Mining electric shovel working device health monitoring method based on multidisciplinary digital twinning

Abstract

The invention provides a mining electric shovel working device health detection method based on multidisciplinary digital twinning, and the method comprises the steps: obtaining attitude data through employing a Latin hypercube sampling method; performing electric shovel simulation by using discrete element and finite element methods, and further obtaining a data set required for neural network training; training and testing the LSTM neural network and the RBF neural network by using the data set, and constructing a mining load prediction model and a mining stress prediction model; constructing a digital twinborn model by using a tilt angle sensor, a stay wire encoder and the virtual model; a laser radar is used for scanning the terrain, the material mass is calculated and obtained, the load and stress of data are sequentially predicted through a neural network model, the stress is transmitted to a digital twinborn model, and real-time monitoring is achieved in a color rendering mode. According to the mining electric shovel, the stress change of the working device is displayed in real time in the working process of the mining electric shovel, and the early warning capacity of the mining electric shovel when excavation exceeds the stress safety range in the complex working environment is improved.

Description

Technical Field

[0001] The present invention relates to the field of computer artificial intelligence technology, and in particular to a health monitoring method for a mining electric shovel working device based on multidisciplinary digital twins. Background Art

[0002] With the continuous development of intelligent mining technology, the structural health monitoring technology of mining electric shovels has made certain progress. However, when dealing with complex working conditions and dynamic operating environments, it still faces the technical bottleneck of being unable to directly measure the stress of the working device of the mining electric shovel.

[0003] On the one hand, the excavation environment of an electric shovel is harsh, and it interacts directly with the rock and soil, making it difficult to directly monitor the structural health of the working device, such as key data such as stress. Sensors are easily interfered with in harsh environments, making it difficult to accurately and stably obtain stress information, and the installation and maintenance costs are high. On the other hand, the excavation process involves the intersection of multiple disciplines such as mechanical, electrical, and geotechnical. The coupling relationship between these disciplines is complex, which poses a great challenge to high-precision indirect measurement modeling. Traditional modeling methods are difficult to accurately describe this complex relationship, affecting the reliability of monitoring. In addition, some current monitoring systems can only monitor situations such as the detachment of the electric shovel bucket, and are insufficient for deep-level health monitoring such as the stress state of the internal structure of the working device. They are also weak in the early warning capability of exceeding the safe stress range, and are unable to promptly detect potential structural failure risks. It is difficult to meet the needs of accurate and real-time structural health monitoring of mining electric shovels under complex working conditions. Summary of the Invention

[0004] The present invention provides a health monitoring method for the working device of a mining electric shovel based on multidisciplinary digital twins. By integrating multidisciplinary collaborative modeling such as mechanical, electrical, and geotechnical, and combining sensor data drive, it realizes real-time visual monitoring of the health of the working structure of the mining electric shovel, improves the timely warning capability of exceeding the safe stress range during the operation of the mining electric shovel, and provides predictive maintenance and intelligent management for the mining electric shovel.

[0005] In the first aspect, the present invention provides a health detection method for a mining electric shovel working device based on a multidisciplinary digital twin, comprising the following steps: S1, according to the posture change range θ∈(θ min ,θ max ), r∈(r min ,r max ), Latin hypercube sampling method is used to obtain the posture data [θ, r] of the mining shovel working device; S2, the posture data [θ, r] is converted into time series data [θ1, r1], ..., [θ n ,r n], in the discrete element simulation tool, set the material parameters of the material particles and bucket, the shape and size of the material particles and their position distribution, and the contact between the material and the bucket. According to the time series data, the discrete element simulation is performed to obtain the lateral excavation load F of the bucket in contact with the soil. t and longitudinal excavation load F n and the material mass m; S3. In the finite element simulation tool, set the material of the mining shovel working device, perform the rotation and extension settings of the mining shovel working device, further perform meshing, use regular tetrahedron meshes, combine the load data obtained by discrete elements, apply the lateral and longitudinal excavation loads and material mass on the bucket, set the parameter set of the finite element simulation tool, and convert the posture data [θ, r] of the mining shovel working device, the material mass m, and the lateral excavation load F t and longitudinal excavation load F n As the input parameter of the data set, the node stress data σ of the mining electric shovel working device is set as the output parameter, and the finite element simulation analysis is performed to collect two data sets for use by the LSTM neural network and the RBF neural network, including the first data set and the second data set; S4, the posture data and material quality in the first data set are used as the input of the LSTM neural network, recorded as The horizontal mining load and the vertical mining load are output from the LSTM neural network and are recorded as mining load data [F t ,F n ], the first data set is divided into a training set and a test set, the LSTM neural network is trained by the training set, the accuracy of the LSTM neural network is tested by the test set, and then the mining load prediction model is constructed based on the LSTM neural network; S5, the posture data, material mass, and horizontal and vertical mining loads in the second data set are used as the input of the RBF neural network, denoted as [θ, r, m, F t ,F n ], the node stress data σ is used as the output of the RBF neural network, the second data set is divided into a training set and a test set, the RBF neural network is trained by the training set, the accuracy of the RBF neural network is tested by the test set, and then the mining stress prediction model is constructed based on the RBF neural network; S6, the inclination sensor and the pull-wire encoder are respectively installed on the bucket arm and the push-pull device of the mining electric shovel working device, and the motion posture data of the mining electric shovel working device are obtained through the inclination sensor and the pull-wire encoder The motion posture data is transmitted through the GATT protocol Input the virtual model of the mining electric shovel working device and the working environment into the game engine to achieve synchronization between the virtual model and the real electric shovel, and complete the construction of the digital twin model of the mining electric shovel working device; S7, install a laser radar on the side of the cab of the mining electric shovel working device to obtain the material surface point cloud, generate the material surface equation, perform difference integration on the material surface equation and the given bucket trajectory surface equation to obtain the material mass; S8, input the posture data and material mass into the excavation load prediction model to obtain the excavation load data, input the excavation load data, posture data and material mass into the excavation stress prediction model to generate node stress data, and based on the node stress data, use the digital twin model to render and display the stress changes of the mining electric shovel working device in real time to achieve real-time health monitoring of the mining electric shovel working device.

