selecting the optimal truncated state classification and recognition model.

[0086] In this embodiment, such as Figure 2 As shown, multi-source data information fusion images are used as a database replacement model for multi-target recognition algorithms. Further, variables for convolution parameters and pooling parameters of feature extraction and recognition classification structures are established to achieve two objectives in perception performance: accuracy and average root mean square error of recognition classification. The correlation between recognition response and model structure and database size is verified by analyzing the linear relationship between the input of augmented model data and the output of recognition model results.

[0087] The specific content of the adaptive control strategy for the coal mining machine in complex interbedded thin coal seams is as follows:

[0088] Define the performance indicators of coal mining machines for complex interbedded thin coal seams, including: cutting area, productivity, cutting energy consumption, cutting resistance, load fluctuation coefficient, and coal loading rate.

[0089] Define the motion parameters of a coal mining machine for complex interbedded thin coal seams, including: traction speed v. q And rotational speed n.

[0090] A multi-task Gaussian algorithm is used to fit the fitting surfaces and fitting errors of each performance index with the motion parameters of the coal mining machine. This yields the functional equations of each performance index with respect to the motion parameters of the coal mining machine, and a comprehensive performance index evaluation model is constructed, expressed as:

[0091] S(v q ,n)=av q 2 -bv q n+cn 2 +dv q -en+f (2)

[0092] Where S represents the comprehensive performance index evaluation function; a, b, c, d, e, f represent the weights of each performance index.

[0093] Based on the actual working conditions, the constraints of the coal mining machine's motion parameters are defined. Based on the auxiliary design and load calculation software of the thin coal seam coal mining machine drum, the comprehensive performance index evaluation model is divided into several regions along the cutting direction. In each region, the piecewise fitting method is used to approximate the constraints of the coal mining machine's motion parameters.

[0094] The constraint condition is expressed as follows:

[0095] X = (x1, x2) T =(v q ,n) T (3)

[0096] Where X represents the constraint condition; x1 and x2 are both constraint terms.

[0097] The optimization objective is defined as maximizing cutting area, productivity, and coal loading rate, while minimizing cutting energy consumption, cutting resistance, and load fluctuation coefficient. Based on the approximation ideal solution ranking theory, the combination of motion parameters of the coal mining machine under different working conditions is used as the evaluation object, and performance indicators are used as evaluation indicators. An n*m decision matrix between the evaluation object and the evaluation indicators is established and then normalized and standardized to generate the optimal combination of motion parameters of the coal mining machine under different working conditions; where n represents the number of evaluation objects and m represents the number of evaluation indicators.

[0098] Based on the optimal combination of motion parameters of the coal mining machine under different working conditions, a mechanical-electrical-hydraulic-control coupling model is constructed, and the control sequence is designed according to the constructed mechanical-electrical-hydraulic-control coupling model; wherein the control sequence is: the control strategy of the coal mining machine with priority to drum speed control model, the control strategy of control model with priority to traction speed control model, and the control strategy of control model with coordinated control model.

[0099] A model is established based on the designed control sequence to correlate the control strategy with the coal seam structure and motion parameters, which is used to automatically select the optimal control strategy based on the coal seam structure and motion parameters.

[0100] A deep reinforcement learning method is used to optimize the model that correlates the control strategy with the coal seam structure and motion parameters, resulting in an adaptive control model. This model is used to generate an adaptive cutting optimal control strategy based on changes in actual working conditions.

[0101] In this embodiment, such as Figure 3 As shown, a comprehensive performance evaluation model for thin coal seam mining machines is constructed using multi-objective parameter optimization design theory. A multi-task Gaussian algorithm is used to fit the cutting area, productivity, cutting ratio energy consumption, cutting resistance, load fluctuation coefficient, coal loading rate, and mining machine motion parameters (v). q By fitting the surface and fitting error between -n), the functional equations between various performance parameters and the motion parameters of the coal mining machine are obtained. Based on the drum-assisted design and load calculation software of the thin coal seam coal mining machine, the comprehensive performance index evaluation model is divided into several regions along the cutting direction. The thickness, location, and hardness of the intercalated rock layer are different in different regions. The constraints of the motion parameters of the coal mining machine are approximated by a piecewise fitting method. Taking the maximum cutting area, productivity, and coal loading rate, and the minimum cutting specific energy consumption, cutting resistance, and load fluctuation coefficient as the optimization objectives, based on the approximation ideal solution ranking theory, an n*m decision matrix between the above evaluation objects and evaluation indicators is established, and positive and standardization processing is performed on it to clarify the optimal and worst schemes corresponding to each optimization objective, thereby obtaining the optimal combination of motion parameters of the coal mining machine under different working conditions, that is, directly taking the optimal scheme corresponding to each optimization objective. The mechanical-electrical-hydraulic-control coupled model is constructed to design the control sequence. Based on the changes in the performance indicators of the coal mining machine during the adjustment process, the control strategies of the coal mining machine using the drum speed priority control model, the traction speed priority control model, and the collaborative control model are clarified. A model is formed that correlates the control strategy with the coal seam structure and performance parameters. The applicability and rationality of the motion parameter control strategy with the optimal comprehensive performance of coal mining machines in various complex interbedded thin coal seam structures are obtained. Based on deep learning theory, a deep reinforcement learning adaptive control model is built to obtain the optimal adaptive truncation control strategy.

[0102] Based on the cutting state perception method and adaptive control strategy of the coal mining machine for complex interbedded thin coal seams, an adaptive cutting digital twin model of the coal mining machine for complex interbedded thin coal seams is established to simulate the adaptive cutting process of the coal mining machine for complex interbedded thin coal seams.

[0103] The adaptive cutting digital twin model of the complex interbedded thin coal seam mining machine includes: a physical perception layer, a virtual simulation layer, a twin decision layer, and a data interaction layer.

