Adaptive cutting head optimization method, system and medium for a roadheader

Abstract

The application discloses a tunneling machine adaptive cutting head optimization method and system and a medium, relates to the tunneling control technical field, and the method comprises the following steps: carrying out cutting section hardness change probability analysis and determining a first hardness sensing frequency and a second hardness sensing frequency; tunneling a target area, collecting and obtaining a first hardness sensing data sequence and a second hardness sensing data sequence; cutting section hardness sensing is carried out, and a cutting section hardness sensing result is obtained; when the difference between the cutting section hardness sensing result and a preset cutting surface hardness is greater than or equal to a preset difference threshold, a cutting head adaptive optimization instruction is triggered; and preset cutting head parameters are optimized and determined to obtain optimized cutting head parameters based on the cutting head adaptive optimization instruction and the cutting section hardness sensing result. The application solves the technical problem that the cutting head parameters cannot be dynamically adjusted according to geological changes in the prior art, leading to low tunneling efficiency, achieves adaptive optimization of the cutting head parameters, and improves the technical effect of tunneling efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of tunneling control, and particularly relates to an adaptive cutting head optimization method and system for a tunneling machine and a medium. BACKGROUND

[0002] When the tunneling machine is operated in a complex geological environment, the cutting section hardness often changes. Since the cutting head parameters in the traditional tunneling process are mostly preset static values, the real-time changes of the actual stratum hardness cannot be reflected, and the cutting parameters are prone to mismatch the actual working conditions. This parameter setting method often leads to reduced tunneling efficiency, increased energy consumption, and even causes equipment overload or accelerated wear and tear, etc., which seriously affects the continuity of tunneling operation. SUMMARY

[0003] The present application provides an adaptive cutting head optimization method and system for a tunneling machine and a medium, which is used to solve the technical problem that the cutting head parameters cannot be dynamically adjusted according to the geological changes in the prior art, resulting in low tunneling efficiency.

[0004] In view of the above problems, the present application provides an adaptive cutting head optimization method and system for a tunneling machine and a medium.

[0005] In a first aspect, the present application provides an adaptive cutting head optimization method for a tunneling machine, the method comprising:

[0006] obtaining a historical tunneling data set of a target area for cutting section hardness change probability analysis, and determining a first hardness perception frequency and a second hardness perception frequency according to an obtained initial hardness perception frequency set; tunneling the target area with a target tunneling machine, collecting cutting motor current, cutting head vibration acceleration and rotary oil cylinder pressure according to the first hardness perception frequency and the second hardness perception frequency respectively, obtaining a first hardness perception data sequence and a second hardness perception data sequence; performing cutting section hardness perception based on the first hardness perception data sequence and the second hardness perception data sequence, obtaining a cutting section hardness perception result, triggering a cutting head adaptive optimization instruction when the difference between the cutting section hardness perception result and the preset cutting surface hardness corresponding to the preset cutting head parameter is greater than or equal to a preset difference threshold; and optimizing the preset cutting head parameter based on the cutting head adaptive optimization instruction and the cutting section hardness perception result, and determining an optimized cutting head parameter.

[0007] In a second aspect, the present application provides an adaptive cutting head optimization system for a tunneling machine, which is used to execute the adaptive cutting head optimization method for a tunneling machine provided by the present application, and the system comprises:

[0008] The probability analysis module is configured to acquire a historical tunneling data set of a target region to perform cutting section hardness change probability analysis, and determine a first hardness perception frequency and a second hardness perception frequency according to an obtained initial hardness perception frequency set; the data acquisition module is configured to tunnel the target region by using a target tunneling machine, and acquire cutting motor current, cutting head vibration acceleration and swing cylinder pressure according to the first hardness perception frequency and the second hardness perception frequency respectively, to obtain a first hardness perception data sequence and a second hardness perception data sequence; the hardness perception module is configured to perform cutting section hardness perception based on the first hardness perception data sequence and the second hardness perception data sequence, to obtain a cutting section hardness perception result, and trigger a cutting head adaptive optimization instruction when a difference between the cutting section hardness perception result and a preset cutting surface hardness corresponding to a preset cutting head parameter is greater than or equal to a preset difference threshold; and the optimization module is configured to optimize the preset cutting head parameter based on the cutting head adaptive optimization instruction and the cutting section hardness perception result, to determine an optimized cutting head parameter.

[0009] In a third aspect, the present application provides a computer readable storage medium storing a computer program, which, when executed by a processor, implements the adaptive cutting head optimization method of the tunneling machine.