[0006] In a second aspect, the present invention provides an electronic device comprising a processor and a memory storing a program, wherein the program comprises instructions, which, when executed by the processor, cause the processor to execute the steps of the method according to the first aspect.

[0007] In a third aspect, the present invention provides a computer storage medium having a computer program stored thereon, which implements the method of the first aspect when executed by a processor.

[0008] The present invention uses Latin hypercube sampling to obtain simulation preparation data, namely the posture data of the mining shovel working device; uses discrete element and finite element methods to simulate the shovel excavator, thereby obtaining the data set required for neural network training; uses the data set to train and test the LSTM neural network and the RBF neural network to construct excavation load prediction models and excavation stress prediction models; uses inclination sensors, wire encoders, and virtual models to construct a digital twin model; uses lidar to scan the terrain and calculate the material quality, and uses the data to predict the load and stress in turn through the neural network model. The stress is then transmitted to the digital twin model for real-time monitoring using color rendering. The present invention realizes the real-time display of stress changes in the working device of the mining shovel during operation, which is of great significance for ensuring early warning of dangerous situations during the operation of the shovel and improving the work quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0010] Figure 1 is a flow chart of the steps of the present invention;

[0011] Figure 2For Figure 1 The corresponding overall flow chart;

[0012] Figure 3 This is the LSTM neural network structure diagram;

[0013] Figure 4 It is the structure diagram of RBF neural network;

[0014] Figure 5 Schematic diagram of the installation locations of sensors, encoders, and lidar;

[0015] Figure 6 This is a simplified diagram for material mass calculation;

[0016] Figure 7 This is a simplified diagram of the working device of a mining electric shovel. DETAILED DESCRIPTION

[0017] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0018] It should be noted that the brief descriptions of terms in the present invention are only for the purpose of facilitating the understanding of the embodiments described below, and are not intended to limit the embodiments of the present invention. Unless otherwise specified, these terms should be understood according to their ordinary and usual meanings.

[0019] In the present specification, claims, and drawings, the terms "first," "second," "third," and the like are used to distinguish similar or similar objects or entities and are not necessarily intended to define a particular order or precedence, unless otherwise noted. It should be understood that the terms used in this manner are interchangeable where appropriate.

[0020] The terms "comprise," "include," and "have," and any variations thereof, are intended to cover but not exclude inclusion; for example, a product or device comprising a list of components is not necessarily limited to all the components expressly listed but may include other components not expressly listed or inherent to such product or device.

[0021] This paper proposes a health detection method for the working device of a mining electric shovel based on multidisciplinary digital twins to improve the early warning capability of mining electric shovels exceeding the stress safety range in complex working environments.