[0104] In this embodiment, such as Figure 4 As shown, taking the MG2×55 / 250-BWD complex interbedded thin coal seam mining machine as the research object, an adaptive cutting digital twin model of the mining machine is established. The model structure includes four layers: physical perception layer, virtual simulation layer, twin decision layer, and data interaction layer. The digital space for adaptive cutting of complex interbedded thin coal seam mining machine is built from three-dimensional model, virtual scene and behavioral logic criteria.

[0105] The physical sensing layer is used to construct the physical structure system of the coal mining machine, and to intelligently monitor the physical structure system of the coal mining machine by designing multiple physical sensors, so as to obtain the working status information of the coal mining machine in complex interbedded thin coal seams in real time.

[0106] The working status information includes: attitude parameter information of the coal mining machine in complex interbedded thin coal seams and multi-source data information collected by multiple physical sensors; wherein the attitude parameter information includes: traction speed, rotation speed and rocker arm adjustment height.

[0107] In this embodiment, by designing the physical structure system of the coal mining machine, the relationship and information fusion of multiple physical sensors, and the data acquisition system, the current mode of relying mainly on manual observation in the mining of complex interbedded coal and rock in thin coal seams is transformed into intelligent information acquisition, while strengthening the deep integration of information.

[0108] The physical sensing layer is constructed as follows: the physical sensing layer consists of three parts: a coal mining machine physical structure system, a multi-physical sensor design, and a data acquisition system; wherein the coal mining machine physical structure system is a coal mining machine adaptive cutting control integrated test bench designed based on similarity theory and cutting experiments.

[0109] In this embodiment, based on the publicly disclosed design method of the adaptive cutting control integrated test bench for coal mining machines, the adaptive cutting control integrated test bench for coal mining machines is designed and developed independently according to similarity theory and cutting experiments. This integrated test bench consists of a mechanical system, a hydraulic system, an electrical system, a data acquisition and transmission system, a test bench control and monitoring system, and a host computer. The mechanical system comprises a coal face clamping device, a cutting device, and a cable conveying device. The coal face clamping device includes an experimental gantry and a base. The cutting device includes a spiral drum, a rocker arm, a sliding platform, and guide rails. The spiral drum is modeled after the spiral drum of the MG2×55 / 250-BWD type thin coal seam coal mining machine, establishing a similar drum model. The hydraulic system consists of a top clamping cylinder acting on the coal face, left and right side clamping cylinders and a pushing cylinder, a sliding platform lateral propulsion cylinder, a sliding platform traction cylinder, a rocker arm height adjustment cylinder, a drum drive motor, and auxiliary components such as a pump station. The electrical system includes a motor control pump motor, a main motor, a filter motor, and a cooling fan.

[0110] The design method of the multi-physical sensor is as follows: by analyzing the load law of key components of the coal mining machine cutting section through dynamic analysis, the sensing force points are determined, and the layout of the multi-physical sensor is designed according to the sensing force points; the sensors are arranged according to the layout of the multi-physical sensor, and based on the method of sensing the cutting state of the coal mining machine in complex interbedded thin coal seams, a spiral drum cutting vibration sensing system, a square head vibration sensing system, and a rocker arm vibration sensing system are constructed to obtain the attitude parameter information of the coal mining machine in complex interbedded thin coal seams in real time, and the cutting state of the thin coal seam coal mining machine is identified by using the multi-source data information collected in real time by each physical sensor during the coal mining machine cutting process, so as to realize the dynamic monitoring of the coal mining machine cutting complex thin coal seams.

[0111] The layout of the multi-physical sensors is as follows: For the adaptive cutting control integrated test bench of the coal mining machine, a torque sensor and a vibration acceleration transmitter are installed between the drum and the hydraulic motor to monitor the motor speed, the torque it receives, and the vibration characteristics during the drum cutting of coal and rock; a magnetostrictive displacement sensor is installed at the tail end of the traction cylinder to measure the traction speed; and pressure transmitters are installed in the coal wall clamping cylinder, traction cylinder, height adjustment cylinder, accumulator, pump station, motor circuit, and lateral propulsion locking cylinder to measure the pressure of the coal wall clamping cylinder, traction cylinder, height adjustment cylinder, accumulator, pump station, motor circuit, and lateral propulsion locking cylinder.

[0112] In this embodiment, the spiral drum, square head, and rocker arm are all key components of the coal mining machine's cutting section. Based on the cutting status sensing method for coal mining machines in complex interbedded thin coal seams described earlier, a spiral drum cutting vibration sensing system, a square head vibration sensing system, and a rocker arm vibration sensing system are designed. A torque sensor and a vibration acceleration transmitter are installed between the drum and the hydraulic motor to comprehensively test the motor speed, torque, and vibration characteristics during the drum's cutting of coal and rock. The cutting status of the thin coal seam coal mining machine is identified using multi-source sensor data as characteristic parameters. A magnetostrictive displacement sensor installed at the tail end of the traction cylinder, combined with the time taken for the traction process, is used to obtain the traction speed of the test bench coal mining machine. Pressure transmitters are used to measure the pressure of the coal wall clamping cylinder, traction cylinder, height adjustment cylinder, accumulator, pump station, motor circuit, and lateral propulsion locking cylinder. The combined measurement results enable dynamic monitoring of the coal mining machine's cutting process in complex thin coal seams.

[0113] The data acquisition system is used to monitor and record the operating parameters of the coal mining machine adaptive cutting control integrated test bench and the working status data collected in real time by multiple physical sensors.

[0114] In this embodiment, such as Figure 4As shown, the existing data acquisition system is used to monitor and record the operating parameters of the coal mining machine adaptive cutting control integrated test bench and the working status data collected in real time by multiple physical sensors.

