Remote control method and system for underground filling equipment

Abstract

The present application discloses a remote control method and system for underground filling equipment, which relates to the technical field of equipment remote control. The method comprises: using a pressure sensor to monitor and collect pipeline pressure data in real time, and to monitor and collect filling material flow values ​​in real time, introducing a data processing and analysis algorithm based on artificial intelligence and deep learning in a remote control server to perform time series analysis and information propagation aggregation representation on the pipeline pressure time series data and the filling material flow time series data, and then using the core component fusion semantics between the pipeline pressure time series propagation aggregation characteristics and the filling material flow time series propagation aggregation characteristics to recommend the filling rate at the next time point, and using a variable frequency drive and a servo motor to achieve precise control of the filling pumping system, which can realize more intelligent remote control of underground filling equipment, thereby adaptively controlling the material filling rate based on the time series coordinated changes between the pipeline pressure and the filling material flow.

Description

Technical Field

[0001] This application relates to the technical field of remote control of equipment, and specifically, to a method and system for remotely controlling underground filling equipment. Background Art

[0002] With the continuous pursuit of safety and efficiency in the mining industry, underground filling operations have gradually become an important part of modern mine exploitation. Filling operations can not only effectively improve the resource utilization rate of mines, but also significantly reduce the risk of geological disasters caused by mine cavities. In underground mine filling operations, tailings or other filling materials need to be pumped into the mined-out area to maintain surface stability, control surrounding rock deformation, and recover valuable resources. This process requires precise control of the conveying rate of the filling materials to ensure the quality and stability of the filling body.

[0003] However, traditional filling operations usually rely on the experience of on-site operators and manual adjustment. The experience and judgment ability of operators directly affect the quality of the filling process. Different operators may adopt different operation strategies under the same conditions, resulting in inconsistent and uneven filling effects. In addition, manual adjustment of the filling process cannot respond in a timely manner to changes in on-site conditions. For example, a pipeline rupture may be caused by excessive pressure due to too fast filling rate. When pressure anomalies or flow instability occur, operators may not be able to react quickly, increasing potential safety hazards.

[0004] Therefore, an optimized underground filling control solution is needed to solve the above technical problems. Summary of the Invention

[0005] This Summary of the Invention section is provided to introduce concepts in a concise form that will be described in detail in the following Detailed Description section. This Summary of the Invention section is not intended to identify key features or essential features of the claimed technical solution, nor is it intended to be used to limit the scope of the claimed technical solution.

[0006] In a first aspect, this application provides a method for remotely controlling underground filling equipment, the method comprising:

[0007] Obtaining a time queue of pipeline pressure and a time queue of filling material flow rate values collected by a pressure sensor;

[0008] Transmitting the time queue of the pipeline pressure and the time queue of the filling material flow rate values to a remote control server through a wireless communication module;

[0009] At the remote control server, local temporal encoding is respectively performed on the time queue of the pipeline pressure and the time queue of the filling material flow rate values to obtain a sequence of local temporal correlation feature vectors of the pipeline pressure and a sequence of local temporal correlation feature vectors of the filling material flow rate;

[0010] The sequence of local temporal correlation feature vectors of the pipeline pressure and the sequence of local temporal correlation feature vectors of the filling material flow rate are input into a feature forward propagation network based on significant temporal decay of node energy to obtain a temporal propagation representation vector of the pipeline pressure and a temporal propagation representation vector of the filling material flow rate;

[0011] The temporal propagation representation vector of the pipeline pressure and the temporal propagation representation vector of the filling material flow rate are input into a query matching interaction network based on feature principal components to obtain a significantly fused representation vector of the core components of the pipeline pressure - filling material flow rate;

[0012] The significantly fused representation vector of the core components of the pipeline pressure - filling material flow rate is input into a remote control instruction generator based on a decoder to obtain a remote operation instruction, and the remote operation instruction includes a recommended decoded value of the filling rate at the next time point.

[0013] Optionally, at the remote control server, respectively performing local temporal encoding on the time queue of the pipeline pressure and the time queue of the filling material flow rate values to obtain a sequence of local temporal correlation feature vectors of the pipeline pressure and a sequence of local temporal correlation feature vectors of the filling material flow rate includes: at the remote control server, inputting the time queue of the pipeline pressure and the time queue of the filling material flow rate values into a sequence encoder based on a 1D - CNN model to obtain the sequence of local temporal correlation feature vectors of the pipeline pressure and the sequence of local temporal correlation feature vectors of the filling material flow rate.

[0014] Optionally, inputting the sequence of the local time-series correlation feature vectors of the pipeline pressure and the sequence of the local time-series correlation feature vectors of the filling material flow rate into a feature forward propagation network based on significant time-series decay of node energy to obtain a pipeline pressure time-series propagation representation vector and a filling material flow rate time-series propagation representation vector, including: calculating an energy significant descriptor of each local time-series correlation feature vector of the pipeline pressure based on the mean and standard deviation of each local time-series correlation feature vector of the pipeline pressure in the sequence of the local time-series correlation feature vectors of the pipeline pressure to obtain a sequence of local time-series energy significant descriptors of the pipeline pressure; extracting the timestamps of each local time-series correlation feature vector of the pipeline pressure in the sequence of the local time-series correlation feature vectors of the pipeline pressure to obtain a sequence of pipeline pressure timestamps; taking the local time-series correlation feature vector of the pipeline pressure corresponding to the current time point in the sequence of the local time-series correlation feature vectors of the pipeline pressure as the current local time-series correlation feature vector of the pipeline pressure and taking the local time-series correlation feature vectors of the pipeline pressure corresponding to other time points as the historical local time-series correlation feature vectors of the pipeline pressure to obtain a sequence of the current local time-series correlation feature vector of the pipeline pressure and the historical local time-series correlation feature vectors of the pipeline pressure; calculating an energy-time decay factor of each historical local time-series correlation feature vector of the pipeline pressure in the sequence of the historical local time-series correlation feature vectors of the pipeline pressure relative to the current local time-series correlation feature vector of the pipeline pressure to obtain a sequence of local time-series energy-time decay factors of the pipeline pressure; after calculating the element-wise division of the sequence of the local time-series energy significant descriptors of the pipeline pressure and the sequence of the local time-series energy-time decay factors of the pipeline pressure, taking the obtained local time-series energy weights of the pipeline pressure as weighting coefficients and calculating the element-wise weighted sum of the sequence of the historical local time-series correlation feature vectors of the pipeline pressure to obtain a local time-series historical node propagation aggregation representation feature vector of the pipeline pressure; calculating the element-wise weighted sum of the local time-series historical node propagation aggregation representation feature vector of the pipeline pressure and the current local time-series correlation feature vector of the pipeline pressure based on the energy significant descriptor corresponding to the current local time-series correlation feature vector of the pipeline pressure to obtain the pipeline pressure time-series propagation aggregation representation vector.

[0015] Optionally, based on the mean and standard deviation of each pipeline pressure local temporal correlation feature vector in the sequence of pipeline pressure local temporal correlation feature vectors, calculate the energy significant descriptor of each pipeline pressure local temporal correlation feature vector to obtain a sequence of pipeline pressure local temporal energy significant descriptors, including: calculating the mean and standard deviation of each pipeline pressure local temporal correlation feature vector in the sequence of pipeline pressure local temporal correlation feature vectors to obtain a sequence of pipeline pressure local means and a sequence of pipeline pressure local standard deviations; after performing position-wise difference on each pair of corresponding pipeline pressure local temporal correlation feature vector and pipeline pressure local mean in the sequence of pipeline pressure local temporal correlation feature vectors and the sequence of pipeline pressure local means, performing fourth-power modulation on each position eigenvalue in the obtained sequence of difference feature vectors to obtain a sequence of pipeline pressure local temporal modulation feature vectors; calculating the expected value of each pipeline pressure local temporal modulation feature vector in the sequence of pipeline pressure local temporal modulation feature vectors to obtain a sequence of pipeline pressure local temporal positive expected factors; calculating the fourth-power modulation value of each pipeline pressure local standard deviation in the sequence of pipeline pressure local standard deviations to obtain a sequence of pipeline pressure local temporal negative expected factors; dividing each pair of corresponding pipeline pressure local temporal positive expected factor and pipeline pressure local temporal negative expected factor in the sequence of pipeline pressure local temporal positive expected factors and the sequence of pipeline pressure local temporal negative expected factors to obtain the sequence of pipeline pressure local temporal energy significant descriptors.

[0016] Optionally, calculate the energy time decay factor of each historical pipeline pressure local temporal correlation feature vector in the sequence of historical pipeline pressure local temporal correlation feature vectors relative to the current pipeline pressure local temporal correlation feature vector to obtain a sequence of pipeline pressure local temporal energy time decay factors, including: respectively calculating the floor value after subtracting the pipeline pressure timestamp corresponding to the current pipeline pressure local temporal correlation feature vector from the pipeline pressure timestamps corresponding to each historical pipeline pressure local temporal correlation feature vector to obtain a sequence of time span values; after calculating the square of each time span value in the sequence of time span values, performing position-wise division of the obtained sequence of span square modulation values by the square of the time decay inverse scaling parameter and performing position-wise multiplication by the decay rate positive scaling parameter to obtain a sequence of temporal span energy decay coefficients; using each temporal span energy decay coefficient in the sequence of temporal span energy decay coefficients as the exponential power, calculating the value of the natural exponential function with the natural constant e as the base to obtain the sequence of pipeline pressure local temporal energy time decay factors.