[0010] The one or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0011] The present application acquires a historical tunneling data set of a target region to perform cutting section hardness change probability analysis, and determines a first hardness perception frequency and a second hardness perception frequency according to an obtained initial hardness perception frequency set; the target region is tunneled by using a target tunneling machine, and cutting motor current, cutting head vibration acceleration and swing cylinder pressure are acquired according to the first hardness perception frequency and the second hardness perception frequency respectively, to obtain a first hardness perception data sequence and a second hardness perception data sequence; cutting section hardness perception is performed based on the first hardness perception data sequence and the second hardness perception data sequence, to obtain a cutting section hardness perception result, and a cutting head adaptive optimization instruction is triggered when a difference between the cutting section hardness perception result and a preset cutting surface hardness corresponding to a preset cutting head parameter is greater than or equal to a preset difference threshold; and the preset cutting head parameter is optimized based on the cutting head adaptive optimization instruction and the cutting section hardness perception result, to determine an optimized cutting head parameter. The present application solves the technical problem that the cutting head parameter cannot be dynamically adjusted according to geological changes in the prior art, resulting in low tunneling efficiency, and achieves the technical effect of adaptive optimization of the cutting head parameter and improvement of the tunneling efficiency by analyzing the hardness perception frequency of the historical tunneling data, perceiving the cutting section hardness in real time and triggering parameter optimization. BRIEF DESCRIPTION OF DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort based on these drawings.

[0013] Figure 1 The adaptive cutting head optimization method flowchart of the tunneling machine provided in the embodiment of the present application is shown in the figure.

[0014] Figure 2 The adaptive cutting head optimization system structure schematic diagram of the tunneling machine provided in the embodiment of the present application is shown in the figure.

[0015] Legend: probability analysis module 11, data acquisition module 12, hardness perception module 13, optimization module 14. DETAILED DESCRIPTION

[0016] The present application provides an adaptive cutting head optimization method, system and medium of a tunneling machine. The technical problem of low tunneling efficiency caused by the inability to dynamically adjust the cutting head parameters according to the geological changes in the prior art is solved. The hardness perception frequency analysis of the historical tunneling data is performed, the cutting section hardness is perceived in real time, and the parameter optimization is triggered. The technical effect of adaptive optimization of the cutting head parameters and improvement of the tunneling efficiency is achieved.

[0017] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without any creative effort fall within the scope of protection of the present application.

[0018] It should be noted that any variation of the terms "comprise" and "have" is intended to cover non-exclusive inclusion, for example, a process, method, system, product or server comprising a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or modules not clearly listed or inherent to these processes, methods, products or devices.

[0019] Embodiment one, as shown in the figure, the present application provides an adaptive cutting head optimization method of a tunneling machine, the method comprises: Figure 1

[0020] Step S100: obtaining the historical tunneling data set of the target area to analyze the cutting section hardness change probability, and determining the first hardness perception frequency and the second hardness perception frequency according to the obtained initial hardness perception frequency set. ​

[0021] In the embodiment of the present application, first, the historical tunneling data set of the target area is obtained from the preset historical database, and then the cutting section hardness change probability analysis is performed. Specifically, first, the historical tunneling data set is segmented based on the cutting motor current, the cutting head vibration acceleration and the rotary oil cylinder pressure, and a historical tunneling perception data sequence set is constructed. Then, the historical tunneling perception data sequence set is processed by near neighbor fusion, a near neighbor fusion number set is obtained, and a cutting section hardness change probability set is constructed through proportional calculation and difference conversion. Then, the cutting section hardness change probability set and the preset hardness perception frequency are multiplied one by one to generate an initial hardness perception frequency set. Finally, the initial hardness perception frequency set is screened to obtain a first hardness perception frequency and a second hardness perception frequency representing different hardness perception distribution characteristics.

[0022] Further, the method provided by the embodiment of the application includes obtaining the historical tunneling data set of the target area to perform cutting section hardness change probability analysis, and determining the first hardness perception frequency and the second hardness perception frequency according to the obtained initial hardness perception frequency set.

[0023] The historical tunneling data set is segmented and extracted based on the cutting motor current, the cutting head vibration acceleration and the rotary oil cylinder pressure to determine a historical tunneling perception data sequence set; the historical tunneling perception data sequence set is processed by near neighbor fusion to determine a near neighbor fusion number set; the near neighbor fusion number set is compared with the total number of historical tunneling perception data in each historical tunneling perception data sequence in the historical tunneling perception data sequence set, and the difference between 1 and each ratio value is taken as a cutting section hardness change probability set; the cutting section hardness change probability set is multiplied by a preset hardness perception frequency to obtain an initial hardness perception frequency set; and the initial hardness perception frequency set is screened to determine a first hardness perception frequency and a second hardness perception frequency.

[0024] In the embodiment of the present application, first, the historical tunneling data set is time series segmented based on the cutting motor current, the cutting head vibration acceleration and the rotary oil cylinder pressure. This process uses a sliding window method, sets a fixed length time window (such as 10 seconds) to cut the original data stream, and each window forms a historical tunneling perception data sequence, and finally a historical tunneling perception data sequence set is obtained.

[0025] Next, the historical tunneling perception data sequence set is fused by neighborhood, in which, similarity analysis is performed on the historical tunneling perception data sequence set, and corresponding neighborhood similarity sequence sets are obtained based on cutting motor current similarity, cutting head vibration acceleration similarity and rotary oil cylinder pressure similarity respectively. Then, whether the similarity values in the neighborhood similarity sequence set are greater than or equal to a preset neighborhood similarity threshold is judged in sequence, if the condition is met, the corresponding neighborhood fusion number in the neighborhood fusion number set initially set as 0 is increased by 1, and finally the complete neighborhood fusion number set is determined.