[0022] First, to efficiently cover the full range of motion postures of a mining shovel, the present invention employs a Latin hypercube sampling method for sample collection. This method, through a stratified balanced sampling strategy, ensures comprehensiveness of posture parameters (including dipper arm inclination angle and thrust displacement) while significantly reducing the number of samples, thereby fully reflecting the characteristics of actual operating conditions. Furthermore, discrete element modeling software (EDEM) is used to construct a material particle model, simulating the particle-level mechanical properties of soil and rock, and accurately capturing load data such as the interaction forces between the bucket and the material. Using the discrete element modeling software, given posture sample data, data is captured regarding the excavation load and interaction between the mining shovel and the material during excavation. Finite element modeling software (Ansys) is used to construct a structural model of the mining shovel, simulate the stress-strain characteristics of its components, and calculate its nodal stress data. Using the finite element modeling software, given the excavation load, posture sample data, and material mass, nodal stress data is obtained for the mining shovel's working device. The collected data includes motion information of the mining shovel (such as the angle, angular velocity, elongation, and elongation velocity of the working device), material mass, excavation load data, and nodal stress data. After completing data collection, the obtained mining data, material quality and load data are processed based on the requirements of model training and testing. According to the input and output of the following two neural networks, the data is divided into two data sets, where data set 1 corresponds to the LSTM neural network model and data set 2 corresponds to the RBF neural network model.

[0023] Next, the data from the two datasets was divided, with 80% used as the training set and 20% as the test set. Two neural network models were constructed: an LSTM neural network model and an RBF neural network model. These neural networks were trained and validated using the datasets, capturing the inherent connections between input and output during the mining process, extracting key feature data, and forming accurate and efficient feature maps. The LSTM neural network model used the training set from dataset 1 to learn the nonlinear mapping relationship between the mining shovel's motion information and load response, and the accuracy of this mapping relationship was verified using the test set. The RBF neural network model used the training set from dataset 2 to learn the nonlinear mapping relationship between the mining shovel's motion information, load, and nodal stress, and the accuracy of this mapping relationship was verified using the test set. Based on these trained neural networks, highly accurate mining load and mining stress prediction models were constructed.

[0024] Finally, the mining shovel is equipped with an inclination sensor, a cable encoder, and a lidar. The inclination sensor and cable encoder acquire the shovel's posture data in real time. This data is transmitted to a virtual model of the mining shovel and the working environment within the virtual game engine via the GATT protocol of the Unity BLE plug-in, synchronizing the virtual model with the real-world shovel to form a digital twin. The lidar scans the material surface to obtain the material surface equation. Combined with the given excavation trajectory equation, the material mass is calculated using differential integration. The posture data and material mass are input into the excavation load prediction model to obtain load data. Furthermore, the load data, posture data, and material mass are input into the excavation stress prediction model to predict node stress data. The node stress data and other data are then input into the digital twin model, and the node stress data is rendered onto the mining shovel, enabling real-time monitoring of the mining shovel's stress conditions.

[0025] See also Figure 1 、 Figure 2 The present invention provides a health detection method for a mining shovel working device based on multidisciplinary digital twins, which can be specifically divided into the following steps:

[0026] S1, according to the posture change range θ∈(θ min ,θ max ), r∈(r min ,r max ), Latin hypercube sampling method is used to obtain the posture data [θ, r] of the mining shovel working device;

[0027] S2, convert the posture data [θ, r] into time series data [θ1, r1], ..., [θ n ,r n ], in the discrete element simulation tool, set the material parameters of the material particles and bucket, the shape and size of the material particles and their position distribution, and the contact between the material and the bucket. According to the time series data, the discrete element simulation is performed to obtain the lateral excavation load F of the bucket in contact with the soil. t and longitudinal excavation load F n and the material mass m;

[0028] S3. In the finite element simulation tool, set the material of the mining shovel working device, perform the rotation and extension settings of the mining shovel working device, further perform meshing, use regular tetrahedron meshes, combine the load data obtained by discrete elements, apply the lateral and longitudinal excavation loads and material mass on the bucket, set the parameter set of the finite element simulation tool, and convert the posture data [θ, r] of the mining shovel working device, the material mass m, and the lateral excavation load F t and longitudinal excavation load F nAs the input parameter of the data set, the node stress data σ of the mining shovel working device is set as the output parameter, and finite element simulation analysis is performed to collect two data sets for use by the LSTM neural network and the RBF neural network, including the first data set and the second data set, namely data set 1 and data set 2;

[0029] S4. The posture data and material quality in the first data set are used as the input of the LSTM neural network, denoted as The horizontal mining load and the vertical mining load are output from the LSTM neural network and are recorded as mining load data [F t ,F n ], the first data set is divided into a training set and a test set, the LSTM neural network is trained by the training set, the accuracy of the LSTM neural network is tested by the test set, and then the mining load prediction model is constructed based on the LSTM neural network;