[0017] Optionally, input the pipeline pressure time series propagation representation vector and the filling material flow rate time series propagation representation vector into a query matching interaction network based on feature principal components to obtain a significantly fused representation vector of the pipeline pressure - filling material flow rate core components, including: performing normalization processing on the pipeline pressure time series propagation representation vector and the filling material flow rate time series propagation representation vector to obtain a normalized pipeline pressure time series propagation representation vector and a normalized filling material flow rate time series propagation representation vector; calculating the sample covariance matrices of the normalized pipeline pressure time series propagation representation vector and the normalized filling material flow rate time series propagation representation vector to obtain a pipeline pressure time series sample covariance matrix and a filling material flow rate time series sample covariance matrix; performing eigenvector extraction based on matrix decomposition on the pipeline pressure time series sample covariance matrix and the filling material flow rate time series sample covariance matrix to obtain a set of pipeline pressure time series principal component eigenvectors and a set of filling material flow rate time series principal component eigenvectors; inputting the set of pipeline pressure time series principal component eigenvectors and the set of filling material flow rate time series principal component eigenvectors into a maximum approximate query matching network to obtain a set of best matching pairs of the pipeline pressure time series principal component eigenvectors and the filling material flow rate time series principal component eigenvectors; inputting each best matching pair of the pipeline pressure time series principal component eigenvectors and the filling material flow rate time series principal component eigenvectors in the set of best matching pairs into a semantic fine - grained gating joint module to obtain a set of pipeline pressure - filling material flow rate time series principal component fusion feature vectors; concatenating each pipeline pressure - filling material flow rate time series principal component fusion feature vector in the set of pipeline pressure - filling material flow rate time series principal component fusion feature vectors to obtain the significantly fused representation vector of the pipeline pressure - filling material flow rate core components; wherein, calculating the sample covariance matrices of the normalized pipeline pressure time series propagation representation vector and the normalized filling material flow rate time series propagation representation vector to obtain a pipeline pressure time series sample covariance matrix and a filling material flow rate time series sample covariance matrix includes: calculating the vector multiplication between the transposed vector of the normalized pipeline pressure time series propagation representation vector and the normalized pipeline pressure time series propagation representation vector, and then performing element - wise division by the number of eigenvalues of the normalized pipeline pressure time series propagation representation vector minus one to obtain the pipeline pressure time series sample covariance matrix; calculating the vector multiplication between the transposed vector of the normalized filling material flow rate time series propagation representation vector and the normalized filling material flow rate time series propagation representation vector, and then performing element - wise division by the number of eigenvalues of the normalized filling material flow rate time series propagation representation vector minus one to obtain the filling material flow rate time series sample covariance matrix.

[0018] Optionally, normalizing the pipeline pressure time series propagation representation vector and the filling material flow rate time series propagation representation vector to obtain a normalized pipeline pressure time series propagation representation vector and a normalized filling material flow rate time series propagation representation vector includes: calculating the mean and standard deviation of the pipeline pressure time series propagation representation vector to obtain a pipeline pressure time series mean and a pipeline pressure time series standard deviation; calculating the element-wise difference between the pipeline pressure time series propagation representation vector and the pipeline pressure time series mean and then dividing element-wise by the pipeline pressure time series standard deviation to obtain the normalized pipeline pressure time series propagation representation vector; calculating the mean and standard deviation of the filling material flow rate time series propagation representation vector to obtain a filling material flow rate time series mean and a filling material flow rate time series standard deviation; calculating the element-wise difference between the filling material flow rate time series propagation representation vector and the filling material flow rate time series mean and then dividing element-wise by the filling material flow rate time series standard deviation to obtain the normalized filling material flow rate time series propagation representation vector.

[0019] Optionally, inputting the set of pipeline pressure time series principal component feature vectors and the set of filling material flow rate time series principal component feature vectors into a maximum approximate query matching network to obtain a set of best matching pairs of pipeline pressure time series principal component feature vectors and filling material flow rate time series principal component feature vectors, includes: taking the pipeline pressure time series principal component feature vector at a predetermined position in the set of pipeline pressure time series principal component feature vectors to obtain a predetermined pipeline pressure time series principal component feature vector; calculating the inner product between the predetermined pipeline pressure time series principal component feature vector and each filling material flow rate time series principal component feature vector in the set of filling material flow rate time series principal component feature vectors to obtain a sequence of pipeline pressure - filling material flow rate inner product values; calculating the L1 norm of the predetermined pipeline pressure time series principal component feature vector to obtain a predetermined pipeline pressure time series principal component L1 norm; calculating the L1 norm of each filling material flow rate time series principal component feature vector in the set of filling material flow rate time series principal component feature vectors to obtain a sequence of filling material flow rate time series principal component L1 norms; multiplying the predetermined pipeline pressure time series principal component L1 norm and the sequence of filling material flow rate time series principal component L1 norms element-wise to obtain a sequence of pipeline pressure - filling material flow rate semantic interaction values; calculating the element-wise division between the sequence of pipeline pressure - filling material flow rate inner product values and the sequence of pipeline pressure - filling material flow rate semantic interaction values, and taking the filling material flow rate time series principal component feature vector corresponding to the maximum eigenvalue in the obtained sequence of eigenvalues and the predetermined pipeline pressure time series principal component feature vector to form a set of best matching pairs.

[0020] Optionally, input each best matching pair of the pipeline pressure time series principal component feature vectors and the filling material flow rate time series principal component feature vectors in the set of best matching pairs into the semantic fine-grained gating joint module to obtain a set of pipeline pressure-filling material flow rate time series principal component fusion feature vectors, including: performing position-based difference, position-based dot product, and position-based summation processing on each of the best matching pairs of the pipeline pressure time series principal component feature vectors and the filling material flow rate time series principal component feature vectors to obtain a set of pipeline pressure-filling material flow rate time series principal component difference feature vectors, a set of pipeline pressure-filling material flow rate time series principal component dot product feature vectors, and a set of pipeline pressure-filling material flow rate time series principal component summation feature vectors; performing position-based concatenation on the set of pipeline pressure-filling material flow rate time series principal component difference feature vectors, the set of pipeline pressure-filling material flow rate time series principal component dot product feature vectors, and the set of pipeline pressure-filling material flow rate time series principal component summation feature vectors to obtain a set of pipeline pressure-filling material flow rate time series principal component multi-dimensional fusion expression vectors; and processing each of the pipeline pressure-filling material flow rate time series principal component multi-dimensional fusion expression vectors in the set of pipeline pressure-filling material flow rate time series principal component multi-dimensional fusion expression vectors through a one-dimensional convolutional layer and then through a max pooling layer to obtain the set of pipeline pressure-filling material flow rate time series principal component fusion feature vectors.

[0021] In a second aspect, the present application provides a remote control system for an underground filling device, the system including:

[0022] A data acquisition module, configured to acquire a time queue of pipeline pressure and a time queue of filling material flow rate values collected by a pressure sensor;

[0023] A data transmission module, configured to transmit the time queue of the pipeline pressure and the time queue of the filling material flow rate values to a remote control server through a wireless communication module;

[0024] A local time series encoding module, configured to perform local time series encoding on the time queue of the pipeline pressure and the time queue of the filling material flow rate values respectively at the remote control server to obtain a sequence of pipeline pressure local time series correlation feature vectors and a sequence of filling material flow rate local time series correlation feature vectors;

[0025] A feature forward propagation module, configured to input the sequence of pipeline pressure local time series correlation feature vectors and the sequence of filling material flow rate local time series correlation feature vectors into a feature forward propagation network based on significant time series attenuation of node energy to obtain a pipeline pressure time series propagation representation vector and a filling material flow rate time series propagation representation vector;

[0026] A significant fusion module for inputting the pipeline pressure time-series propagation representation vector and the filling material flow rate time-series propagation representation vector into a query matching interaction network based on feature principal components to obtain a significantly fused representation vector of the core components of pipeline pressure - filling material flow rate;

[0027] A remote operation instruction determination module for inputting the significantly fused representation vector of the core components of pipeline pressure - filling material flow rate into a remote control instruction generator based on a decoder to obtain a remote operation instruction, where the remote operation instruction includes a recommended decoded value of the filling rate at the next time point.

[0028] The present application has at least the following technical effects:

[0029] Compared with the prior art, the present application real-time monitors and collects pipeline pressure data through a pressure sensor, and real-time monitors and collects the filling material flow rate value. An artificial intelligence and deep learning-based data processing and analysis algorithm is introduced into the remote control server to perform time-series analysis and information propagation aggregation representation on the pipeline pressure time-series data and the filling material flow rate time-series data. Then, the core component fusion semantics between the pipeline pressure time-series propagation aggregation feature and the filling material flow rate time-series propagation aggregation feature are used to recommend the filling rate at the next time point, and a variable frequency drive and a servo motor are used to achieve precise control of the filling and pumping system. Based on the time-series co-variation between the pipeline pressure and the filling material flow rate, the filling rate of the material is adaptively controlled, enabling more intelligent remote control of underground filling equipment, thereby avoiding problems such as unstable filling rate, excessive pressure, and uneven filling caused by the traditional manual operation method, and improving the safety and efficiency of the overall underground filling operation.