[0026] Then, each neighborhood fusion number in the neighborhood fusion number set is divided by the total number of historical tunneling perception data in the historical tunneling perception data sequence in the corresponding historical tunneling perception data sequence set to calculate the ratio. Then, the ratio is calculated by formula 1-ratio, and finally the cutting section hardness change probability set reflecting the fluctuation degree is obtained. The closer each value in the set to 1, the more unstable the perception data of the data segment, that is, the more frequent the change of stratum hardness; on the contrary, close to 0 indicates that the geological conditions of this segment are relatively uniform and stable. For example, if a historical tunneling perception data sequence contains 100 historical tunneling perception data, and the corresponding neighborhood fusion number is 20, then the ratio is 20 / 100=0.2, and the corresponding cutting section hardness change probability is 1-0.2=0.8.

[0027] After obtaining the cutting section hardness change probability set, the preset hardness perception frequency set is obtained by multiplying the preset hardness perception frequency set set by the technical expert in sequence.

[0028] Finally, the initial hardness perception frequency set is screened, in which, the frequency with the highest occurrence frequency is first determined as the hardness perception frequency screening center, and its initial hardness perception frequency screening center neighborhood is constructed based on the preset screening radius, and the corresponding neighborhood density is calculated. Then, the neighborhood is iteratively diffused according to the preset diffusion radius, and the density of the diffused hardness perception frequency screening center neighborhood is continuously compared with the initial density, if the density is improved, the diffusion continues until the preset diffusion times are reached, and finally the target diffusion hardness perception frequency screening center neighborhood is obtained, and the maximum value thereof is taken as the first hardness perception frequency and the minimum value thereof is taken as the second hardness perception frequency. If the diffusion density is not improved, the maximum value and the minimum value are directly extracted from the initial neighborhood as the first hardness perception frequency and the second hardness perception frequency.

[0029] Further, the method provided by the application embodiment further comprises:

[0030] The near neighbor historical tunneling perception data near neighbor similarity analysis is performed on the historical tunneling perception data sequence set from the cutting motor current similarity, the cutting head vibration acceleration similarity, and the swing oil cylinder pressure similarity, to obtain a near neighbor similarity sequence set; whether the near neighbor similarity in the near neighbor similarity sequence set is greater than or equal to a preset near neighbor similarity threshold is sequentially judged, and if yes, the near neighbor fusion quantity in the near neighbor fusion quantity set initially set as 0 is increased by 1, to obtain the near neighbor fusion quantity set.

[0031] In the embodiments of the present application, when the historical tunneling perception data sequence set is processed, firstly, the near neighbor historical tunneling perception data near neighbor similarity analysis is performed on each historical tunneling perception data sequence in the historical tunneling perception data sequence set and the remaining historical tunneling perception data sequences in the set from the three dimensions of the cutting motor current similarity, the cutting head vibration acceleration similarity, and the swing oil cylinder pressure similarity. The analysis adopts the Euclidean distance method, point-by-point difference calculation is performed on the perception data of each pair of historical tunneling perception data sequences in the same dimension, and normalization processing is performed based on the overall difference level, so as to convert the distance value between each pair of historical tunneling perception data sequences into a similarity score between 0 and 1, to form a near neighbor similarity sequence set.

[0032] After the near neighbor similarity sequence set is obtained, each near neighbor similarity in the near neighbor similarity sequence set is sequentially judged, and the threshold comparison method is adopted to compare whether the near neighbor similarity is greater than or equal to the preset near neighbor similarity threshold set in the system. If a near neighbor similarity is greater than or equal to the preset near neighbor similarity threshold, it is considered that the two historical tunneling perception data sequences corresponding to the near neighbor similarity form an effective near neighbor in the current perception dimension.

[0033] Each pair of historical tunneling perception data sequences that meet the condition is judged, and the near neighbor fusion quantity corresponding to the historical tunneling perception data sequences is increased by 1 in the near neighbor fusion quantity set initially set as 0. The operation is performed in the three dimensions of the cutting motor current similarity, the cutting head vibration acceleration similarity, and the swing oil cylinder pressure similarity, to finally form the near neighbor fusion quantity set.

[0034] Further, the method provided in the embodiments of the application further includes the following steps of:

[0035] The frequency of the initial hardness perception frequency with the highest occurrence in the set of initial hardness perception frequencies is counted, and a hardness perception frequency screening center is obtained; an initial hardness perception frequency screening center neighborhood of the hardness perception frequency screening center is constructed according to a preset screening radius, and an initial hardness perception frequency screening center neighborhood density is calculated; the initial hardness perception frequency screening center neighborhood is diffused according to a preset diffusion radius, a diffused hardness perception frequency screening center neighborhood is obtained, and a diffused hardness perception frequency screening center neighborhood density of the diffused hardness perception frequency screening center neighborhood is calculated; when the diffused hardness perception frequency screening center neighborhood density is greater than or equal to the initial diffused hardness perception frequency screening center neighborhood density, the diffused hardness perception frequency screening center neighborhood is continuously diffused according to the preset diffusion radius until a preset diffusion number is met, and a target diffused hardness perception frequency screening center neighborhood is obtained; the maximum initial hardness perception frequency in the target diffused hardness perception frequency screening center neighborhood is taken as a first hardness perception frequency, and the minimum initial hardness perception frequency in the target diffused hardness perception frequency screening center neighborhood is taken as a second hardness perception frequency.