[0030] S5. The posture data, material mass, and lateral and longitudinal mining loads in the second data set are used as the input of the RBF neural network, denoted as [θ, r, m, F t ,F n ], the node stress data σ is used as the output of the RBF neural network, the second data set is divided into a training set and a test set, the RBF neural network is trained with the training set, the accuracy of the RBF neural network is tested with the test set, and then a mining stress prediction model is constructed based on the RBF neural network;

[0031] S6. Install an inclination sensor and a cable encoder on the bucket arm and push-pull device of the mining electric shovel working device respectively, and obtain the motion posture data of the mining electric shovel working device through the inclination sensor and cable encoder. The motion posture data is transmitted through the GATT protocol Input the virtual model of the mining shovel working device and the working environment into the game engine, synchronize the virtual model with the real shovel, and complete the construction of the digital twin model of the mining shovel working device;

[0032] S7. Install a laser radar on the side of the cab of the mining electric shovel working device to obtain a material surface point cloud, generate a material surface equation, perform a difference integral between the material surface equation and a given bucket trajectory surface equation, and obtain the material mass;

[0033] S8. Input the posture data and material mass into the mining load prediction model to obtain the mining load data. Input the mining load data, posture data and material mass into the mining stress prediction model to generate the node stress data. Based on the node stress data, the stress changes of the mining electric shovel working device are displayed in real time through the digital twin model to realize the real-time health monitoring of the mining electric shovel working device.

[0034] Optionally, the first data set is divided into a training set and a test set, where 80% of the data is the training set and 20% of the data is the test set. The training set is put into the LSTM neural network for training, and the accuracy of the trained LSTM neural network is tested with the data of the test set:

[0035]

[0036] Where x is the first data set input, θ represents the rotation angle of the stick, represents the angular velocity of the boom, r represents the extension of the boom, represents the extension speed of the bucket arm, m represents the material mass, y is the output of the first data set, F t Indicates the longitudinal excavation load, F n represents the lateral excavation load;

[0037] Based on deep learning theory, an LSTM neural network model is constructed, and the mapping relationship between input and output is established as follows:

[0038] y0,...,y t-1 =f a (x0,...,x t-1 )

[0039] After training and testing, the mining load prediction model is constructed as follows:

[0040]

[0041] in, Represents the actual input of the LSTM neural network, represents the LSTM neural network prediction value, t represents the time step, and f a Represents an LSTM neural network model.

[0042] Optionally, the second data set is divided into a training set and a test set, where 80% of the data is the training set and 20% of the data is the test set. The training set is put into the RBF neural network for training, and the accuracy of the trained RBF neural network is tested with the data of the test set:

[0043] u=[θ,r,m,F t ,F n ],v=σ

[0044] Where u is the second data set input, θ represents the rotation angle of the bucket arm, r represents the extension of the bucket arm, m represents the material mass, and F t Indicates the longitudinal excavation load, F n represents the lateral excavation load, v is the output of the second data set, and σ is the nodal stress data during the excavation process;

[0045] Based on deep learning theory, the RBF neural network model is constructed, and the mapping relationship between input and output is established as follows:

[0046] v=f c (u)

[0047] After training and testing, the excavation stress prediction model constructed is as follows:

[0048]

[0049] in, Represents the actual input of the RBF neural network, represents the predicted value of RBF neural network, f c Represents the RBF neural network model.

[0050] Optionally, the diagonal sensor and the cable encoder are installed on the bucket arm and the push-pull device of the mining electric shovel working device respectively, and the motion posture data of the mining electric shovel working device is obtained through the diagonal sensor and the cable encoder. The invention comprises: installing an inclination sensor on the side of the bucket baffle of the mining electric shovel working device to obtain a rotation angle signal of the bucket arm; installing a pull wire encoder between the fixed end and the movable end of the pushing mechanism of the mining electric shovel working device to obtain an elongation signal of the bucket arm, wherein the rotation angle signal of the bucket arm includes the rotation angle of the bucket arm and the angular velocity of the bucket arm, and the elongation signal of the bucket arm includes the elongation amount of the bucket arm and the elongation speed of the bucket arm.

[0051] Optionally, the motion posture data is transmitted via the GATT protocol The virtual mining electric shovel working device and the virtual model of the working environment are input into the game engine to achieve synchronization between the virtual model and the real electric shovel, including: transmitting the rotation angle signal of the bucket arm and the extension signal of the bucket arm to the game engine through the GATT protocol; using virtual simulation software in the game engine to build a virtual mining electric shovel working device and a virtual model of the working environment used by it; inputting the rotation angle signal of the bucket arm and the extension signal of the bucket arm into the virtual model of the working environment and converting them into motion control instructions of the lifting motor and the pushing motor respectively, so as to drive the virtual mining electric shovel working device to move synchronously with the actual mining electric shovel working device.