[0115] The virtual simulation layer is used to construct a high-precision three-dimensional twin model of a dynamic, complex, thin coal seam with interbedded gangue. By combining the real-time acquired working status information of the coal mining machine in the complex, thin coal seam with interbedded gangue, an adaptive control model is used to generate an adaptive cutting optimal control strategy, which is used to adaptively adjust and control the actions of the physical structure system of the coal mining machine in the physical perception layer.

[0116] The method for constructing the virtual simulation layer is as follows: coal and rock samples are collected from typical working faces and their physical and mechanical properties are tested. The physical parameters of the coal and rock mass based on discrete element method are determined and a discrete element model of the coal and rock mass is constructed. Based on application programming interface (API) technology, the discrete element model of the coal and rock mass is corrected by compiling a replacement API for multiple particle type clusters. A high-precision three-dimensional twin model of dynamic complex interbedded thin coal seam with real-time correction and replacement of particle sets is constructed.

[0117] In this embodiment, such as Figure 4 As shown, for Yangcun Mine of Yanzhou Coal Industry and Tashan Mine of Datong Coal Industry, coal and rock samples from typical working faces containing complex interbedded rock, hard nodules, and faults with different hardness and number of layers were selected according to sampling standards for physical and mechanical property testing to determine the physical parameters of coal and rock mass based on discrete element method (DEM), including: intrinsic parameters of the material, basic contact parameters, and contact model parameters. Application Program Interface (API) technology was integrated to compile a replacement API for multiple particle type clusters. Real-time model correction was achieved through the replacement API, dynamically adjusting the particle set in the model according to the working state of the coal mining machine and the actual response of the coal and rock mass. Through behavioral logic programming, interactive simulation between the coal mining machine and the coal and rock model was realized, simulating the cutting process of the coal mining machine under different coal seam conditions. A high-precision three-dimensional twin model of complex interbedded thin coal seams with real-time correction and replacement of particle sets was constructed, achieving high-fidelity reconstruction of complex interbedded thin coal seams.

[0118] Based on the attitude parameter information of the coal mining machine in the complex interbedded thin coal seam acquired in real time, the cutting action of the coal mining machine is simulated and a high-precision three-dimensional twin model of the dynamic complex interbedded thin coal seam with real-time correction and replacement of particle sets is generated. Based on the adaptive control strategy of the coal mining machine in the complex interbedded thin coal seam, an adaptive control model is constructed using multi-source data information collected from multiple physical sensors. The adaptive control model is used to generate the optimal adaptive cutting control strategy of the coal mining machine. The adjustment action of the coal mining machine is virtually executed according to the generated optimal adaptive cutting control strategy, and the expected working state information is generated.

[0119] The virtually executed adjustment actions are fed back to the physical perception layer to guide the physical structure system of the coal mining machine to perform actions, obtain the working status information after the action, evaluate the effectiveness of the adaptive cutting optimal control strategy by comparing the expected working status information and the working status information after the action, adjust the adaptive cutting optimal control strategy according to the evaluation results, and then virtually execute the adjustment action of the coal mining machine again to form a closed-loop control process.

[0120] In this embodiment, such as Figure 4 As shown, by acquiring the attitude parameter information of the coal mining machine, the corresponding action simulation is completed in the virtual simulation layer, including speed adjustment and height adjustment. Then, the adaptive control model is used to virtually execute the adjustment action to guide the adjustment of the attitude and position of the coal mining machine in the physical sensing layer. That is, the traction speed and rotation speed of the coal mining machine and the height of the rocker arm are adjusted in real time according to the structural distribution of the coal seam, so as to realize closed-loop adaptive adjustment control.

[0121] A rigid-flexible coupled virtual prototype model of a coal mining machine was established to cut coal and rock. The hydraulic system and electrical control system of the coal mining machine were integrated into the virtual simulation layer. Interface technology was used to realize the interconnection between the models, determine the input and output of each model, and perform joint simulation to simulate the cutting process of a coal mining machine in a complex thin coal seam with interbedded gangue.

[0122] In this embodiment, such as Figure 4 As shown, a rigid-flexible coupled virtual prototype model of a coal mining machine is established to cut coal and rock, along with a hydraulic system and an electrical-control system. Interconnection and interoperability between the various system models are achieved based on interface technology, the inputs and outputs of each system model are determined, and joint simulation is performed.

[0123] The twin decision layer is used to optimize the cutting process of the coal mining machine in complex interbedded thin coal seams by using the cutting state classification and recognition model to identify the cutting state of the coal mining machine in real time based on the working state information obtained by the physical perception layer and the dynamic complex interbedded thin coal seam high-precision three-dimensional twin model constructed by the virtual simulation layer.

[0124] The method for constructing the twin decision layer is as follows: using the discrete element model of coal and rock mass constructed in the virtual simulation layer as the basis, a test model for coal mining machine cutting coal seams is constructed by changing the position, thickness, and rock properties of the interbedded rock layer in the discrete element model of coal and rock mass in the virtual simulation layer. The test model is then used to analyze the stress law, vibration characteristics, and cutting tooth state of the spiral drum cutting coal and rock in the coal mining machine. A process model for coal and rock crushing, collapse, flow, and the formation of interface pressure at the coupling surface during the coal mining machine system cutting interbedded coal and rock is also constructed. This model is used to analyze the coal and rock crushing, collapse, flow, and interface pressure formation during the coal mining machine cutting process.

[0125] Based on the model of the formation process of coal and rock crushing, collapse, flow and coupling interface pressure during the cutting of interbedded coal and rock by the coal mining machine system, a cutting state classification and identification model is constructed using the cutting state perception method of coal mining machine in complex interbedded thin coal seams, which is used to monitor the cutting state of thin coal seam coal mining machine.