[0030] Other features and advantages of the present application will be described in detail in the subsequent specific implementation section. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Combined with the drawings and referring to the following specific implementation manners, the above and other features, advantages, and aspects of the embodiments of the present application will become more obvious. Throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic, and the original elements and elements are not necessarily drawn to scale. In the drawings:

[0032] Figure 1 is a flowchart of a method for remotely controlling an underground filling device shown according to an exemplary embodiment.

[0033] Figure 2 is a block diagram of a remote control system for an underground filling device shown according to an exemplary embodiment. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0034] Embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although some embodiments of the present application are shown in the drawings, it should be understood that the present application can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. On the contrary, these embodiments are provided to more thoroughly and completely understand the present application. It should be understood that the drawings and embodiments of the present application are only for exemplary purposes and are not used to limit the protection scope of the present application.

[0035] It should be understood that the various steps recited in the method embodiments of the present application can be executed in a different order and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present application is not limited in this regard.

[0036] As used herein, the term "including" and its variants are open-ended, that is, "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0037] It should be noted that the concepts such as "first" and "second" mentioned in the present application are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0038] It should be noted that the modifications of "one" and "multiple" mentioned in the present application are illustrative rather than restrictive. Those skilled in the art should understand that unless otherwise clearly specified in the context, it should be understood as "one or more".

[0039] The names of the messages or information exchanged between multiple devices in the embodiments of the present application are only for illustrative purposes and are not used to limit the scope of these messages or information.

[0040] The following will describe the specific embodiments of the present application in detail with reference to the accompanying drawings.

[0041] It should be noted that all the acquisition and processing of information or data in the present application are carried out on the premise of complying with the corresponding data protection regulations and policies of the country where the location is located and obtaining the authorization given by the corresponding device owner.

[0042] It should be understood that the control of the filling rate is an important part of the filling operation in underground mines. A reasonable filling rate can not only ensure the quality of the filling body, but also effectively improve the efficiency of the filling operation, while reducing resource waste and potential safety risks. Specifically, an appropriate filling rate helps to ensure the uniform distribution of the filling material, avoiding the occurrence of voids or over-dense areas, thereby improving the overall stability and bearing capacity of the filling body. In addition, reasonable control of the filling rate helps to prevent accidents, such as pipeline rupture caused by excessive pressure due to too fast filling rate.

[0043] Based on this, in the technical solution of this application, a remote control method for underground filling equipment is proposed. It can real-time monitor and collect pipeline pressure data through a pressure sensor, and real-time monitor and collect the flow rate value of the filling material. Then, the collected pipeline pressure time-series data and filling material flow rate time-series data are transmitted to a remote control server through a wireless communication module, so as to introduce data processing and analysis algorithms based on artificial intelligence and deep learning in the remote control server to perform time-series analysis and information propagation aggregation representation on the pipeline pressure time-series data and the filling material flow rate time-series data. Then, the core component fusion semantics between the pipeline pressure time-series propagation aggregation feature and the filling material flow rate time-series propagation aggregation feature are used to recommend the filling rate at the next time point, and a variable frequency drive and a servo motor are used to achieve precise control of the filling pumping system to reach the required recommended filling rate. It can use the wireless communication module and the remote control server to realize more intelligent remote control of underground filling equipment, so as to adaptively control the material filling rate based on the time-series collaborative change situation between the pipeline pressure and the filling material flow rate, thereby avoiding the problems of unstable filling rate, excessive pressure and uneven filling caused by the traditional manual operation method, and improving the safety and efficiency of the overall underground filling operation.

[0044] Figure 1 is a flowchart of a remote control method for underground filling equipment shown according to an exemplary embodiment, as Figure 1 shown, the method includes:

[0045] Step S101, obtain the time queue of the pipeline pressure collected by the pressure sensor and the time queue of the filling material flow rate value;

[0046] Step S102, transmit the time queue of the pipeline pressure and the time queue of the filling material flow rate value to the remote control server through the wireless communication module;

[0047] Step S103: At the remote control server, perform local temporal encoding on the time queue of the pipeline pressure and the time queue of the filling material flow rate respectively to obtain a sequence of local temporal correlation feature vectors of the pipeline pressure and a sequence of local temporal correlation feature vectors of the filling material flow rate;

[0048] Step S104: Input the sequence of local temporal correlation feature vectors of the pipeline pressure and the sequence of local temporal correlation feature vectors of the filling material flow rate into a feature forward propagation network based on significant temporal decay of node energy to obtain a temporal propagation representation vector of the pipeline pressure and a temporal propagation representation vector of the filling material flow rate;

[0049] Step S105: Input the temporal propagation representation vector of the pipeline pressure and the temporal propagation representation vector of the filling material flow rate into a query matching interaction network based on feature principal components to obtain a significantly fused representation vector of the core components of the pipeline pressure - filling material flow rate;

[0050] Step S106: Input the significantly fused representation vector of the core components of the pipeline pressure - filling material flow rate into a remote control instruction generator based on a decoder to obtain a remote operation instruction, where the remote operation instruction includes a recommended decoded value of the filling rate at the next time point.

[0051] Specifically, in the technical solution of this application, first, obtain the time queue of the pipeline pressure and the time queue of the filling material flow rate collected by the pressure sensor, and transmit the time queue of the pipeline pressure and the time queue of the filling material flow rate to the remote control server through a wireless communication module.

[0052] At the remote control server, input the time queue of the pipeline pressure and the time queue of the filling material flow rate into a sequence encoder based on a 1D - CNN model for local temporal encoding, so as to extract the local temporal correlation feature information of the pipeline pressure and the filling material flow rate in the time dimension respectively, thereby obtaining a sequence of local temporal correlation feature vectors of the pipeline pressure and a sequence of local temporal correlation feature vectors of the filling material flow rate.

[0053] In an embodiment of this application, performing local temporal encoding on the time queue of the pipeline pressure and the time queue of the filling material flow rate respectively at the remote control server to obtain a sequence of local temporal correlation feature vectors of the pipeline pressure and a sequence of local temporal correlation feature vectors of the filling material flow rate includes: at the remote control server, inputting the time queue of the pipeline pressure and the time queue of the filling material flow rate into a sequence encoder based on a 1D - CNN model to obtain the sequence of local temporal correlation feature vectors of the pipeline pressure and the sequence of local temporal correlation feature vectors of the filling material flow rate.

[0054] Then, in the sequence of the local temporal correlation feature vectors of the pipeline pressure, considering that each local temporal correlation feature vector of the pipeline pressure contains the local temporal feature information of the pipeline pressure within a local time period, there are correlation relationships and interaction effects based on the global time sequence among these local temporal semantics of the pipeline pressure, and this influence will decay continuously over time. Therefore, in order to more comprehensively and accurately represent the current pressure state of the pipeline and provide a basis for subsequent speculation of the filling rate, it is necessary to utilize the correlation features and decay importance effects among the local temporal semantics of each pipeline pressure. Based on this, in the technical solution of this application, the sequence of the local temporal correlation feature vectors of the pipeline pressure is further input into a feature forward propagation network based on significant temporal decay of node energy to obtain a pipeline pressure temporal propagation representation vector. Specifically, the feature forward propagation network based on significant temporal decay of node energy can focus on aggregating the local temporal features of each pipeline pressure through the significance and temporal decay characteristics of node energy to generate a propagation aggregation representation vector. Specifically, in the feature forward propagation network based on significant temporal decay of node energy, the evaluation of node energy significance can help the network determine which pipeline pressure features within which local time periods are more important. For the filling process, the pipeline pressure will change continuously over time, and the local temporal semantics of the pipeline pressure closer to the current time period will be more important for the temporal information aggregation representation of the pipeline pressure and the subsequent pressure speculation. Therefore, through the evaluation of energy significance and the consideration of temporal decay, this network can finely adjust the feature weights of each local temporal correlation feature vector of the pipeline pressure, strengthen the influence of the local temporal features of the pipeline pressure at key nodes, and appropriately consider the time factor. This network can learn the variation law of the pipeline pressure during the filling process faster, helps to strengthen the semantic expression of key nodes during the aggregation of the local temporal semantics of the pipeline pressure, and has a significant effect in improving the accuracy of the model's prediction of time series, that is, it can predict the subsequent pipeline pressure change, thereby helping to speculate and control the filling rate at the next time point. Correspondingly, the process of processing the sequence of the local temporal correlation feature vectors of the filling material flow rate to obtain a filling material flow rate temporal propagation representation vector is the same as the above method.