[0052] Optionally, the material surface equation and the given bucket track surface equation are differentially integrated to obtain the material mass. The material mass calculation formula is:

[0053]

[0054] Among them, ρ is the material density, and the material surface equation is expressed as f s(x,y), the bucket track surface equation is expressed as f tr (x,y).

[0055] Optionally, the real-time rendering and display of stress changes of the mining electric shovel working device based on the node stress data through the digital twin model includes: inputting the node stress data into the digital twin model, and rendering the stress by gradually transitioning from small to large using different colors, to obtain a cloud map that changes in real time on the game engine to monitor the stress changes of the mining electric shovel working device during the excavation process.

[0056] Specifically, the solution of the present invention is further described according to the following examples:

[0057] 1. Detailed steps for dataset construction:

[0058] 1.1 Using the Latin hypercube sampling method, according to the change range of the mining shovel posture θ∈(θ min ,θ max ), r∈(r min ,r max ), obtaining multiple posture samples, namely posture data [θ, r], to provide data support for discrete element simulation. Here, θ represents the rotation angle of the mining shovel bucket arm, and r represents the extension of the mining shovel bucket arm.

[0059] 1.2 Calculate each posture sample and the initial posture to obtain the time series data [θ1,r1],...,[θ n ,r n In the discrete element software EDEM, the material surface is established, the material parameters of the material and working device, the shape and size of the material, the position distribution, and the contact properties between the material and the bucket are set, and the time series data is given to simulate and calculate the lateral excavation load F of the mining electric shovel. n , longitudinal excavation load F t and material quality m .

[0060] 1.3 Set the parameter set in the finite element software Ansys, where [θ, r, F n ,F t ,m] as input, and the nodal stress data σ of the mining shovel as output. The working device's rotation and extension settings were set, and the mesh was further divided into tetrahedral meshes with a cell size of 50 mm. Combined with the load data obtained by discrete element method, the lateral and longitudinal excavation loads on the bucket, as well as the material mass, were applied. The nodal stress data σ of the mining shovel's working device during the excavation process was calculated under the given operating parameters.

[0061] 1.4 Derivative of the data [θ, r] to obtain The data obtained from discrete element and finite element simulations were classified and sorted to construct two data sets. Data set 1 is is input, output [F t ,F n ]; Dataset 2 is [θ,r,F t ,F n ,m] is the input and the output is σ. is the angular velocity of the shovel arm, The extension speed of the electric shovel bucket arm.

[0062] 2. Detailed steps for mining load prediction model training:

[0063] 2.1 The data in Dataset 1 is used for training and testing the LSTM neural network. Dataset 1 is randomly divided into two parts: 80% of the data is put into the training set for training the neural network model, and 20% of the data is put into the test set for testing the accuracy of the trained neural network prediction.

[0064]

[0065] Among them, x is the input of data set 1, θ represents the rotation angle of the stick, represents the angular velocity of the boom, r represents the extension of the boom, represents the extension speed of the bucket arm, m represents the material mass, y is the output of data set 1, F t Indicates the longitudinal excavation load, F n Indicates the lateral excavation load.

[0066] 2.2 Based on deep learning theory, a neural network model is constructed and the mapping relationship between input and output is established as follows:

[0067] y0,...,y t-1 =f a (x0,...,x t-1 )

[0068] The training set data is input into the LSTM neural network. The neural network learns from the training set data and continuously iterates and optimizes, gradually approaching the desired objective function. The test set data is input into the trained neural network to verify the accuracy of the neural network's predicted data. The mining load prediction model is further constructed as follows:

[0069]

[0070] in, Represents the actual input of the LSTM neural network, represents the LSTM neural network prediction value, t represents the time step, and f a Represents an LSTM neural network model, such as Figure 3 shown.

[0071] 3. Detailed steps for constructing the mining stress prediction model:

[0072] 3.1 The data in Dataset 2 are used for training and testing of RBF neural network. Dataset 2 is randomly divided, with 80% of the data placed in the training set for training the neural network model and 20% of the data placed in the test set for testing the accuracy of the trained neural network prediction.

[0073] u=[θ,r,m,F t ,F n ],v=σ

[0074] Among them, u is the input of data set 2, θ represents the rotation angle of the bucket arm, r represents the extension of the bucket arm, m represents the material mass, F t Indicates the longitudinal excavation load, F n represents the lateral excavation load, v is the output of dataset 2, and σ is the nodal stress data during the excavation process.

[0075] 3.2 Based on deep learning theory, a neural network model is constructed and the mapping relationship between input and output is established as follows:

[0076] v=f c (u)

[0077] The training set data is input into the RBF neural network. The neural network learns from the training set data and continuously iterates and optimizes, gradually approaching the desired objective function. The test set data is input into the trained neural network to verify the accuracy of the neural network's prediction data and further build the mining stress prediction model.