[0126] In this embodiment, such as Figure 4 As shown, a discrete element model with varying rock interlayer hardness is used as the basis. A test model for coal seam cutting by a coal mining machine is constructed by changing the location, thickness, and rock properties of the interlayer. This test model is then used to analyze the stress patterns, vibration characteristics, and cutting tooth state of the spiral drum during coal and rock cutting. A model is constructed to model the formation process of coal and rock fracturing, collapse, flow, and interface pressure at the coupling surface during the coal mining machine system's cutting of interlayered coal and rock. A multi-source data information fusion network model for the perception, identification, and monitoring of the cutting state of a thin coal seam coal mining machine is also constructed. This model includes three modules: data generation, accuracy assessment, and online monitoring, used to achieve multi-mode identification and monitoring of the cutting state of a thin coal seam coal mining machine under different data sources.

[0127] During the cutting process of a coal mining machine in a complex thin coal seam with interbedded rock, the characteristics of the coal and rock flow velocity field are obtained by extracting the running speed and displacement parameters of coal and rock particles. The dynamic load distribution of the coal mining machine is simulated by optimizing the running speed and displacement parameters of coal and rock particles.

[0128] Based on the characteristics of the coal and rock flow velocity field and the dynamic load distribution of the coal mining machine, a virtual simulation layer is used to simulate the cutting process of a coal mining machine in a complex thin coal seam with interbedded rock. The cutting state classification and identification model is used to identify the cutting state of the coal mining machine in real time. Based on the identification results and the working state information obtained from the physical perception layer, the cutting parameters and cutting path are dynamically adjusted to optimize the cutting process of the coal mining machine in a complex thin coal seam with interbedded rock.

[0129] In this embodiment, such as Figure 4 As shown, the velocity and displacement parameters of coal and rock particles are extracted to obtain the characteristics of the coal and rock flow velocity field, including velocity magnitude, direction, and variation trend. An optimization program for the velocity and displacement parameters of coal and rock particles is developed to obtain the dynamic load distribution of the coal mining machine. Simulation provides a data stream, namely cutting parameters and cutting paths, to the digital twin model, thereby improving the accuracy of the twin model and enabling real-time monitoring of the cutting status of the coal mining machine during the cutting process. Optimizing motion parameters makes the cutting process more efficient.

[0130] The data interaction layer is used to transmit and interact data between the physical perception layer, the virtual simulation layer, and the twin decision layer, so as to realize the dynamic adjustment of adaptive cutting of the coal mining machine in complex interbedded thin coal seams.

[0131] The data interaction layer is constructed as follows: For the multiple physical sensors set in the physical perception layer, each physical sensor is treated as a sensor node. A cluster head node is selected, and a clustering strategy is used to divide all sensor nodes into different clusters. A network topology adapted to the underground scenario is designed. A routing protocol is formulated based on the network topology and clustering strategy. A multi-objective algorithm is used to optimize the routing protocol. The network topology is optimized by clustering, optimizing cluster heads, and constructing a fitness function model. The optimized network topology is used as the communication structure of the data interaction layer. The routing protocol is loaded by reading the external file of the import model of the real-time information of coal and rock cutting status.

[0132] In this embodiment, such as Figure 4 As shown, the data interaction layer model, centered on a large number of sensor nodes and integrated with a host computer and switches, is divided into a transmission data model and an interaction data model. This model effectively connects the underground physical sensing layer, the twin decision-making layer of the central control system, and the above-ground virtual simulation layer. Aiming to improve the timeliness of the information transmission network system and reduce system energy consumption, this study analyzes the characteristics of wireless sensor networks, such as their large scale, self-organization, and high fault tolerance. Considering constraints such as the underground working environment of coal mines and the distribution patterns of sensor nodes, a network topology suitable for underground scenarios, i.e., the communication structure of the data interaction layer, is designed. A multi-objective algorithm is used to improve the routing protocol. Through clustering, cluster head optimization, and the construction of a fitness function model, it ensures that nodes can quickly find the optimal transmission path during information transmission. An import model for real-time information on coal and rock cutting status is established. By reading external files from the import model, the routing protocol is loaded, establishing a high-efficiency, low-loss data interaction layer model.

[0133] Based on the adaptive cutting digital twin model of the coal mining machine in complex thin coal seams with interbedded gangue, a digital twin model driven mechanism characteristic life prediction model and a dynamic reliability digital twin monitoring model of the adaptive cutting of the coal mining machine in complex thin coal seams with interbedded gangue are constructed to evaluate the reliability of the adaptive cutting process of the coal mining machine in complex thin coal seams with interbedded gangue in real time.

[0134] The method for constructing the numerical model-driven mechanism-based lifetime prediction model is as follows:

[0135] The wear of cutting teeth and helical blades of coal mining machines under different interbedded coal seams was determined by multi-domain collaborative coupling simulation method. The wear distribution characteristics of cutting teeth and helical blades of coal mining machines under different interbedded coal seams were analyzed based on the measured wear.

[0136] Based on the wear distribution characteristics of coal mining machine cutting teeth and spiral blades under different interbedded coal seams, the single-factor method was used to analyze the influence of coal and rock compressive strength, drum structural parameters, and coal mining machine kinematic parameters on spiral drum wear.

[0137] Operational data of complex interbedded thin coal seams mining machines under different interbedded coal seams were collected as historical big data, including cutting parameters, cutting damage, cutting impact signal frequency, and cutting time. Based on the results of single-factor analysis, a big data model was constructed for cutting parameters, cutting damage, cutting impact signal frequency, and cutting time. The big data model was then used to fit the historical big data to generate the initial life prediction curve of the complex interbedded thin coal seam mining machine.

[0138] Acceleration signals and operating data of coal mining machines in complex interbedded thin coal seams are collected in real time, and the current life prediction curve of the coal mining machine in complex interbedded thin coal seams is generated using a big data model.