[0055] In one embodiment of the present application, inputting the sequence of the local time-series correlation feature vectors of the pipeline pressure and the sequence of the local time-series correlation feature vectors of the filling material flow rate into a feature forward propagation network based on significant time-series decay of node energy to obtain a pipeline pressure time-series propagation representation vector and a filling material flow rate time-series propagation representation vector, including: calculating an energy significant descriptor for each local time-series correlation feature vector of the pipeline pressure in the sequence of the local time-series correlation feature vectors of the pipeline pressure based on the mean and standard deviation of each local time-series correlation feature vector of the pipeline pressure to obtain a sequence of local time-series energy significant descriptors of the pipeline pressure; extracting the timestamps of each local time-series correlation feature vector of the pipeline pressure in the sequence of the local time-series correlation feature vectors of the pipeline pressure to obtain a sequence of pipeline pressure timestamps; taking the local time-series correlation feature vector of the pipeline pressure corresponding to the current time point in the sequence of the local time-series correlation feature vectors of the pipeline pressure as the current local time-series correlation feature vector of the pipeline pressure and taking the local time-series correlation feature vectors of the pipeline pressure corresponding to other time points as the historical local time-series correlation feature vectors of the pipeline pressure to obtain a sequence of the current local time-series correlation feature vector of the pipeline pressure and the historical local time-series correlation feature vectors of the pipeline pressure; calculating an energy time decay factor of each historical local time-series correlation feature vector of the pipeline pressure in the sequence of the historical local time-series correlation feature vectors of the pipeline pressure relative to the current local time-series correlation feature vector of the pipeline pressure to obtain a sequence of local time-series energy time decay factors of the pipeline pressure; calculating the local time-series energy weight obtained after the position-wise division of the sequence of the local time-series energy significant descriptors of the pipeline pressure and the sequence of the local time-series energy time decay factors of the pipeline pressure as a weighting coefficient, and calculating the position-wise weighted sum of the sequence of the historical local time-series correlation feature vectors of the pipeline pressure to obtain a local time-series historical node propagation aggregation representation feature vector of the pipeline pressure; calculating the position-wise weighted sum of the local time-series historical node propagation aggregation representation feature vector of the pipeline pressure and the current local time-series correlation feature vector of the pipeline pressure based on the energy significant descriptor corresponding to the current local time-series correlation feature vector of the pipeline pressure to obtain the pipeline pressure time-series propagation aggregation representation vector.

[0056] Further, in an embodiment of the present application, based on the mean and standard deviation of each pipeline pressure local temporal correlation feature vector in the sequence of pipeline pressure local temporal correlation feature vectors, calculate the energy significant descriptor of each pipeline pressure local temporal correlation feature vector to obtain a sequence of pipeline pressure local temporal energy significant descriptors, including: calculating the mean and standard deviation of each pipeline pressure local temporal correlation feature vector in the sequence of pipeline pressure local temporal correlation feature vectors to obtain a sequence of pipeline pressure local means and a sequence of pipeline pressure local standard deviations; after performing position-wise difference on each pair of corresponding pipeline pressure local temporal correlation feature vector and pipeline pressure local mean in the sequence of pipeline pressure local temporal correlation feature vectors and the sequence of pipeline pressure local means, then performing fourth-power modulation on each position eigenvalue in the obtained sequence of difference feature vectors to obtain a sequence of pipeline pressure local temporal modulation feature vectors; calculating the expected value of each pipeline pressure local temporal modulation feature vector in the sequence of pipeline pressure local temporal modulation feature vectors to obtain a sequence of pipeline pressure local temporal positive expectation factors; calculating the fourth-power modulation value of each pipeline pressure local standard deviation in the sequence of pipeline pressure local standard deviations to obtain a sequence of pipeline pressure local temporal negative expectation factors; dividing each pair of corresponding pipeline pressure local temporal positive expectation factor and pipeline pressure local temporal negative expectation factor in the sequence of pipeline pressure local temporal positive expectation factors and the sequence of pipeline pressure local temporal negative expectation factors to obtain the sequence of pipeline pressure local temporal energy significant descriptors.

[0057] Furthermore, in an embodiment of the present application, calculate the energy time decay factor of each historical pipeline pressure local temporal correlation feature vector in the sequence of historical pipeline pressure local temporal correlation feature vectors relative to the current pipeline pressure local temporal correlation feature vector to obtain a sequence of pipeline pressure local temporal energy time decay factors, including: respectively calculating the pipeline pressure timestamp corresponding to the current pipeline pressure local temporal correlation feature vector minus the pipeline pressure timestamp corresponding to each historical pipeline pressure local temporal correlation feature vector and then taking the floor to obtain a sequence of time span values; after calculating the square of each time span value in the sequence of time span values, performing position-wise division of the obtained sequence of span square modulation values by the square of the time decay inverse scaling parameter and performing position-wise multiplication by the decay rate positive scaling parameter to obtain a sequence of temporal span energy decay coefficients; using each temporal span energy decay coefficient in the sequence of temporal span energy decay coefficients as the exponential power, calculating the natural exponential function value with the natural constant e as the base to obtain the sequence of pipeline pressure local temporal energy time decay factors.

[0058] Specifically, input the sequence of the local time-series correlation feature vectors of the pipeline pressure into the feature forward propagation network based on the significant time-series decay of node energy and process it with the following node feature propagation formula to obtain the time-series propagation representation vector of the pipeline pressure;

[0059] Among them, the node feature propagation formula is:

[0060]

[0061]

[0062]

[0063]

[0064]

[0065] Among them, is the sequence of the local time-series correlation feature vectors of the pipeline pressure, and are respectively the 1st, 2nd,... ([[]] )th and the th local time-series correlation feature vectors of the pipeline pressure in the sequence of the local time-series correlation feature vectors of the pipeline pressure, is the th local time-series correlation feature vector of the pipeline pressure in the sequence of the local time-series correlation feature vectors of the pipeline pressure, is the th position feature value in the th local time-series correlation feature vector of the pipeline pressure, is the length of the th local time-series correlation feature vector of the pipeline pressure, is the mean value of the th local time-series correlation feature vector of the pipeline pressure, is the standard deviation of the th local time-series correlation feature vector of the pipeline pressure, represents the calculation of the expected value, is the energy significant descriptor of the th local time-series correlation feature vector of the pipeline pressure, represents the value of the exponential function with the natural constant e as the base, is the positive scaling parameter for controlling the decay rate, is the inverse scaling parameter for controlling the time decay period, and respectively represent the timestamps of the and th local time-series correlation feature vectors of the pipeline pressure, Denotes the floor operation, is the energy significant descriptor of the -th local temporal correlation feature vector of the pipeline pressure, and is a trainable weight hyperparameter, is the temporal propagation representation vector of the pipeline pressure.

[0066] It should be understood that the temporal propagation representation vector of the pipeline pressure and the temporal propagation representation vector of the filling material flow respectively contain the full-time domain propagation aggregation representation information of the local temporal features of the pipeline pressure and the local temporal features of the filling material flow. Such information also respectively contains the temporal decay information and inference information of the local temporal features of the pipeline pressure and the filling material flow in time series, which is beneficial to inferring the filling rate at the next time point. Moreover, due to the implicit correlation and mutual influence between the pipeline pressure and the filling material flow, this interaction information is of great significance for the control of the filling rate. For example, problems such as pipeline rupture caused by excessive pipeline pressure due to too fast filling rate. Based on this, in the technical solution of this application, the temporal propagation representation vector of the pipeline pressure and the temporal propagation representation vector of the filling material flow are further input into a query matching interaction network based on feature principal components to obtain a significantly fused representation vector of the pipeline pressure - filling material flow core components. The query matching interaction network based on feature principal components is a feature extraction and fusion technology based on feature sparsification processing, principal component analysis, and significant fusion, which is used to generate a sparse and significant feature correlation representation between feature vectors.

[0067] In one embodiment of the present application, inputting the pipeline pressure time series propagation representation vector and the filling material flow rate time series propagation representation vector into a query matching interaction network based on feature principal components to obtain a significantly fused representation vector of the pipeline pressure - filling material flow rate core components, including: performing normalization processing on the pipeline pressure time series propagation representation vector and the filling material flow rate time series propagation representation vector to obtain a normalized pipeline pressure time series propagation representation vector and a normalized filling material flow rate time series propagation representation vector; calculating the sample covariance matrices of the normalized pipeline pressure time series propagation representation vector and the normalized filling material flow rate time series propagation representation vector to obtain a pipeline pressure time series sample covariance matrix and a filling material flow rate time series sample covariance matrix; performing eigenvector extraction based on matrix decomposition on the pipeline pressure time series sample covariance matrix and the filling material flow rate time series sample covariance matrix to obtain a set of pipeline pressure time series principal component eigenvectors and a set of filling material flow rate time series principal component eigenvectors; inputting the set of pipeline pressure time series principal component eigenvectors and the set of filling material flow rate time series principal component eigenvectors into a maximum approximate query matching network to obtain a set of best matching pairs of the pipeline pressure time series principal component eigenvectors and the filling material flow rate time series principal component eigenvectors; inputting each best matching pair of the pipeline pressure time series principal component eigenvectors and the filling material flow rate time series principal component eigenvectors in the set of best matching pairs into a semantic fine - grained gating joint module to obtain a set of pipeline pressure - filling material flow rate time series principal component fusion feature vectors; concatenating each pipeline pressure - filling material flow rate time series principal component fusion feature vector in the set of pipeline pressure - filling material flow rate time series principal component fusion feature vectors to obtain the significantly fused representation vector of the pipeline pressure - filling material flow rate core components; wherein, calculating the sample covariance matrices of the normalized pipeline pressure time series propagation representation vector and the normalized filling material flow rate time series propagation representation vector to obtain a pipeline pressure time series sample covariance matrix and a filling material flow rate time series sample covariance matrix includes: calculating the vector multiplication between the transposed vector of the normalized pipeline pressure time series propagation representation vector and the normalized pipeline pressure time series propagation representation vector, and then performing element - by - element division by the number of eigenvalues of the normalized pipeline pressure time series propagation representation vector minus one to obtain the pipeline pressure time series sample covariance matrix; calculating the vector multiplication between the transposed vector of the normalized filling material flow rate time series propagation representation vector and the normalized filling material flow rate time series propagation representation vector, and then performing element - by - element division by the number of eigenvalues of the normalized filling material flow rate time series propagation representation vector minus one to obtain the filling material flow rate time series sample covariance matrix.