[0078]

[0079] in, Represents the actual input of the RBF neural network, represents the predicted value of RBF neural network, f c Represents the RBF neural network model, such as Figure 4 shown.

[0080] 4. Detailed steps for building a health monitoring system for mining shovel operation:

[0081] 4.1 Install the inclination sensor and the wire encoder on the side of the bucket arm baffle and between the fixed end and the moving end of the pushing mechanism of the mining electric shovel to record the movement posture data of the electric shovel in real time. Signals are then transmitted over the network and the GATT protocol in the BLE plug-in in Unity, transferring the posture data to the virtual model of the mining shovel and terrain in Unity, thus connecting the digital model with the physical mining shovel. The virtual model is then debugged to ensure consistent motion with the mining shovel, completing the construction of the digital twin model. Figure 5 Installation location diagram of the inclination sensor, wire encoder and lidar.

[0082] 4.2 Install a laser radar on the body of the mining electric shovel. The laser radar can obtain the material surface point cloud and further reconstruct the material surface equation. The material mass calculation is to obtain the material mass m by performing the difference integral between the pile surface equation and the given bucket track surface equation, as shown in the following example: Figure 6 shown.

[0083] The mathematical expression for material mass calculation is:

[0084]

[0085] Where, ρ is the material density, and the material surface equation is expressed as f s (x,y), the bucket track surface equation is expressed as f tr (x,y).

[0086] 4.3 The inclination sensor and the wire encoder installed on the mining electric shovel are used to obtain the posture data x1 of the mining electric shovel during the excavation process, and the bucket arm angle, bucket arm angular velocity, bucket arm extension amount and bucket arm extension speed are obtained.

[0087]

[0088] 4.4 Combine the mining shovel posture data with the material quality As input, it is imported into the excavation load prediction model to predict the lateral excavation load F n and longitudinal excavation load F t .

[0089] 4.5 The mining shovel posture data, material mass and excavation load u=[θ,r,m,F t ,F n ] is used as input and imported into the excavation stress prediction model to predict the node stress data σ of the mining electric shovel working device.

[0090] 4.6 The acquired node stress data of the shovel working device is transferred to the digital twin model. Stress changes are displayed in real time in the digital twin model through color rendering, enabling real-time monitoring of the shovel working device's structural health and early warning of stress exceeding the safe stress range. Figure 7 This is a simplified diagram of the working posture of an electric shovel.

[0091] This invention utilizes modeling and simulation methods, integrating geotechnical mechanics, structural mechanics, and computer science to collaboratively achieve accurate modeling of the multidisciplinary excavation process involving mechanical, electrical, and geotechnical disciplines. Combined with data from inclination sensors and wire encoders, it accurately characterizes the dynamic response of mining shovels during operation and indirectly measures stress at the working device nodes, thus resolving the technical bottleneck of stress measurement. This solution not only enhances the warning capability of mining shovels for exceeding safe stress ranges during structural health monitoring but also provides important support for predictive maintenance and intelligent management of mining shovels.

[0092] As another example, an embodiment of the present invention further provides an electronic device, which is an example of a hardware device that can be applied to various aspects of the present invention. The electronic device is intended to represent various forms of digital electronic computer equipment, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0093] The electronic device may include: a processor, a communication interface, a memory, and a communication bus.

[0094] The processor, communication interface, and memory communicate with each other through a communication bus. The communication interface is used to communicate with other electronic devices or servers.

[0095] The processor is used to execute the program, and specifically can execute the relevant steps in the above method embodiment.

[0096] Specifically, the program may include program codes including computer operation instructions.

[0097] The processor may be a CPU, an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present invention. The one or more processors included in the smart device may be processors of the same type, such as one or more CPUs, or processors of different types, such as one or more CPUs and one or more ASICs.

[0098] Memory is used to store programs. The memory may include high-speed RAM memory, and may also include non-volatile memory (non-volatile memory), such as at least one disk storage.

[0099] When the program is executed by a processor, it is used to enable the electronic device to execute the method of the present invention.

[0100] In addition, the specific implementation of each step in the program can refer to the corresponding description of the corresponding steps and units in the above method embodiments, and will not be repeated here. Those skilled in the art will clearly understand that for the convenience and brevity of description, the specific working process of the above-described devices and modules can refer to the corresponding process description in the above method embodiments, and will not be repeated here.