[0139] The method based on numerical modeling is used to mine the life decay characteristics of coal mining machines. Based on the generated initial life prediction curve and the current life prediction curve, the big data model is adjusted using meta-learning theory and the fitted historical life prediction curve is corrected in real time by combining the life decay characteristics.

[0140] In this embodiment, such as Figure 5 As shown, the wear distribution characteristics of cutting teeth and spiral blades of coal mining machines under different interbedded coal seams are determined using existing multi-domain collaborative coupling simulation methods. Based on the single-factor method, the influence of coal and rock compressive strength, drum structural parameters, and coal mining machine kinematic parameters on spiral drum wear is analyzed. A big data model is established for relevant cutting parameters, cutting damage, cutting impact signal frequency, and cutting time. Based on this model, the initial life prediction curve of the thin coal seam coal mining machine is derived by fitting curves from historical big data. Using real-time acceleration signals and combining the obtained mechanistic model, meta-learning theory is used in conjunction with the big data model to predict and correct the historical life prediction curve in real time. Based on the numerical model-driven mining of coal mining machine life decay characteristics, a numerical model-driven mechanistic characteristic life prediction model is established.

[0141] The method for constructing the adaptive cutting dynamic reliability digital twin monitoring model of the coal mining machine in complex interbedded thin coal seams is as follows:

[0142] The cutting process of the coal mining machine under different coal seam conditions is simulated by using a rigid-flexible coupled virtual prototype model of the coal mining machine in the virtual simulation layer. Based on the simulation results, the structural characteristics of the coal mining machine in real-time service are obtained, and the regions of interest of key components of the cutting part of the coal mining machine are identified by using the structural characteristics. The state functions of the maximum stress and maximum amplitude of the key components are established.

[0143] Based on the state functions of the maximum stress and maximum amplitude of key components, the stress reliability, frequency reliability and amplitude reliability of key components are evaluated by stress-intensity interference theory, resonance failure theory and reliability sensitivity theory, respectively. Then, the reliability of the design variables of key components of the coal mining machine cutting section is evaluated by sensitivity analysis and used as the reliability sensitivity index of the adaptive cutting system of the coal mining machine.

[0144] Based on the Copula function, failure mode correlation and key component correlation are established using historical failure data and simulation results of key components of the coal mining machine cutting section. The reliability of key components and the reliability of the coal mining machine adaptive cutting system are obtained, and the reliability of key components and the reliability of the coal mining machine adaptive cutting system are used as the reliability indicators of the coal mining machine adaptive cutting system.

[0145] Based on the simulation results, the equivalent stress index of the key components of the coal mining machine's cutting section and the dynamic characteristic index of the coal mining machine were obtained.

[0146] A comprehensive dynamic reliability performance evaluation model for the adaptive cutting of complex interbedded coal and rock was constructed using equivalent stress indices of key components of the coal mining machine's cutting section, state functions of maximum stress and maximum amplitude of key components, stress reliability, frequency reliability, and amplitude reliability of key components, system reliability sensitivity index, system reliability index, and dynamic characteristic index of the coal mining machine. Based on genetic algorithm theory, genetic encoding, fitness analysis, and multi-generation evolution were performed on the comprehensive dynamic reliability performance evaluation model for the adaptive cutting of complex interbedded coal and rock to obtain the inherent frequencies of key components of the coal mining machine, which were then used to optimize the key components of the coal mining machine's cutting section and the adaptive cutting system of the coal mining machine.

[0147] In this embodiment, such as Figure 6 As shown, based on the structural characteristics of the coal mining machine during real-time service, the areas of interest for key components of the cutting section are determined. The relationship between the areas of interest and the maximum amplitude of these key components is fitted using MATLAB, establishing state functions for the maximum stress and maximum amplitude of the key components to simulate the stress and vibration conditions of these components during service. Combining stress-strength interference theory, resonance failure theory, and reliability sensitivity theory, the stress reliability, frequency reliability, and amplitude reliability of the key components are analyzed, along with the reliability sensitivity of the design variables. Based on the Copula function, failure mode correlation and key component correlation are established to obtain the reliability of the key components and the reliability of the adaptive cutting system of the coal mining machine. By forming a comprehensive dynamic reliability performance evaluation model for the adaptive cutting of complex interbedded coal and rock by the coal mining machine, including equivalent stress indices of key components, state functions for the maximum stress and maximum amplitude of key components, stress reliability, frequency reliability, and amplitude reliability of key components, system reliability sensitivity indices, system reliability indices, and dynamic characteristic indices of the coal mining machine, a comprehensive dynamic reliability performance evaluation model is developed. Genetic encoding, fitness analysis, and multi-generational evolution are then implemented to complete the evolutionary design of the key components and the system.

[0148] Historical cutting data, cutting parameters, and reliability evaluation data of the coal mining machine are acquired and fused using multi-domain collaborative control technology. The fused data is then used to train a particle swarm optimization long short-term memory (PSO-LSTM) model to obtain a predictive model for evaluating the reliability of key components and the adaptive cutting system of the coal mining machine.

[0149] A digital twin monitoring model for the dynamic reliability of adaptive cutting of coal mining machines in complex interbedded thin coal seams was built using this prediction model. This model is used to optimize key components and the adaptive cutting system of the coal mining machine based on the evaluation results of the prediction model, so as to realize real-time evaluation and monitoring of the dynamic reliability of key components and the adaptive cutting of the coal mining machine during the adaptive cutting process of coal mining machines in complex interbedded thin coal seams.

[0150] In this embodiment, such as Figure 6 As shown, a digital twin monitoring model for the dynamic reliability of adaptive cutting of a coal mining machine in a complex thin coal seam with interbedded gangue was constructed. Through multi-domain collaborative control technology, virtual simulation data and historical cutting data were integrated. Combining the coal mining machine cutting parameters and dynamic reliability evaluation data, the Particle Swarm Optimization Long Short-Term Memory (PSO-LSTM) neural network algorithm was used as an alternative model to the multi-objective genetic algorithm. The cutting parameters and reliability evaluation data were written into the PSO-LSTM neural network algorithm. The reliability of key components and systems was evaluated through algorithm simulation results, and the dynamic reliability of key components and systems during the adaptive cutting process was evaluated and monitored in real time.