[0068] Specifically, first, the time-series propagation representation vectors of the pipeline pressure and the filling material flow rate are normalized to ensure that the time-series propagation characteristics of the pipeline pressure and the filling material flow rate are distributed on a scale with zero mean and unit variance. This can eliminate the influence of different dimensions and magnitudes between the two, enabling analysis and comparison between the time-series propagation representation vector of the pipeline pressure and the time-series propagation representation vector of the filling material flow rate.

[0069] In an embodiment of the present application, normalizing the time-series propagation representation vector of the pipeline pressure and the time-series propagation representation vector of the filling material flow rate to obtain a normalized time-series propagation representation vector of the pipeline pressure and a normalized time-series propagation representation vector of the filling material flow rate includes: calculating the mean and standard deviation of the time-series propagation representation vector of the pipeline pressure to obtain the pipeline pressure time-series mean and the pipeline pressure time-series standard deviation; calculating the position-wise difference between the time-series propagation representation vector of the pipeline pressure and the pipeline pressure time-series mean and then dividing by the pipeline pressure time-series standard deviation position-wise to obtain the normalized time-series propagation representation vector of the pipeline pressure; calculating the mean and standard deviation of the time-series propagation representation vector of the filling material flow rate to obtain the filling material flow rate time-series mean and the filling material flow rate time-series standard deviation; calculating the position-wise difference between the time-series propagation representation vector of the filling material flow rate and the filling material flow rate time-series mean and then dividing by the filling material flow rate time-series standard deviation position-wise to obtain the normalized time-series propagation representation vector of the filling material flow rate.

[0070] Next, calculate the covariance matrix of the normalized time-series propagation representation vector of the pipeline pressure and the normalized time-series propagation representation vector of the filling material flow rate. The covariance matrix reflects the linear correlation between the time-series propagation characteristics of the pipeline pressure and the time-series propagation characteristics of the filling material flow rate. This step helps to understand how the characteristics in the time-series propagation representation vectors of the pipeline pressure and the filling material flow rate depend on each other.

[0071] Then, use the principal component analysis algorithm to perform feature sparsification processing on the normalized time-series propagation representation vector of the pipeline pressure and the normalized time-series propagation representation vector of the filling material flow rate. Principal component analysis can identify the most important change directions in the time-series characteristics of the pipeline pressure and the time-series characteristics of the filling material flow rate and represent them as principal components. By principal component analysis, the dimension of the data can be reduced while retaining the key information in the data feature set, which helps to extract the time-series core characteristics of the pipeline pressure and the filling material flow rate from the normalized time-series propagation representation vector of the pipeline pressure and the normalized time-series propagation representation vector of the filling material flow rate.

[0072] After extracting the principal component feature vectors, it is next necessary to determine which feature vectors are the most similar to each other, that is, they have similar principal component features. Furthermore, based on the maximum approximate query matching between the set of principal component feature vectors of the pipeline pressure time series and the set of principal component feature vectors of the packing material flow rate time series, by calculating and comparing the similarity between the sets of principal component feature vectors of the pipeline pressure and the packing material flow rate, the best matching pairs are found, which involves a distance metric or a similarity scoring mechanism. The maximum approximate query matching network helps to understand the internal relationship between the pipeline pressure and the packing material flow rate by identifying the most similar pairs of principal component feature vectors, and provides an accurate correspondence for the fusion of these two features.

[0073] In an embodiment of the present application, the set of principal component feature vectors of the pipeline pressure time series and the set of principal component feature vectors of the packing material flow rate time series are input into the maximum approximate query matching network to obtain a set of best matching pairs of the principal component feature vectors of the pipeline pressure time series and the principal component feature vectors of the packing material flow rate time series, including: taking the principal component feature vector of the pipeline pressure time series at a predetermined position in the set of principal component feature vectors of the pipeline pressure time series to obtain a predetermined principal component feature vector of the pipeline pressure time series; calculating the inner product between the predetermined principal component feature vector of the pipeline pressure time series and each principal component feature vector of the packing material flow rate time series in the set of principal component feature vectors of the packing material flow rate time series to obtain a sequence of pipeline pressure-packing material flow rate inner product values; calculating the first norm of the predetermined principal component feature vector of the pipeline pressure time series to obtain a first norm of the predetermined principal component feature of the pipeline pressure time series; calculating the first norm of each principal component feature vector of the packing material flow rate time series in the set of principal component feature vectors of the packing material flow rate time series to obtain a sequence of first norms of the principal component features of the packing material flow rate time series; multiplying the first norm of the predetermined principal component feature of the pipeline pressure time series and the sequence of first norms of the principal component features of the packing material flow rate time series at the position points to obtain a sequence of pipeline pressure-packing material flow rate semantic interaction values; after calculating the point-by-point division between the sequence of pipeline pressure-packing material flow rate inner product values and the sequence of pipeline pressure-packing material flow rate semantic interaction values, taking the principal component feature vector of the packing material flow rate time series corresponding to the maximum eigenvalue in the obtained sequence of eigenvalue and the predetermined principal component feature vector of the pipeline pressure time series to form a group of best matching pairs.

[0074] Next, a gating mechanism is used to perform fine-grained semantic combination on the best matching pairs of the pipeline pressure time series principal component feature vectors and the filling material flow rate time series principal component feature vectors. The gating mechanism can dynamically adjust the weights of the features according to the importance and context information of the pipeline pressure time series principal components and the filling material flow rate time series principal components. Through gating combination, not only the information with high significance in each best matching pair of the pipeline pressure time series principal component feature vectors and the filling material flow rate time series principal component feature vectors is enhanced, but also through fine-grained control, the model's ability to identify the importance of features is improved, providing rich semantic information for generating the final pipeline pressure - filling material flow rate time series principal component fusion feature representation.

[0075] Finally, the set of pipeline pressure - filling material flow rate time series principal component fusion feature vectors obtained through gating combination is cascaded and fused to generate the significant fusion representation vector of the pipeline pressure - filling material flow rate core components. This vector contains the key information of the pipeline pressure and the filling material flow rate and their interrelationships, and this representation helps to improve the prediction accuracy of the model for the filling rate, especially when dealing with complex underground mine filling operation tasks.

[0076] In one embodiment of the present application, inputting each best matching pair of the pipeline pressure time series principal component feature vectors and the filling material flow rate time series principal component feature vectors in the set of best matching pairs into a semantic fine-grained gating combination module to obtain a set of pipeline pressure - filling material flow rate time series principal component fusion feature vectors includes: respectively performing position-based difference, position-based dot product, and position-based summation processing on each best matching pair of the pipeline pressure time series principal component feature vectors and the filling material flow rate time series principal component feature vectors to obtain a set of pipeline pressure - filling material flow rate time series principal component difference feature vectors, a set of pipeline pressure - filling material flow rate time series principal component dot product feature vectors, and a set of pipeline pressure - filling material flow rate time series principal component summation feature vectors; performing position-based concatenation on the set of pipeline pressure - filling material flow rate time series principal component difference feature vectors, the set of pipeline pressure - filling material flow rate time series principal component dot product feature vectors, and the set of pipeline pressure - filling material flow rate time series principal component summation feature vectors to obtain a set of pipeline pressure - filling material flow rate time series principal component multi-dimensional fusion expression vectors; respectively processing each pipeline pressure - filling material flow rate time series principal component multi-dimensional fusion expression vector in the set of pipeline pressure - filling material flow rate time series principal component multi-dimensional fusion expression vectors through a one-dimensional convolutional layer and then through a maximum pooling layer to obtain the set of pipeline pressure - filling material flow rate time series principal component fusion feature vectors.