[0036] In the embodiments of the present application, firstly, the frequency of all the initial hardness perception frequencies appearing in the set of initial hardness perception frequencies is counted using a frequency counting method, and the initial hardness perception frequency with the highest occurrence frequency is identified and selected as the hardness perception frequency screening center.

[0037] Subsequently, an interval centered on the hardness perception frequency screening center is constructed according to a preset screening radius, and an initial hardness perception frequency screening center neighborhood is formed. This step adopts an interval construction method, that is, the hardness perception frequency screening center is taken as the center, and the preset screening radius is added and subtracted to the center value respectively, to form a symmetrical closed interval. For example, if the hardness perception frequency screening center is 10 Hz and the preset screening radius is 1 Hz, the initial hardness perception frequency screening center neighborhood is [9 Hz, 11 Hz].

[0038] Then, the density of the constructed initial hardness perception frequency screening center neighborhood is calculated, and a unit interval density calculation method is adopted, that is, the number of initial hardness perception frequencies in the neighborhood interval is counted, and then divided by the area of the neighborhood interval (that is, the interval length, twice the screening radius). For example, if there are 6 initial hardness perception frequencies in the [9 Hz, 11 Hz] interval, and the interval length is 2 Hz, then the initial hardness perception frequency screening center neighborhood density is 3 / Hz.

[0039] In order to enhance the screening flexibility and range, the neighborhood is expanded by using a fixed step expansion strategy. According to the preset diffusion radius set by the system, the original initial hardness perception frequency screening center neighborhood is extended by a corresponding frequency width on both sides to generate a diffusion hardness perception frequency screening center neighborhood. For example, based on the above [9Hz, 11Hz], if the preset diffusion radius is 1Hz, the diffusion neighborhood is [8Hz, 12Hz].

[0040] After obtaining the diffusion hardness perception frequency screening center neighborhood, the number of initial hardness perception frequencies in the neighborhood is re-counted, and the diffusion hardness perception frequency screening center neighborhood density is calculated using the same method. The judgment logic is that if the diffusion hardness perception frequency screening center neighborhood density is greater than or equal to the initial hardness perception frequency screening center neighborhood density of the previous round, it means that the expansion area still has significant frequency aggregation. At this time, continue to take the current neighborhood as the basis and expand again according to the preset diffusion radius. This operation is carried out by iterative diffusion method, and the above expansion and density calculation process is repeated until the set preset diffusion times limit is reached, and the final target diffusion hardness perception frequency screening center neighborhood is obtained. If the diffusion hardness perception frequency screening center neighborhood density of a certain round is less than the diffusion hardness perception frequency screening center neighborhood density of the last round, it is considered that the frequency aggregation begins to decline and no longer has diffusion value. At this time, the diffusion operation is terminated, and the diffusion hardness perception frequency screening center neighborhood with higher density in the last round is determined as the final target diffusion hardness perception frequency screening center neighborhood.

[0041] Finally, the frequency range boundary is screened from the target diffusion hardness perception frequency screening center neighborhood. Specifically, the extreme value extraction method is used, that is, the maximum initial hardness perception frequency in the neighborhood is taken as the first hardness perception frequency, and the minimum initial hardness perception frequency is taken as the second hardness perception frequency, to complete the entire frequency screening process.

[0042] Further, the method provided by the application embodiment further comprises:

[0043] When the diffusion hardness perception frequency screening center neighborhood density is less than the initial diffusion hardness perception frequency screening center neighborhood density, the maximum initial hardness perception frequency in the initial diffusion hardness perception frequency screening center neighborhood is taken as the first hardness perception frequency, and the minimum initial hardness perception frequency in the initial diffusion hardness perception frequency screening center neighborhood is taken as the second hardness perception frequency.

[0044] In the embodiment of the present application, when the diffusion hardness perception frequency screening center neighborhood density is less than or equal to the initial diffusion hardness perception frequency screening center neighborhood density, it indicates that with the expansion of the diffusion range, the concentration degree of the frequency distribution decreases, and further diffusion will lead to the weakening of the representative of the perceived frequency. Therefore, the diffusion operation is terminated, and the initial diffusion hardness perception frequency screening center neighborhood constructed in the last round is directly reserved as the final reference interval. Next, the extreme value extraction method is used to extract the maximum and minimum values from the initial diffusion hardness perception frequency screening center neighborhood, wherein the maximum value is determined as the first hardness perception frequency, and the minimum value is determined as the second hardness perception frequency.

[0045] Step S200: tunneling the target region by using the target tunneling machine, collecting the cutting motor current, the cutting head vibration acceleration and the swing cylinder pressure according to the first hardness perception frequency and the second hardness perception frequency respectively, and obtaining the first hardness perception data sequence and the second hardness perception data sequence.