[0151] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope defined by the claims of the present invention.

Claims

1. A method for constructing an adaptive cutting and control system for a coal mining machine based on digital twins, characterized in that, This method includes the following steps: Design a cutting state perception method and an adaptive control strategy for a coal mining machine with complex interbedded thin coal seams. The specific content of the adaptive control strategy for the coal mining machine in complex interbedded thin coal seams is as follows: Define the performance indicators of coal mining machines for complex interbedded thin coal seams, including: cutting area, productivity, cutting ratio energy consumption, cutting resistance, load fluctuation coefficient, and coal loading rate; Define the motion parameters of a coal mining machine for complex interbedded thin coal seams, including: traction speed. and rotational speed ; A multi-task Gaussian algorithm was used to fit the fitting surfaces and fitting errors of each performance index and the motion parameters of the coal mining machine, so as to obtain the functional equations of each performance index and the motion parameters of the coal mining machine and construct a comprehensive performance index evaluation model. Based on the actual working conditions, the constraints of the motion parameters of the coal mining machine are defined. Based on the auxiliary design and load calculation software of the drum of the thin coal seam coal mining machine, the comprehensive performance index evaluation model is divided into several regions along the cutting direction. In each region, the piecewise fitting method is used to approximate the constraints of the motion parameters of the coal mining machine. The optimization objective is defined as maximizing cutting area, productivity, and coal loading rate, while minimizing cutting specific energy consumption, cutting resistance, and load fluctuation coefficient. Based on the approximation ideal solution ranking theory, the combination of motion parameters of the coal mining machine under different working conditions is used as the evaluation object, and performance indicators are used as evaluation indicators. An n*m decision matrix between the evaluation object and the evaluation indicators is established and then normalized and standardized to generate the optimal combination of motion parameters of the coal mining machine under different working conditions; where n represents the number of evaluation objects and m represents the number of evaluation indicators. A mechanical-electrical-hydraulic-control coupling model is built based on the optimal combination of motion parameters of the coal mining machine under different working conditions. The control sequence is designed based on the built mechanical-electrical-hydraulic-control coupling model. The control sequence is as follows: the control strategy of the coal mining machine is: the drum speed control model takes priority over the traction speed control model, the traction speed control model takes priority over the drum speed control model, and the coordinated control model. A model is established to correlate the control strategy with the coal seam structure and motion parameters based on the designed control sequence, which is used to automatically select the optimal control strategy based on the coal seam structure and motion parameters. A deep reinforcement learning method is used to optimize the model that correlates the control strategy with the coal seam structure and motion parameters, resulting in an adaptive control model, which is used to generate an adaptive cutting optimal control strategy based on changes in actual working conditions. Based on the cutting state perception method and adaptive control strategy of the coal mining machine for complex interbedded thin coal seams, an adaptive cutting digital twin model of the coal mining machine for complex interbedded thin coal seams is established to simulate the adaptive cutting process of the coal mining machine for complex interbedded thin coal seams. Based on the adaptive cutting digital twin model of the coal mining machine with complex interbedded thin coal seams, a digital twin model driven mechanism characteristic life prediction model and a dynamic reliability digital twin monitoring model of the adaptive cutting of the coal mining machine with complex interbedded thin coal seams are constructed to evaluate the reliability of the adaptive cutting process of the coal mining machine with complex interbedded thin coal seams in real time. The method for constructing the numerical model-driven mechanism-based lifetime prediction model is as follows: The wear of cutting teeth and helical blades of coal mining machines under different interbedded coal seams was determined by multi-domain collaborative coupling simulation method. The wear distribution characteristics of cutting teeth and helical blades of coal mining machines under different interbedded coal seams were analyzed based on the measured wear. Based on the wear distribution characteristics of coal mining machine cutting teeth and spiral blades under different interbedded coal seams, the single-factor method was used to analyze the influence of coal and rock compressive strength, drum structure parameters, and coal mining machine kinematic parameters on spiral drum wear. Operational data of complex interbedded thin coal seams were collected as historical big data, including cutting parameters, cutting damage, cutting impact signal frequency, and cutting time. Based on the results of single-factor analysis, a big data model was constructed for cutting parameters, cutting damage, cutting impact signal frequency, and cutting time. The big data model was then used to fit the historical big data to generate the initial life prediction curve of the complex interbedded thin coal seam mining machine. Acceleration signals and operating data of coal mining machines in complex interbedded thin coal seams are collected in real time, and the current life prediction curve of coal mining machines in complex interbedded thin coal seams is generated using big data models. The method based on numerical modeling is used to mine the life decay characteristics of coal mining machines. Based on the generated initial life prediction curve and the current life prediction curve, the big data model is adjusted using meta-learning theory and the fitted historical life prediction curve is corrected in real time by combining the life decay characteristics.