[0077] Specifically, input the pipeline pressure time series propagation representation vector and the filling material flow rate time series propagation representation vector into the query matching interaction network based on feature principal components and process them according to the following feature principal component query matching formula to obtain the pipeline pressure - filling material flow rate core component significantly fused representation vector;

[0078] Among them, the feature principal component query matching formula is:

[0079]

[0080]

[0081]

[0082]

[0083]

[0084]

[0085]

[0086]

[0087]

[0088]

[0089]

[0090]

[0091]

[0092] Among them, is the pipeline pressure time series propagation representation vector, is the filling material flow rate time series propagation representation vector, is the mean value of the pipeline pressure time series propagation representation vector, is the standard deviation of the pipeline pressure time series propagation representation vector, is the mean value of the filling material flow rate time series propagation representation vector, is the standard deviation of the filling material flow rate time series propagation representation vector, represents differential by position, is the standardized pipeline pressure time series propagation representation vector, is the standardized filling material flow rate time series propagation representation vector, is the number of eigenvalues in the standardized pipeline pressure time series propagation representation vector, is the number of eigenvalues in the eigenvector time - series propagation representation of the standardized filling material flow is the transpose operation is the sample covariance matrix of the pipeline pressure time series is the sample covariance matrix of the filling material flow time series is each pipeline pressure time - series principal component eigenvector in the set of pipeline pressure time - series principal component eigenvectors is the pipeline pressure time - series principal component orthogonal matrix represents that the elements on the diagonal of the matrix are diagonal matrix respectively represent the weight parameters of the respective pipeline pressure time - series principal component eigenvectors is the pipeline pressure time - series diagonal matrix is each filling material flow time - series principal component eigenvector in the set of filling material flow time - series principal component eigenvectors is the filling material flow time - series principal component orthogonal matrix represents that the elements on the diagonal of the matrix are diagonal matrix is the filling material flow time - series diagonal matrix respectively represent the weight parameters of the respective filling material flow time - series principal component eigenvectors is the th pipeline pressure time - series principal component eigenvector in the set of pipeline pressure time - series principal component eigenvectors and are respectively the th and the th filling material flow time - series principal component eigenvectors in the set of filling material flow time - series principal component eigenvectors represents the inner product of vectors represents the one - norm of a vector represents returning the value corresponding to the maximum point is the maximum approximate query match value and respectively represent element - wise multiplication and element - wise addition represents a one - dimensional convolutional layer represents max - pooling is the th pipeline pressure - filling material flow time - series principal component fusion eigenvector in the set of pipeline pressure - filling material flow time - series principal component fusion eigenvectors is the number of eigenvectors in the set of pipeline pressure - filling material flow time - series principal component fusion eigenvectors It is the significant fusion representation vector of the core components of the pipeline pressure - filling material flow rate.

[0093] Subsequently, the significant fusion representation vector of the core components of the pipeline pressure - filling material flow rate is input into a decoder - based remote control instruction generator to obtain a remote operation instruction, and the remote operation instruction includes a recommended decoded value of the filling rate at the next time point. That is to say, the core component significance fusion characterization information between the pipeline pressure time - series propagation aggregation feature and the filling material flow rate time - series propagation aggregation feature is used for decoding regression, so as to recommend the filling rate at the next time point, and a variable - frequency drive and a servo motor are used to achieve precise control of the filling and pumping system to reach the required recommended filling rate. In this way, the wireless communication module and the remote control server can be used to achieve more intelligent remote control of underground filling equipment, adaptively control the material filling rate based on the time - series co - variation between the pipeline pressure and the filling material flow rate, thus avoiding the problems of unstable filling rate, excessive pressure, and uneven filling caused by the traditional manual operation method, and improving the safety and efficiency of the overall underground filling operation.

[0094] In a preferred example, considering that the pipeline pressure time - series propagation representation vector and the filling material flow rate time - series propagation representation vector respectively represent the significant attenuation propagation features based on the time - series node energy of the one - dimensional local time - series correlation features of the pipeline pressure and the filling material flow rate values. Therefore, when performing significant fusion based on feature principal component query matching, due to the difference in time - series correlation feature distribution caused by the source time - series distribution difference, superimposed with the significant propagation difference across time - series nodes, the significant fusion representation vector of the core components of the pipeline pressure - filling material flow rate will also have differences in the time - series feature - diverse spatial structure, affecting the convergence consistency of the decoder. Therefore, it is expected that the regression mapping consistency of the significant fusion representation vector of the core components of the pipeline pressure - filling material flow rate when input into the decoder - based remote control instruction generator for decoding regression, so as to improve the accuracy of the obtained remote operation instruction. Therefore, in one example, the significant fusion representation vector of the core components of the pipeline pressure - filling material flow rate is optimized.

[0095] Based on this, in the preferred example, inputting the significant fusion representation vector of the core components of the pipeline pressure - filling material flow rate into an optimal bit - weight recommender based on a decoder to obtain an optimal bit - weight recommended decoded value specifically includes:

[0096] Calculating the square root of the sum of the absolute values and the sum of the squares of all eigenvalues of the significant fusion representation vector of the core components of the pipeline pressure - filling material flow rate to obtain a first significant fusion spatial structure value of the core components of the pipeline pressure - filling material flow rate and a second significant fusion spatial structure value of the core components of the pipeline pressure - filling material flow rate, that is:

[0097]

[0098]

[0099] Among them, represents the th eigenvalue of the significant fusion representation vector of the pipeline pressure - filling material flow core component, represents the spatial structure value of the significant fusion of the first pipeline pressure - filling material flow core component, represents the spatial structure value of the significant fusion of the second pipeline pressure - filling material flow core component;

[0100] Determine the number of eigenvalues of all eigenvalues of the significant fusion representation vector of the pipeline pressure - filling material flow core component, that is, the length of the significant fusion representation vector of the pipeline pressure - filling material flow core component;

[0101] For each eigenvalue of the significant fusion representation vector of the pipeline pressure - filling material flow core component, calculate the first long - range dependence value of the significant fusion of the pipeline pressure - filling material flow core component obtained by subtracting the product of the eigenvalue and the number of eigenvalues from the spatial structure value of the significant fusion of the first pipeline pressure - filling material flow core component where, represents the spatial structure value of the significant fusion of the first pipeline pressure - filling material flow core component, represents the th eigenvalue of the significant fusion representation vector of the pipeline pressure - filling material flow core component, represents the number of eigenvalues of all eigenvalues of the significant fusion representation vector of the pipeline pressure - filling material flow core component, represents the first long - range dependence value of the significant fusion of the pipeline pressure - filling material flow core component;

[0102] Calculate the second long - range dependence value of the significant fusion of the pipeline pressure - filling material flow core component obtained by subtracting the spatial structure value of the significant fusion of the second pipeline pressure - filling material flow core component from the product of the square root of the number of eigenvalues and the eigenvalue where, represents the spatial structure value of the significant fusion of the second pipeline pressure - filling material flow core component, represents the th eigenvalue of the significant fusion representation vector of the pipeline pressure - filling material flow core component, represents the number of eigenvalues of all eigenvalues of the significant fusion representation vector of the pipeline pressure - filling material flow core component, Represents the significant fusion long-range dependence value of the second pipeline pressure - filling material flow core component;

[0103] Perform a weighted sum of the exponential value obtained by calculating the significant fusion long-range dependence value of the first pipeline pressure - filling material flow core component as the exponent of the natural constant and the reciprocal of the significant fusion long-range dependence value of the second pipeline pressure - filling material flow core component to obtain the optimized eigenvalue corresponding to each eigenvalue , where, Represents the significant fusion long-range dependence value of the first pipeline pressure - filling material flow core component, Represents the significant fusion long-range dependence value of the second pipeline pressure - filling material flow core component, Represents the natural constant, and Represents the weighted hyperparameter, Represents the optimized eigenvalue corresponding to each eigenvalue;

[0104] Form an optimized pipeline pressure - filling material flow core component significant fusion representation vector with the optimized eigenvalues, and input the optimized pipeline pressure - filling material flow core component significant fusion representation vector into a decoder-based remote control instruction generator to obtain a remote operation instruction.

[0105] That is, due to the possible lack of spatial structure in the feature set of the pipeline pressure - filling material flow core component significant fusion representation vector in the high-dimensional space, the weights of the decoder implicitly infer spatial structure information based on features, resulting in inconsistent convergence. By establishing long-distance feature dependencies based on the overall feature scale of the pipeline pressure - filling material flow core component significant fusion representation vector relative to the spatial structure representation of the pipeline pressure - filling material flow core component significant fusion representation vector, to establish the feature local connectivity of the pipeline pressure - filling material flow core component significant fusion representation vector, and by capturing the spatial ambiguity information of the object eigenvalue through the unstructured eigenvalue point prediction of the pipeline pressure - filling material flow core component significant fusion representation vector, thereby enhancing the spatial inductive bias perception ability of the feature set of the pipeline pressure - filling material flow core component significant fusion representation vector, improving the convergence consistency of the decoder, and enhancing the accuracy of the remote operation instruction obtained by inputting the pipeline pressure - filling material flow core component significant fusion representation vector into the decoder-based remote control instruction generator. In this way, it is possible to more accurately adaptively control the material filling rate based on the temporal collaborative change situation between the pipeline pressure and the filling material flow, thereby avoiding the problems of excessive pressure and uneven filling caused by unstable filling rate in the traditional manual operation method, and improving the safety and efficiency of the overall underground filling operation.

[0106] In summary, by adopting the above solution, the pipeline pressure data is collected through real-time monitoring by a pressure sensor, and the flow rate value of the filling material is collected through real-time monitoring. Then, the collected pipeline pressure time-series data and filling material flow rate time-series data are transmitted to a remote control server through a wireless communication module, so as to introduce data processing and analysis algorithms based on artificial intelligence and deep learning in the remote control server to perform time-series analysis and information propagation aggregation representation on the pipeline pressure time-series data and the filling material flow rate time-series data. Furthermore, the core component fusion semantics between the pipeline pressure time-series propagation aggregation feature and the filling material flow rate time-series propagation aggregation feature are used to recommend the filling rate at the next time point, and a variable frequency drive and a servo motor are used to achieve precise control of the filling and pumping system, so as to reach the required recommended filling rate. In this way, the wireless communication module and the remote control server can be used to realize more intelligent remote control of underground filling equipment, and adaptively control the material filling rate based on the time-series coordinated change situation between the pipeline pressure and the filling material flow rate, thereby avoiding the problems of excessive pressure and uneven filling caused by unstable filling rate brought by the traditional manual operation method, and improving the safety and efficiency of the overall underground filling operation.