[0046] In the embodiment of the present application, after the first hardness perception frequency and the second hardness perception frequency are determined, the target tunneling machine is used to perform tunneling operation on the target region. During the tunneling process, the perception collection is simultaneously performed according to the first hardness perception frequency and the second hardness perception frequency. Specifically, the three types of key operating parameters of the cutting motor current, the cutting head vibration acceleration and the swing cylinder pressure are collected in parallel at two different perception frequencies. The data collected according to the first hardness perception frequency constitutes the first hardness perception data sequence, and the data collected according to the second hardness perception frequency constitutes the second hardness perception data sequence.

[0047] Step S300: performing cutting section hardness perception based on the first hardness perception data sequence and the second hardness perception data sequence, and obtaining a cutting section hardness perception result; when a difference between the cutting section hardness perception result and a preset cutting surface hardness corresponding to a preset cutting head parameter is greater than or equal to a preset difference threshold value, triggering a cutting head adaptive optimization instruction.

[0048] In the embodiment of the present application, when the cutting section hardness perception is performed based on the first hardness perception data sequence and the second hardness perception data sequence, the first hardness perception data sequence and the second hardness perception data sequence are respectively input into the first cutting section hardness perception branch and the second cutting section hardness perception branch which are pre-trained to perform time sequence perception, and the first cutting section hardness perception feature and the second cutting section hardness perception feature are determined. Subsequently, the first cutting section hardness perception feature and the second cutting section hardness perception feature are analyzed interactively to determine a unified cutting section hardness perception result.

[0049] Afterwards, the cutting section hardness sensing result is subtracted by the preset cutting surface hardness corresponding to the preset cutting head parameter to obtain a difference value, and then the difference value is compared with a preset difference value threshold. When the difference value is greater than or equal to the preset difference value threshold, it is judged that there is a significant mismatch between the current tunneling environment and the parameter setting, and the cutting head adaptive optimization instruction is automatically triggered.

[0050] Further, the method provided by the application embodiment further comprises:

[0051] The first cutting section hardness sensing branch and the second cutting section hardness sensing branch are respectively used for time sequence sensing of the first hardness sensing data sequence and the second hardness sensing data sequence to determine first cutting section hardness sensing features and second cutting section hardness sensing features; and the first cutting section hardness sensing features and the second cutting section hardness sensing features are interacted to determine the cutting section hardness sensing result.

[0052] In the application embodiment, first, the first cutting section hardness sensing branch and the second cutting section hardness sensing branch which are pre-trained are respectively used for time sequence sensing of the first hardness sensing data sequence and the second hardness sensing data sequence. The first cutting section hardness sensing branch and the second cutting section hardness sensing branch both obtain network parameters with feature extraction capability through training. The data used for training includes historical tunneling sensing samples with labels, wherein the input is historical first hardness sensing data sequence and second hardness sensing data sequence, and the output label is actually measured stratum hardness. Through a large amount of historical tunneling data with hardness annotation for supervised learning, each sensing branch can automatically identify key features related to hardness according to sensing signals. After training, in real-time application, the first hardness sensing data sequence is input into the first cutting section hardness sensing branch to extract first cutting section hardness sensing features. At the same time, the second hardness sensing data sequence is input into the second cutting section hardness sensing branch to extract second cutting section hardness sensing features.

[0053] Then the first cutting section hardness perception feature and the second cutting section hardness perception feature are interacted. In this process, by performing multi-dimensional approximate analysis on the first cutting section hardness perception feature and the second cutting section hardness perception feature, the similarity degrees of the two features in each feature dimension are calculated, and a multi-dimensional approximate analysis similarity coefficient set is obtained. Then, a perception feature adjacency matrix reflecting the feature correlation is constructed based on the multi-dimensional approximate analysis similarity coefficient set. Finally, the perception feature adjacency matrix is used to perform feature enhancement on the second cutting section hardness perception feature, so as to generate the cutting section hardness perception result.

[0054] Further, the method provided by the application embodiment further comprises the following steps:

[0055] The first cutting section hardness perception feature and the second cutting section hardness perception feature are subjected to multi-dimensional approximate analysis to obtain a multi-dimensional approximate analysis similarity coefficient set; a perception feature adjacency matrix is constructed based on the multi-dimensional approximate analysis similarity coefficient set; and the second cutting section hardness perception feature is enhanced through the perception feature adjacency matrix to obtain the cutting section hardness perception result.

[0056] In the application embodiment, first, the first cutting section hardness perception feature and the second cutting section hardness perception feature are subjected to multi-dimensional approximate analysis. This process uses a cosine similarity calculation method, and by calculating the cosine value of the angle between the feature vectors in the same dimension, a similarity score is obtained, which reflects the closeness between the two features in this dimension. All dimensions are calculated in turn, and finally a set of numerical values is obtained, which constitutes a multi-dimensional approximate analysis similarity coefficient set.

[0057] After the calculation of the multi-dimensional approximate analysis similarity coefficient set is completed, the construction of the perception feature adjacency matrix is started. First, all the numerical values in the multi-dimensional approximate analysis similarity coefficient set are subjected to normalization processing, so as to unify the numerical scales between the features and avoid the influence of value differences in different dimensions on the correlation judgment. After normalization, an empty two-dimensional matrix is initialized, and the rows and columns of the matrix correspond to the dimension orders of the first cutting section hardness perception feature and the second cutting section hardness perception feature respectively. Then, the normalized similarity coefficients are sequentially filled into the matrix according to their corresponding feature dimension positions, and the construction of the perception feature adjacency matrix is completed.