2. The method for constructing an adaptive cutting and control system for a coal mining machine based on digital twins according to claim 1, characterized in that, The specific details of the method for sensing the cutting status of a coal mining machine in complex interbedded thin coal seams are as follows: Virtual simulation experiments based on the cutting vibration mechanism of thin coal seam mining machines were conducted to simulate the cutting process of thin coal seam mining machines under different combinations of coal-rock ratios, coal-rock structures, and coal-rock hardness. Based on the simulation results, a nonlinear correlation model between the coal-rock cutting state and the vibration signal of the thin coal seam mining machine was constructed. The nonlinear correlation model between the coal-rock cutting state and the vibration signal of the thin coal seam mining machine was represented by a one-dimensional vibration acceleration curve. The one-dimensional vibration acceleration curve was converted into a two-dimensional time-spectrum image using an information transformation customized model. The cross-term features of the two-dimensional time-spectrum image were extracted using data analysis. The physical meaning of the two-dimensional time-spectrum image was verified by comparing the extracted cross-term features with the physical phenomena in the coal mining machine cutting process. Using the feature information processing module, the two-dimensional time-spectrum image processed by the information transformation customized model is fused with the multi-source data information image that conforms to the structural change distribution of arbitrarily complex interbedded coal seams, and the generated multi-source data information fused image is used to build a basic data sample library for cutting state recognition. A cut-off state classification and recognition model was constructed using a basic data sample library for cut-off state recognition.

3. The method for constructing an adaptive cutting and control system for a coal mining machine based on digital twins according to claim 2, characterized in that, The adaptive cutting digital twin model of the complex interbedded thin coal seam mining machine includes: a physical perception layer, a virtual simulation layer, a twin decision layer, and a data interaction layer; The physical sensing layer is used to construct the physical structure system of the coal mining machine, and to intelligently monitor the physical structure system of the coal mining machine by designing multiple physical sensors, so as to obtain the working status information of the coal mining machine in complex interbedded thin coal seams in real time. The virtual simulation layer is used to construct a high-precision three-dimensional twin model of a dynamic, complex, thin coal seam with interbedded gangue. By combining the real-time working status information of the coal mining machine in the complex, thin coal seam with interbedded gangue, the adaptive control model is used to generate an adaptive cutting optimal control strategy, which is used to adaptively adjust and control the action of the physical structure system of the coal mining machine in the physical perception layer. The twin decision layer is used to optimize the cutting process of the coal mining machine in complex interbedded thin coal seams by using the cutting state classification and recognition model to identify the cutting state of the coal mining machine in real time based on the working state information obtained by the physical perception layer and the dynamic complex interbedded thin coal seam high-precision three-dimensional twin model constructed by the virtual simulation layer. The data interaction layer is used to transmit and interact data between the physical perception layer, the virtual simulation layer, and the twin decision layer, so as to realize the dynamic adjustment of adaptive cutting of the coal mining machine in complex interbedded thin coal seams.

4. The method for constructing an adaptive cutting and control system for a coal mining machine based on digital twins according to claim 3, characterized in that, The physical sensing layer is constructed as follows: the physical sensing layer consists of three parts, namely: the coal mining machine physical structure system, the multi-physical sensor design and the data acquisition system; The physical structure system of the coal mining machine is a comprehensive test bench for adaptive cutting control of coal mining machines, designed based on similarity theory and cutting experiments. The design method of the multi-physical sensor is as follows: by analyzing the load law of key components of the coal mining machine cutting section through dynamic analysis, the sensing force points are determined, and the layout of the multi-physical sensor is designed according to the sensing force points; the sensors are arranged according to the layout of the multi-physical sensor, and based on the sensing method of the cutting state of the coal mining machine in complex interbedded thin coal seams, a spiral drum cutting vibration sensing system, a square head vibration sensing system, and a rocker arm vibration sensing system are constructed to obtain the attitude parameter information of the coal mining machine in complex interbedded thin coal seams in real time, and the cutting state of the thin coal seam coal mining machine is identified by using the multi-source data information collected in real time by each physical sensor during the coal mining machine cutting process, so as to realize the dynamic monitoring of the coal mining machine cutting complex thin coal seams. The attitude parameter information includes: traction speed, rotation speed, and rocker arm adjustment height; The attitude parameter information of the complex interbedded thin coal seam mining machine and the multi-source data information collected by multiple physical sensors are used; wherein the attitude parameter information is used as the working status information of the complex interbedded thin coal seam mining machine. The data acquisition system is used to monitor and record the operating parameters of the coal mining machine adaptive cutting control integrated test bench and the working status data collected in real time by multiple physical sensors.

5. The method for constructing an adaptive cutting and control system for a coal mining machine based on digital twins according to claim 4, characterized in that, The method for constructing the virtual simulation layer is as follows: coal and rock samples are collected from typical working faces and their physical and mechanical properties are tested. The physical parameters of coal and rock mass based on discrete element method are determined and a discrete element model of coal and rock mass is constructed. Based on application programming interface (API) technology, the discrete element model of coal and rock mass is corrected by compiling a replacement API for multiple particle type clusters. A high-precision three-dimensional twin model of dynamic complex interbedded thin coal seam with real-time correction and replacement of particle sets is constructed. Based on the attitude parameter information of the coal mining machine in the complex interbedded thin coal seam acquired in real time, the cutting action of the coal mining machine is simulated and the dynamic complex interbedded thin coal seam particle set is corrected and replaced in real time. Based on the adaptive control strategy of the coal mining machine in the complex interbedded thin coal seam, an adaptive control model is constructed using multi-source data information collected by multiple physical sensors. The adaptive control model is used to generate the optimal adaptive cutting control strategy of the coal mining machine. The adjustment action of the coal mining machine is virtually executed according to the generated optimal adaptive cutting control strategy, and the expected working state information is generated. The virtually executed adjustment actions are fed back to the physical perception layer to guide the physical structure system of the coal mining machine to perform actions, obtain the working status information after the action, evaluate the effectiveness of the adaptive cutting optimal control strategy by comparing the expected working status information and the working status information after the action, adjust the adaptive cutting optimal control strategy according to the evaluation results, and then virtually execute the adjustment action of the coal mining machine again to form a closed-loop control process. A rigid-flexible coupled virtual prototype model of a coal mining machine was established to cut coal and rock. The hydraulic system and electrical control system of the coal mining machine were integrated into the virtual simulation layer. Interface technology was used to realize the interconnection between the models, determine the input and output of each model, and perform joint simulation to simulate the cutting process of a coal mining machine in a complex thin coal seam with interbedded gangue.