[0107] Figure 2 is a block diagram of a remote control system for an underground filling device shown according to an exemplary embodiment. As Figure 2 shown, the system 200 includes:

[0108] A data acquisition module 201, configured to acquire a time queue of pipeline pressure collected by a pressure sensor and a time queue of filling material flow rate values;

[0109] A data transmission module 202, configured to transmit the time queue of the pipeline pressure and the time queue of the filling material flow rate values to a remote control server through a wireless communication module;

[0110] A local time-series encoding module 203, configured to perform local time-series encoding on the time queue of the pipeline pressure and the time queue of the filling material flow rate values respectively in the remote control server to obtain a sequence of pipeline pressure local time-series correlation feature vectors and a sequence of filling material flow rate local time-series correlation feature vectors;

[0111] A feature forward propagation module 204, configured to input the sequence of pipeline pressure local time-series correlation feature vectors and the sequence of filling material flow rate local time-series correlation feature vectors into a feature forward propagation network based on significant time-series attenuation of node energy to obtain a pipeline pressure time-series propagation representation vector and a filling material flow rate time-series propagation representation vector;

[0112] A significant fusion module 205, configured to input the pipeline pressure time series propagation representation vector and the filling material flow rate time series propagation representation vector into a query matching interaction network based on feature principal components to obtain a significantly fused representation vector of the pipeline pressure - filling material flow rate core components;

[0113] A remote operation instruction determination module 206, configured to input the significantly fused representation vector of the pipeline pressure - filling material flow rate core components into a remote control instruction generator based on a decoder to obtain a remote operation instruction, where the remote operation instruction includes a recommended decoded value of the filling rate at the next time point.

[0114] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code that contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, may be implemented by a dedicated hardware - based system for performing the specified functions or operations, or may be implemented by a combination of dedicated hardware and computer instructions.

[0115] The above description is only a preferred embodiment of the present application and an explanation of the applied technical principles. Those skilled in the art should understand that the scope of disclosure involved in the present application is not limited to the technical solutions formed by the specific combination of the above - mentioned technical features, and should also cover other technical solutions formed by any combination of the above - mentioned technical features or their equivalent features without departing from the above - mentioned disclosure concept. For example, technical solutions formed by mutually replacing the above - mentioned features with technical features having similar functions (but not limited to) disclosed in the present application.

[0116] In addition, although the operations are depicted in a specific order, this should not be construed as requiring that the operations be performed in the specific order shown or in a sequential order. In certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of the present application. Certain features described in the context of separate embodiments may also be implemented combinatorially in a single embodiment. Conversely, the various features described in the context of a single embodiment may also be implemented separately or in any suitable sub - combination in multiple embodiments.

[0117] Although the subject matter has been described in language specific to structural features and / or methodological logical acts, it should be understood that the claimed subject matter is not necessarily limited to the specific features or acts described above. On the contrary, the specific features and acts described above are merely example forms of implementation. Regarding the devices in the above embodiments, the specific manner in which each module performs operations has been described in detail in the embodiments related to the method, and will not be elaborated here.

Claims

1. A remote control method for underground filling equipment, characterized in that: include: Obtaining a time queue of pipeline pressure and a time queue of filling material flow values ​​collected by a pressure sensor; Transmitting the time queue of the pipeline pressure and the time queue of the filling material flow value to the remote control server through the wireless communication module; On the remote control server, locally time-series encoding is performed on the time queue of the pipeline pressure and the time queue of the filling material flow value to obtain a sequence of local time-series associated feature vectors of the pipeline pressure and a sequence of local time-series associated feature vectors of the filling material flow; Inputting the sequence of the pipeline pressure local time series associated feature vectors and the sequence of the filling material flow local time series associated feature vectors into a feature forward propagation network based on significant time series attenuation of node energy to obtain a pipeline pressure time series propagation representation vector and a filling material flow time series propagation representation vector; Inputting the pipeline pressure time series propagation representation vector and the filling material flow rate time series propagation representation vector into a query matching interaction network based on characteristic principal components to obtain a pipeline pressure-filling material flow rate core component significant fusion representation vector; The pipeline pressure-filling material flow rate core component significant fusion representation vector is input into a decoder-based remote control instruction generator to obtain a remote operation instruction, and the remote operation instruction includes a recommended decoded value of the filling rate at the next time point.

2. The remote control method of underground filling equipment according to claim 1, characterized in that: On the remote control server, local time series encoding is performed on the time queue of the pipeline pressure and the time queue of the filling material flow value respectively to obtain a sequence of local time series correlation feature vectors of the pipeline pressure and a sequence of local time series correlation feature vectors of the filling material flow, including: on the remote control server, the time queue of the pipeline pressure and the time queue of the filling material flow value are input into a sequence encoder based on a 1D-CNN model to obtain a sequence of local time series correlation feature vectors of the pipeline pressure and a sequence of local time series correlation feature vectors of the filling material flow.

3. The remote control method of underground filling equipment according to claim 2, characterized in that: Inputting the sequence of the pipeline pressure local time series associated feature vectors and the sequence of the filling material flow local time series associated feature vectors into a feature forward propagation network based on significant time series attenuation of node energy to obtain a pipeline pressure time series propagation representation vector and a filling material flow time series propagation representation vector, including: Based on the mean and standard deviation of each pipeline pressure local time series associated feature vector in the sequence of pipeline pressure local time series associated feature vectors, calculating the energy significant descriptor of each pipeline pressure local time series associated feature vector to obtain a sequence of pipeline pressure local time series energy significant descriptors; Extracting the timestamp of each pipeline pressure local time series correlation feature vector in the sequence of pipeline pressure local time series correlation feature vectors to obtain a sequence of pipeline pressure timestamps; The pipeline pressure local time series correlation feature vector corresponding to the current time point in the sequence of the pipeline pressure local time series correlation feature vectors is used as the current pipeline pressure local time series correlation feature vector and the pipeline pressure local time series correlation feature vectors corresponding to other time points are used as the historical pipeline pressure local time series correlation feature vectors to obtain a sequence of the current pipeline pressure local time series correlation feature vector and the historical pipeline pressure local time series correlation feature vector; Calculating the energy time decay factor of each historical pipeline pressure local time series associated characteristic vector in the sequence of historical pipeline pressure local time series associated characteristic vectors relative to the current pipeline pressure local time series associated characteristic vector to obtain a sequence of pipeline pressure local time series energy time decay factors; After calculating the positional division of the sequence of the pipeline pressure local time series energy significant descriptors and the sequence of the pipeline pressure local time series energy time decay factors, the obtained pipeline pressure local time series energy weight is used as a weighting coefficient, and the positional weighted sum of the sequence of the historical pipeline pressure local time series associated feature vectors is calculated to obtain the pipeline pressure local time series historical node propagation aggregation representation feature vector; Based on the energy significant descriptor corresponding to the current pipeline pressure local time series associated feature vector, the position-weighted sum of the pipeline pressure local time series historical node propagation aggregation representation feature vector and the current pipeline pressure local time series associated feature vector is calculated to obtain the pipeline pressure time series propagation aggregation representation vector.

4. The remote control method of underground filling equipment according to claim 3, characterized in that: Based on the mean and standard deviation of each pipeline pressure local time series associated feature vector in the sequence of pipeline pressure local time series associated feature vectors, calculating the energy significant descriptors of each pipeline pressure local time series associated feature vector to obtain a sequence of pipeline pressure local time series energy significant descriptors, including: Calculating the mean and standard deviation of each pipeline pressure local time series associated characteristic vector in the sequence of pipeline pressure local time series associated characteristic vectors to obtain a sequence of pipeline pressure local means and a sequence of pipeline pressure local standard deviations; After performing positional differentiation on each group of corresponding local time series correlation feature vectors of pipeline pressure and local mean values ​​in the sequence of local time series correlation feature vectors of pipeline pressure and the sequence of local mean values ​​of pipeline pressure, each positional feature value in the obtained sequence of differential feature vectors is modulated to the fourth power to obtain a sequence of local time series modulation feature vectors of pipeline pressure; Calculating the expected value of each pipeline pressure local time series modulation characteristic vector in the sequence of pipeline pressure local time series modulation characteristic vectors to obtain a sequence of pipeline pressure local time series positive expected factors; Calculating the fourth power modulation value of each local standard deviation of pipeline pressure in the sequence of local standard deviations of pipeline pressure to obtain a sequence of local time series inverse expectation factors of pipeline pressure; Each group of corresponding pipeline pressure local timing positive expected factors and pipeline pressure local timing negative expected factors in the sequence of pipeline pressure local timing positive expected factors and the sequence of pipeline pressure local timing negative expected factors are divided to obtain the sequence of pipeline pressure local timing energy significant descriptors.