[0058] Then, a trained neural network model is used to enhance the second cutting section hardness perception feature. The neural network model is a graph structure modeling network, and its training process is based on known perception features and actual stratum hardness labels in historical tunneling data, and is performed in a supervised learning manner. During training, the historical second cutting section hardness perception features and their adjacency matrix with other features are input, and the output is the corresponding stratum hardness value. By continuously adjusting the network parameters through back propagation, the model can learn how to optimize feature representation based on the association structure between features. After training, the model has the ability to enhance feature representation using adjacency matrices. In actual application, the neural network model takes the perception feature adjacency matrix as the structure input and the second cutting section hardness perception feature as the feature input. Through the information propagation mechanism between nodes in the graph, the associated information from the first cutting section hardness perception feature is integrated to update and enhance the original feature. Finally, the second cutting section hardness perception feature is enhanced by the neural network model, and the cutting section hardness perception result is obtained.

[0059] Step S400: optimizing the preset cutting head parameters based on the cutting head adaptive optimization instruction and the cutting section hardness perception result, and determining the optimized cutting head parameters.

[0060] In the embodiments of the present application, when the preset cutting head parameters are optimized based on the cutting head adaptive optimization instruction and the cutting section hardness perception result, the cutting head adaptive optimization instruction is used to call the pre-constructed parameter optimizer. Then, the cutting section hardness perception result and the preset cutting head parameters are input into the parameter optimizer for analysis, and finally the optimized cutting head parameters matched with the current stratum condition are output.

[0061] Further, the method provided by the embodiments of the present application, based on the cutting head adaptive optimization instruction and the cutting section hardness perception result, the preset cutting head parameters are optimized to determine the optimized cutting head parameters, further comprising:

[0062] Pre-constructing a parameter optimizer, wherein the parameter optimizer is obtained by training a support vector machine based on training data; the parameter optimizer is called based on the cutting head adaptive optimization instruction, and the cutting section hardness perception result and the preset cutting head parameters are analyzed to obtain the optimized cutting head parameters.

[0063] In the embodiment of the present application, first, a parameter optimizer is constructed in advance, which is based on a support vector machine model and is trained by a supervised learning method. In the training process, historical tunneling operation data is extracted from a historical database to construct training samples. Each training sample contains three elements: the input of stratum conditions, i.e., the historical cutting face hardness perception result; the control configuration used in the operation, i.e., the corresponding preset cutting head parameters, including cutting head speed, advance speed, roller spacing, etc.; and the actual effect of the tunneling operation, which is used to mark whether the parameter configuration is effective under the hardness condition. According to the effect, the samples are divided into positive samples and negative samples, and the support vector machine model is trained to identify the matching relationship between the optimal parameter combination and the stratum condition.

[0064] After the training is completed, the current cutting face hardness perception result and the preset cutting head parameters are input into the parameter optimizer for analysis to obtain optimized cutting head parameters that are more matched with the current stratum condition.

[0065] In the embodiment of the present application, as described above, the embodiment of the present application has at least the following technical effects:

[0066] The present application obtains a historical tunneling data set of a target region for cutting face hardness change probability analysis, and determines a first hardness perception frequency and a second hardness perception frequency according to an obtained initial hardness perception frequency set; tunnels the target region using a target tunneling machine, collects cutting motor current, cutting head vibration acceleration and rotary oil cylinder pressure according to the first hardness perception frequency and the second hardness perception frequency respectively, and obtains a first hardness perception data sequence and a second hardness perception data sequence; performs cutting face hardness perception based on the first hardness perception data sequence and the second hardness perception data sequence, obtains a cutting face hardness perception result, triggers a cutting head adaptive optimization instruction when the difference between the cutting face hardness perception result and the preset cutting face hardness corresponding to the preset cutting head parameters is greater than or equal to a preset difference threshold, optimizes the preset cutting head parameters based on the cutting head adaptive optimization instruction and the cutting face hardness perception result, and determines optimized cutting head parameters. The present application solves the technical problem that the cutting head parameters cannot be dynamically adjusted according to geological changes in the prior art, resulting in low tunneling efficiency, and achieves the technical effects of adaptive optimization of cutting head parameters and improvement of tunneling efficiency by analyzing the hardness perception frequency of historical tunneling data, real-time perception of cutting face hardness and triggering of parameter optimization.