[0101] An exemplary embodiment of the present invention further provides a computer storage medium storing a computer program, wherein when the computer program is executed by a processor, the methods of the various embodiments of the present invention are implemented. The corresponding process descriptions in the aforementioned method embodiments can be referred to and will not be repeated here.

[0102] The method according to the embodiment of the present invention described above can be implemented in hardware, firmware, or as software or computer code that can be stored in a recording medium (such as a CD ROM, RAM, floppy disk, hard disk or magneto-optical disk), or as computer code that is originally stored in a remote recording medium or a non-temporary machine-readable medium downloaded via a network and will be stored in a local recording medium, so that the method described herein can be stored in such software processing on a recording medium using a general-purpose computer, a dedicated processor or programmable or dedicated hardware (such as an ASIC or FPGA). It can be understood that a computer, a processor, a microprocessor controller or programmable hardware includes a storage component (e.g., RAM, ROM, flash memory, etc.) that can store or receive software or computer code, and when the software or computer code is accessed and executed by a computer, a processor or hardware, the method described herein is implemented. In addition, when a general-purpose computer accesses the code for implementing the method shown here, the execution of the code converts the general-purpose computer into a dedicated computer for executing the method shown here.

[0103] Thus far, specific embodiments of the present invention have been described. Other embodiments are within the scope of the appended claims. In some cases, the actions recited in the claims can be performed in a different order and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order shown or sequential order to achieve the desired results. In certain embodiments, multitasking and parallel processing may be advantageous.

[0104] It should be understood that although this specification is described according to various embodiments, not every embodiment contains only one independent technical solution. This narrative method of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation methods that can be understood by those skilled in the art.

[0105] Finally, it should be noted that the above implementation methods are only used to illustrate the embodiments of the present invention, and are not limitations on the embodiments of the present invention. Ordinary technicians in the relevant technical field can make various changes and modifications without departing from the spirit and scope of the embodiments of the present invention. Therefore, all equivalent technical solutions also fall within the scope of the embodiments of the present invention, and the scope of patent protection of the embodiments of the present invention should be defined by the claims.

Claims

1. A health detection method for a mining shovel working device based on multidisciplinary digital twins, characterized in that: The following steps are involved: S1, according to the posture change range θ∈(θ min ,θ max ), r∈(r min ,r max ), Latin hypercube sampling method is used to obtain the posture data [θ, r] of the mining shovel working device; S2, convert the posture data [θ, r] into time series data [θ1, r1], ..., [θ n ,r n ], in the discrete element simulation tool, set the material parameters of the material particles and bucket, the shape and size of the material particles and their position distribution, and the contact between the material and the bucket. According to the time series data, the discrete element simulation is performed to obtain the lateral excavation load F of the bucket in contact with the soil. t and longitudinal excavation load F n and the material mass m; S3. In the finite element simulation tool, set the material of the mining shovel working device, perform the rotation and extension settings of the mining shovel working device, further perform meshing, use regular tetrahedron meshes, combine the load data obtained by discrete elements, apply the lateral and longitudinal excavation loads and material mass on the bucket, set the parameter set of the finite element simulation tool, and convert the posture data [θ, r] of the mining shovel working device, the material mass m, and the lateral excavation load F t and longitudinal excavation load F n As the input parameter of the data set, the node stress data σ of the mining electric shovel working device is set as the output parameter, and the finite element simulation analysis is performed to collect two data sets for use by the LSTM neural network and the RBF neural network, including the first data set and the second data set; S4. The posture data and material quality in the first data set are used as the input of the LSTM neural network, denoted as The horizontal mining load and the vertical mining load are output from the LSTM neural network and are recorded as mining load data [F t ,F n ], the first data set is divided into a training set and a test set, the LSTM neural network is trained by the training set, the accuracy of the LSTM neural network is tested by the test set, and then the mining load prediction model is constructed based on the LSTM neural network; S5. The posture data, material mass, and lateral and longitudinal mining loads in the second data set are used as the input of the RBF neural network, denoted as [θ, r, m, F t ,F n ], the node stress data σ is used as the output of the RBF neural network, the second data set is divided into a training set and a test set, the RBF neural network is trained with the training set, the accuracy of the RBF neural network is tested with the test set, and then a mining stress prediction model is constructed based on the RBF neural network; S6. Install an inclination sensor and a cable encoder on the bucket arm and push-pull device of the mining electric shovel working device respectively, and obtain the motion posture data of the mining electric shovel working device through the inclination sensor and cable encoder. The motion posture data is transmitted through the GATT protocol Input the virtual model of the mining shovel working device and the working environment into the game engine, synchronize the virtual model with the real shovel, and complete the construction of the digital twin model of the mining shovel working device; S7. Install a laser radar on the side of the cab of the mining electric shovel working device to obtain a material surface point cloud, generate a material surface equation, perform a difference integral between the material surface equation and a given bucket trajectory surface equation, and obtain the material mass; S8. Input the posture data and material mass into the mining load prediction model to obtain the mining load data. Input the mining load data, posture data and material mass into the mining stress prediction model to generate the node stress data. Based on the node stress data, the stress changes of the mining electric shovel working device are displayed in real time through the digital twin model to realize the real-time health monitoring of the mining electric shovel working device.