5. The remote control method of underground filling equipment according to claim 4, characterized in that: Calculating the energy time decay factor of each historical pipeline pressure local time series associated feature vector in the sequence of the historical pipeline pressure local time series associated feature vectors relative to the current pipeline pressure local time series associated feature vector to obtain a sequence of pipeline pressure local time series energy time decay factors, including: Respectively calculate the pipeline pressure timestamp corresponding to the current pipeline pressure local time series correlation feature vector and the pipeline pressure timestamp corresponding to each historical pipeline pressure local time series correlation feature vector, subtract them and round them down to obtain a sequence of time span values; After calculating the square of each time span value in the sequence of time span values, the obtained sequence of span square modulation values ​​is divided by the square of the time decay inverse scaling parameter by position and multiplied by the decay rate positive scaling parameter by position to obtain a sequence of time series span energy decay coefficients; Taking each time series span energy attenuation coefficient in the sequence of the time series span energy attenuation coefficient as an exponential power, a natural exponential function value with a natural constant e as a base is calculated to obtain a sequence of the local time series energy time attenuation factors of the pipeline pressure.

6. The remote control method of underground filling equipment according to claim 5, characterized in that: Inputting the pipeline pressure time series propagation representation vector and the filling material flow time series propagation representation vector into a query matching interaction network based on characteristic principal components to obtain a pipeline pressure-filling material flow core component significant fusion representation vector, including: Standardizing the pipeline pressure time series propagation representation vector and the filling material flow rate time series propagation representation vector to obtain a standardized pipeline pressure time series propagation representation vector and a standardized filling material flow rate time series propagation representation vector; Calculating the sample covariance matrix of the standardized pipeline pressure time series propagation representation vector and the standardized filling material flow time series propagation representation vector to obtain a pipeline pressure time series sample covariance matrix and a filling material flow time series sample covariance matrix; Performing eigenvector extraction based on matrix decomposition on the pipeline pressure time series sample covariance matrix and the filling material flow time series sample covariance matrix to obtain a set of pipeline pressure time series principal component eigenvectors and a set of filling material flow time series principal component eigenvectors; Inputting the set of pipeline pressure time series principal component feature vectors and the set of filling material flow time series principal component feature vectors into a maximum approximate query matching network to obtain a set of best matching pairs of pipeline pressure time series principal component feature vectors and filling material flow time series principal component feature vectors; Inputting each best matching pair of the pipeline pressure time series principal component feature vector and the filling material flow time series principal component feature vector in the set of best matching pairs of the pipeline pressure time series principal component feature vector and the filling material flow time series principal component feature vector into the semantic fine-grained gating joint module to obtain a set of pipeline pressure-filling material flow time series principal component fusion feature vectors; Cascading each pipeline pressure-filling material flow time series principal component fusion feature vector in the set of pipeline pressure-filling material flow time series principal component fusion feature vectors to obtain the pipeline pressure-filling material flow core component significant fusion representation vector; The step of calculating the sample covariance matrix of the standardized pipeline pressure time series propagation representation vector and the standardized filling material flow time series propagation representation vector to obtain the pipeline pressure time series sample covariance matrix and the filling material flow time series sample covariance matrix includes: Calculate the vector multiplication between the transposed vector of the standardized pipeline pressure time series propagation representation vector and the standardized pipeline pressure time series propagation representation vector, and then divide it by the number of eigenvalues ​​of the standardized pipeline pressure time series propagation representation vector minus one to obtain the pipeline pressure time series sample covariance matrix; The filling material flow time series sample covariance matrix is ​​obtained by calculating the vector multiplication between the transposed vector of the standardized filling material flow time series propagation representation vector and the standardized filling material flow time series propagation representation vector and then dividing it by the number of eigenvalues ​​of the standardized filling material flow time series propagation representation vector minus one by position point.

7. The remote control method of underground filling equipment according to claim 6, characterized in that: The pipeline pressure time series propagation representation vector and the filling material flow rate time series propagation representation vector are normalized to obtain a normalized pipeline pressure time series propagation representation vector and a normalized filling material flow rate time series propagation representation vector, including: Calculating the mean and standard deviation of the pipeline pressure time series propagation representation vector to obtain the pipeline pressure time series mean and pipeline pressure time series standard deviation; Calculating the position difference between the pipeline pressure time series propagation representation vector and the pipeline pressure time series mean value, and then dividing the position difference by the pipeline pressure time series standard deviation to obtain the standardized pipeline pressure time series propagation representation vector; Calculating the mean and standard deviation of the filling material flow time series propagation representation vector to obtain the filling material flow time series mean and the filling material flow time series standard deviation; The positional difference between the filling material flow timing propagation representation vector and the filling material flow timing mean is calculated and then divided by the filling material flow timing standard deviation to obtain the standardized filling material flow timing propagation representation vector.

8. The remote control method of underground filling equipment according to claim 7, characterized in that: Inputting the set of pipeline pressure time series principal component feature vectors and the set of filling material flow time series principal component feature vectors into a maximum approximate query matching network to obtain a set of best matching pairs of pipeline pressure time series principal component feature vectors and filling material flow time series principal component feature vectors, including: Taking the pipeline pressure time series principal component feature vector at a predetermined position in the set of pipeline pressure time series principal component feature vectors to obtain a predetermined pipeline pressure time series principal component feature vector; Calculating the inner products of the predetermined pipeline pressure time series principal component eigenvector and each filling material flow time series principal component eigenvector in the set of filling material flow time series principal component eigenvectors to obtain a sequence of pipeline pressure-filling material flow inner product values; Calculating a norm of a principal component characteristic vector of the predetermined pipeline pressure time series to obtain a norm of a principal component characteristic of the predetermined pipeline pressure time series; Calculating a norm of each filling material flow time series principal component characteristic vector in the set of filling material flow time series principal component characteristic vectors to obtain a sequence of a norm of the filling material flow time series principal component characteristic; Multiplying the sequence of the predetermined pipeline pressure time series principal component feature one norm and the filling material flow time series principal component feature one norm by position points to obtain a sequence of pipeline pressure-filling material flow semantic interaction values; After calculating the point-by-point division between the sequence of pipeline pressure-filling material flow inner product values ​​and the sequence of pipeline pressure-filling material flow semantic interaction values, the filling material flow time series principal component eigenvector corresponding to the maximum eigenvalue in the obtained eigenvalue sequence and the predetermined pipeline pressure time series principal component eigenvector form a set of optimal matching pairs.

9. The remote control method of underground filling equipment according to claim 8, characterized in that: Inputting each best matching pair of the pipeline pressure time series principal component feature vector and the filling material flow time series principal component feature vector in the set of best matching pairs of the pipeline pressure time series principal component feature vector and the filling material flow time series principal component feature vector into the semantic fine-grained gating joint module to obtain a set of pipeline pressure-filling material flow time series principal component fusion feature vectors, including: The best matching pairs of the pipeline pressure time series principal component eigenvectors and the filling material flow time series principal component eigenvectors are respectively subjected to position difference, position dot multiplication and position summation processing to obtain a set of pipeline pressure-filling material flow time series principal component difference eigenvectors, a set of pipeline pressure-filling material flow time series principal component dot product eigenvectors and a set of pipeline pressure-filling material flow time series principal component summation eigenvectors; The set of pipeline pressure-filling material flow time series principal component difference eigenvectors, the set of pipeline pressure-filling material flow time series principal component dot product eigenvectors and the set of pipeline pressure-filling material flow time series principal component sum eigenvectors are cascaded by position to obtain a set of pipeline pressure-filling material flow time series principal component multi-dimensional fusion expression vectors; Each pipeline pressure-filling material flow time series principal component multidimensional fusion expression vector in the set of the pipeline pressure-filling material flow time series principal component multidimensional fusion expression vector is processed by a one-dimensional convolution layer and then by a maximum pooling layer to obtain a set of pipeline pressure-filling material flow time series principal component fusion feature vectors.

10. A remote control system for underground filling equipment, characterized in that: include: A data acquisition module, used to acquire a time queue of pipeline pressure and a time queue of filling material flow value collected by a pressure sensor; A data transmission module, used for transmitting the time queue of the pipeline pressure and the time queue of the filling material flow value to the remote control server through a wireless communication module; A local time series encoding module is used to perform local time series encoding on the time series of the pipeline pressure and the time series of the filling material flow value on the remote control server to obtain a sequence of local time series associated feature vectors of the pipeline pressure and a sequence of local time series associated feature vectors of the filling material flow; A feature forward propagation module, used for inputting the sequence of the pipeline pressure local time series associated feature vectors and the sequence of the filling material flow local time series associated feature vectors into a feature forward propagation network based on significant time series attenuation of node energy to obtain a pipeline pressure time series propagation representation vector and a filling material flow time series propagation representation vector; A significant fusion module, used for inputting the pipeline pressure time series propagation representation vector and the filling material flow rate time series propagation representation vector into a query matching interaction network based on characteristic principal components to obtain a pipeline pressure-filling material flow rate core component significant fusion representation vector; A remote operation instruction determination module is used to input the pipeline pressure-filling material flow core component significant fusion representation vector into a decoder-based remote control instruction generator to obtain a remote operation instruction, wherein the remote operation instruction includes a recommended decoded value of the filling rate at the next time point.

Images