[0067] Embodiment two, based on the same inventive concept as the adaptive cutting head optimization method of the tunneling machine in the foregoing embodiments, as shown in Figure 2 The present application provides an adaptive cutting head optimization system for a tunneling machine, and the system and method embodiments in the present application are based on the same inventive concept. The system includes:

[0068] The probability analysis module 11 is used to acquire a historical tunneling data set of a target area to perform cutting section hardness change probability analysis, and to determine a first hardness perception frequency and a second hardness perception frequency according to an obtained initial hardness perception frequency set; the data acquisition module 12 is used to tunnel the target area by using a target tunneling machine, to acquire cutting motor current, cutting head vibration acceleration and swing cylinder pressure according to the first hardness perception frequency and the second hardness perception frequency respectively, and to obtain a first hardness perception data sequence and a second hardness perception data sequence; the hardness perception module 13 is used to perform cutting section hardness perception based on the first hardness perception data sequence and the second hardness perception data sequence, to obtain a cutting section hardness perception result, and to trigger a cutting head adaptive optimization instruction when a difference between the cutting section hardness perception result and a preset cutting surface hardness corresponding to a preset cutting head parameter is greater than or equal to a preset difference threshold value; and the optimization module 14 is used to optimize the preset cutting head parameter based on the cutting head adaptive optimization instruction and the cutting section hardness perception result, and to determine an optimized cutting head parameter.

[0069] Further, the system is also used to implement the following functions:

[0070] The historical tunneling data set is segmented and extracted based on the cutting motor current, the cutting head vibration acceleration and the swing cylinder pressure to determine a historical tunneling perception data sequence set; the historical tunneling perception data sequence set is subjected to near neighbor fusion to determine a near neighbor fusion quantity set; the near neighbor fusion quantity set is compared with a total number of historical tunneling perception data in each historical tunneling perception data sequence in the historical tunneling perception data sequence set respectively, and a difference between 1 and each comparison value is taken as a cutting section hardness change probability set; the cutting section hardness change probability set is multiplied by a preset hardness perception frequency respectively to obtain an initial hardness perception frequency set; and the initial hardness perception frequency set is screened to determine a first hardness perception frequency and a second hardness perception frequency.

[0071] Further, the system is also used to implement the following functions:

[0072] The historical tunneling perception data sequence set is subjected to near neighbor historical tunneling perception data near neighbor similarity analysis respectively based on the cutting motor current similarity, the cutting head vibration acceleration similarity and the swing cylinder pressure similarity to obtain a near neighbor similarity sequence set; whether a near neighbor similarity in the near neighbor similarity sequence set is greater than or equal to a preset near neighbor similarity threshold value is judged in sequence, if yes, a near neighbor fusion quantity in a near neighbor fusion quantity set initially set as 0 is increased by 1 to obtain the near neighbor fusion quantity set.

[0073] Further, the system is also used to implement the following functions:

[0074] The initial hardness sensing frequency with the highest frequency in the initial hardness sensing frequency set is statistically analyzed to obtain a hardness sensing frequency screening center. A neighborhood of the initial hardness sensing frequency screening center is constructed according to a preset screening radius, and the density of this neighborhood is calculated. The neighborhood of the initial hardness sensing frequency screening center is diffused according to a preset diffusion radius to obtain a diffused hardness sensing frequency screening center neighborhood, and its density is calculated. When the density of the diffused hardness sensing frequency screening center neighborhood is greater than or equal to the density of the initial diffused hardness sensing frequency screening center neighborhood, diffusion continues according to the preset diffusion radius until a preset number of diffusions is met, obtaining a target diffused hardness sensing frequency screening center neighborhood. The maximum initial hardness sensing frequency in the target diffused hardness sensing frequency screening center neighborhood is taken as the first hardness sensing frequency, and the minimum initial hardness sensing frequency in the target diffused hardness sensing frequency screening center neighborhood is taken as the second hardness sensing frequency.

[0075] Furthermore, the system is also used to implement the following functions:

[0076] When the density of the neighborhood of the diffusion hardness sensing frequency screening center is less than the density of the initial diffusion hardness sensing frequency screening center, the maximum initial hardness sensing frequency in the neighborhood of the initial diffusion hardness sensing frequency screening center is taken as the first hardness sensing frequency, and the minimum initial hardness sensing frequency in the neighborhood of the initial diffusion hardness sensing frequency screening center is taken as the second hardness sensing frequency.

[0077] Furthermore, the system is also used to implement the following functions:

[0078] The first and second hardness sensing branches of the cutting section are used to perform time-series sensing on the first hardness sensing data sequence and the second hardness sensing data sequence, respectively, to determine the hardness sensing features of the first and second cutting sections; the first and second hardness sensing features of the cutting section are interacted to determine the hardness sensing result of the cutting section.

[0079] Furthermore, the system is also used to implement the following functions:

[0080] A multidimensional approximation analysis is performed on the hardness perception features of the first and second cut surfaces to determine a set of similarity coefficients for the multidimensional approximation analysis. A perception feature adjacency matrix is ​​constructed based on the set of similarity coefficients for the multidimensional approximation analysis. The hardness perception features of the second cut surface are enhanced by the perception feature adjacency matrix to obtain the hardness perception result of the cut surface.

[0081] Furthermore, the system is also used to implement the following functions:

[0082] A pre-constructed parameter optimizer is obtained by training a support vector machine with positive and negative samples based on training data. The parameter optimizer is invoked based on the adaptive optimization instruction of the cutting head to analyze the hardness perception result of the cutting section and the preset cutting head parameters to obtain the optimized cutting head parameters.

[0083] In Embodiment 3, based on the adaptive cutting head optimization method for tunneling machines in the foregoing embodiments and using the same inventive concept, this application also provides a computer-readable storage medium storing a computer program that, when executed, implements the steps of any of the methods described in Embodiment 1 above.