6. The method for constructing an adaptive cutting and control system for a coal mining machine based on digital twins according to claim 5, characterized in that, The construction method of the twin decision layer is as follows: using the discrete element model of coal and rock mass constructed in the virtual simulation layer as the basis, the test model of coal mining machine cutting coal seam is constructed by changing the position, thickness and rock properties of the interbedded gangue layer in the discrete element model of coal and rock mass in the virtual simulation layer. The test model is then used to analyze the stress law, vibration characteristics and cutting tooth state of the spiral drum cutting coal and rock in the coal mining machine. A process model of coal and rock crushing, collapse, flow and interface pressure formation during the coal mining machine system cutting interbedded gangue coal and rock is also constructed for coal and rock crushing, collapse, flow and interface pressure formation during the coal mining machine cutting process. Based on the model of the formation process of coal and rock crushing, collapse, flow and coupling interface pressure during the cutting of interbedded coal and rock by the coal mining machine system, a cutting state classification and identification model is constructed using the cutting state perception method of coal mining machine in complex interbedded thin coal seams, which is used to monitor the cutting state of thin coal seam coal mining machine. During the cutting process of a coal mining machine in a complex thin coal seam with interbedded rock, the characteristics of the coal and rock flow velocity field are obtained by extracting the running speed and displacement parameters of coal and rock particles. The dynamic load distribution of the coal mining machine is simulated by optimizing the running speed and displacement parameters of coal and rock particles. Based on the characteristics of the coal and rock flow velocity field and the dynamic load distribution of the coal mining machine, a virtual simulation layer is used to simulate the cutting process of a coal mining machine in a complex thin coal seam with interbedded rock. The cutting state classification and identification model is used to identify the cutting state of the coal mining machine in real time. Based on the identification results and the working state information obtained from the physical perception layer, the cutting parameters and cutting path are dynamically adjusted to optimize the cutting process of the coal mining machine in a complex thin coal seam with interbedded rock.

7. The method for constructing an adaptive cutting and control system for a coal mining machine based on digital twins according to claim 6, characterized in that, The data interaction layer is constructed as follows: For the multiple physical sensors set in the physical perception layer, each physical sensor is treated as a sensor node. A cluster head node is selected, and a clustering strategy is used to divide all sensor nodes into different clusters. A network topology adapted to the underground scenario is designed. A routing protocol is formulated based on the network topology and clustering strategy. A multi-objective algorithm is used to optimize the routing protocol. The network topology is optimized by clustering, optimizing cluster heads, and constructing a fitness function model. The optimized network topology is used as the communication structure of the data interaction layer. The routing protocol is loaded by reading the external file of the import model of the real-time information of coal and rock cutting status.

8. The method for constructing an adaptive cutting and control system for a coal mining machine based on digital twins according to claim 7, characterized in that, The method for constructing the adaptive cutting dynamic reliability digital twin monitoring model of the coal mining machine in complex interbedded thin coal seams is as follows: The cutting process of the coal mining machine under different coal seam conditions is simulated by using a rigid-flexible coupled virtual prototype model of the coal mining machine in the virtual simulation layer. The structural characteristics of the coal mining machine in real-time service are obtained based on the simulation results. The structural characteristics are used to identify the areas of interest of key components of the coal mining machine cutting section, and the state functions of the maximum stress and maximum amplitude of the key components are established. Based on the state functions of the maximum stress and maximum amplitude of key components, the stress reliability, frequency reliability and amplitude reliability of key components are evaluated by stress-intensity interference theory, resonance failure theory and reliability sensitivity theory respectively. Then, the reliability of the design variables of key components of the coal mining machine cutting section is evaluated by sensitivity analysis and used as the reliability sensitivity index of the coal mining machine adaptive cutting system. Based on the Copula function, the failure mode correlation and key component correlation are established by using historical failure data and simulation results of key components of the coal mining machine cutting section. The reliability of key components and the reliability of the coal mining machine adaptive cutting system are obtained, and the reliability of key components and the reliability of the coal mining machine adaptive cutting system are used as the reliability indicators of the coal mining machine adaptive cutting system. Based on the simulation results, the equivalent stress index of the key components of the coal mining machine's cutting section and the dynamic characteristic index of the coal mining machine were obtained. A comprehensive performance evaluation model for dynamic reliability of a coal mining machine during adaptive cutting of complex interbedded coal and rock was constructed using equivalent stress indices of key components in the cutting section, state functions of maximum stress and maximum amplitude of key components, stress reliability, frequency reliability and amplitude reliability of key components, system reliability sensitivity index, system reliability index, and dynamic characteristic index of the coal mining machine. Based on the theory of genetic algorithm, genetic encoding, fitness analysis and multi-generation evolution were performed on the comprehensive performance evaluation model for dynamic reliability of the coal mining machine during adaptive cutting of complex interbedded coal and rock to obtain the inherent frequencies of key components of the coal mining machine, which were used to optimize key components of the cutting section and the adaptive cutting system of the coal mining machine. Historical cutting data, cutting parameters, and reliability evaluation data of the coal mining machine are acquired and fused through multi-domain collaborative control technology. The fused data is then used to train a particle swarm optimization long short-term memory (PSO-LSTM) model to obtain a predictive model for evaluating the reliability of key components and the adaptive cutting system of the coal mining machine. A digital twin monitoring model for the dynamic reliability of adaptive cutting of coal mining machines in complex interbedded thin coal seams was built using this prediction model. This model is used to optimize key components and the adaptive cutting system of the coal mining machine based on the evaluation results of the prediction model, so as to realize real-time evaluation and monitoring of the dynamic reliability of key components and the adaptive cutting of the coal mining machine during the adaptive cutting process of coal mining machines in complex interbedded thin coal seams.

Images