2. The method according to claim 1, characterized in that The first data set is divided into a training set and a test set, where 80% of the data is the training set and 20% of the data is the test set. The training set is put into the LSTM neural network for training, and the data of the test set is used to test the accuracy of the trained LSTM neural network: y=[F t ,F n ] Where x is the first data set input, θ represents the rotation angle of the stick, represents the angular velocity of the boom, r represents the extension of the boom, represents the extension speed of the bucket arm, m represents the material mass, y is the output of the first data set, F t Indicates the longitudinal excavation load, F n represents the lateral excavation load; Based on deep learning theory, an LSTM neural network model is constructed, and the mapping relationship between input and output is established as follows: y0,...,y t-1 =f a (x0,...,x t-1 ) After training and testing, the mining load prediction model is constructed as follows: in, Represents the actual input of the LSTM neural network, represents the LSTM neural network prediction value, t represents the time step, and f a Represents an LSTM neural network model.

3. The method according to claim 1, characterized in that The second data set is divided into a training set and a test set, where 80% of the data is the training set and 20% of the data is the test set. The training set is put into the RBF neural network for training, and the data of the test set is used to test the accuracy of the trained RBF neural network: u=[θ,r,m,F t ,F n ],v=σ Where u is the second data set input, θ represents the rotation angle of the bucket arm, r represents the extension of the bucket arm, m represents the material mass, and F t Indicates the longitudinal excavation load, F n represents the lateral excavation load, v is the output of the second data set, and σ is the nodal stress data during the excavation process; Based on deep learning theory, the RBF neural network model is constructed, and the mapping relationship between input and output is established as follows: v=f c (u) After training and testing, the excavation stress prediction model constructed is as follows: in, Represents the actual input of the RBF neural network, represents the predicted value of RBF neural network, f c Represents the RBF neural network model.

4. The method according to claim 1, wherein The inclination sensor and the wire encoder are respectively installed on the bucket arm and the push-pull device of the mining electric shovel working device, and the motion posture data of the mining electric shovel working device is obtained through the inclination sensor and the wire encoder. include: Install a tilt sensor on the side of the bucket baffle of the mining electric shovel working device to obtain the rotation angle signal of the bucket arm; A wire encoder is installed between the fixed end and the movable end of the pushing mechanism of the mining electric shovel working device to obtain the extension signal of the bucket arm. The rotation angle signal of the bucket arm includes the rotation angle of the bucket arm and the angular velocity of the bucket arm. The extension signal of the bucket arm includes the extension amount of the bucket arm and the extension speed of the bucket arm.

5. The method according to claim 4, characterized in that The motion posture data is transmitted through the GATT protocol Input the virtual mining shovel working device and the virtual model of the working environment into the game engine to achieve synchronization between the virtual model and the real shovel, including: Transmitting the rotation angle signal of the stick and the extension signal of the stick to the game engine via the GATT protocol; Use virtual simulation software in the game engine to build a virtual model of the mining shovel working device and the working environment in which it is used; The rotation angle signal of the bucket arm and the extension signal of the bucket arm are input into the virtual model of the working environment and converted into motion control instructions of the lifting motor and the pushing motor respectively, so as to drive the virtual mining electric shovel working device to move synchronously with the actual mining electric shovel working device.

6. The method according to claim 1, characterized in that The material surface equation and the given bucket track surface equation are differentially integrated to obtain the material mass. The material mass calculation formula is: Among them, ρ is the material density, and the material surface equation is expressed as f s (x,y), the bucket track surface equation is expressed as f tr (x,y).

7. The method according to claim 6, characterized in that The stress changes of the mining shovel working device are displayed in real time through the digital twin model based on the node stress data, including: The node stress data is input into the digital twin model, and the stress is rendered by gradually transitioning from small to large using different colors. A cloud map that changes in real time on the game engine is obtained to monitor the stress changes of the mining electric shovel working device during the excavation process.

8. An electronic device, characterized in that: include: processor; Memory for storing programs; The program includes instructions, which, when executed by the processor, cause the processor to perform the steps of the method according to any one of claims 1 to 7.

9. A computer storage medium, characterized in that A computer program is stored thereon, and when the program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images