[0084] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0085] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0086] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. An adaptive cutting head optimization method for tunneling machines, characterized in that, The method includes: The historical tunneling data set of the target area is obtained to perform probability analysis of the hardness change of the cutting section, and the first hardness sensing frequency and the second hardness sensing frequency are determined based on the obtained initial hardness sensing frequency set. The target area is tunneled using a target tunneling machine. The current of the cutting motor, the vibration acceleration of the cutting head and the pressure of the rotary cylinder are collected according to the first hardness sensing frequency and the second hardness sensing frequency, respectively, to obtain the first hardness sensing data sequence and the second hardness sensing data sequence. Based on the first hardness sensing data sequence and the second hardness sensing data sequence, the hardness of the cutting section is sensed to obtain the hardness sensing result of the cutting section. When the difference between the hardness sensing result of the cutting section and the hardness of the preset cutting surface corresponding to the preset cutting head parameters is greater than or equal to the preset difference threshold, the cutting head adaptive optimization command is triggered. The preset cutting head parameters are optimized based on the cutting head adaptive optimization command and the cutting section hardness sensing result to determine the optimized cutting head parameters; Historical tunneling data sets for the target area are acquired for probability analysis of hardness variation at the cutting section. Based on the obtained initial hardness sensing frequency set, the first and second hardness sensing frequencies are determined, including: Based on the cutting motor current, cutting head vibration acceleration and rotary cylinder pressure, the historical tunneling data set is segmented and extracted to determine the historical tunneling sensing data sequence set. The historical tunneling sensing data sequence set is subjected to nearest neighbor fusion to determine the set of nearest neighbor fusion quantities; The nearest neighbor fusion quantity set is compared with the total number of historical tunneling sensing data in each historical tunneling sensing data sequence in the historical tunneling sensing data sequence set, and the difference between 1 and each ratio is taken as the set of probability of change of cutting section hardness. The set of probabilities of hardness change of the cut surface is multiplied by the preset hardness sensing frequency to obtain the initial set of hardness sensing frequencies. Filtering the initial set of hardness sensing frequencies to determine the first hardness sensing frequency and the second hardness sensing frequency includes: The initial hardness sensing frequency that appears most frequently in the initial hardness sensing frequency set is counted to obtain the hardness sensing frequency screening center. According to the preset screening radius, the initial hardness sensing frequency screening center neighborhood is constructed, and the density of the initial hardness sensing frequency screening center neighborhood is calculated. According to the preset diffusion radius, the initial hardness sensing frequency screening center neighborhood is diffused to obtain the diffused hardness sensing frequency screening center neighborhood, and the diffusion hardness sensing frequency screening center neighborhood density is calculated. When the density of the diffusion hardness sensing frequency screening center neighborhood is greater than or equal to the initial diffusion hardness sensing frequency screening center neighborhood density, the diffusion hardness sensing frequency screening center neighborhood continues to be diffused according to the preset diffusion radius until the preset diffusion number is met, and the target diffusion hardness sensing frequency screening center neighborhood is obtained. The maximum initial hardness sensing frequency in the target diffusion hardness sensing frequency screening center neighborhood is taken as the first hardness sensing frequency, and the minimum initial hardness sensing frequency in the target diffusion hardness sensing frequency screening center neighborhood is taken as the second hardness sensing frequency.

2. The adaptive cutting head optimization method for a tunneling machine as described in claim 1, characterized in that, The historical tunneling sensing data sequence set is subjected to nearest neighbor fusion to determine the set of nearest neighbor fusion quantities, including: The similarity analysis of the historical tunneling sensing data sequence set is performed on the similarity of the cutting motor current, the similarity of the cutting head vibration acceleration, and the similarity of the rotary cylinder pressure to obtain a set of near-neighbor similarity sequences. The nearest neighbor similarity sequence set is sequentially determined to be greater than or equal to a preset nearest neighbor similarity threshold. If so, the nearest neighbor fusion number in the nearest neighbor fusion number set, which is initially 0, is incremented by 1 to obtain the nearest neighbor fusion number set.

3. The adaptive cutting head optimization method for a tunneling machine as described in claim 1, characterized in that, When the density of the neighborhood of the diffusion hardness sensing frequency screening center is less than the density of the initial diffusion hardness sensing frequency screening center, the maximum initial hardness sensing frequency in the neighborhood of the initial diffusion hardness sensing frequency screening center is taken as the first hardness sensing frequency, and the minimum initial hardness sensing frequency in the neighborhood of the initial diffusion hardness sensing frequency screening center is taken as the second hardness sensing frequency.

4. The adaptive cutting head optimization method for a tunneling machine as described in claim 1, characterized in that, Based on the first hardness sensing data sequence and the second hardness sensing data sequence, the hardness of the cutting section is sensed to obtain the hardness sensing result of the cutting section. When the difference between the hardness sensing result of the cutting section and the preset cutting surface hardness corresponding to the preset cutting head parameters is greater than or equal to a preset difference threshold, an adaptive optimization instruction for the cutting head is triggered, including: The first and second hardness sensing data sequences are time-series sensing data using the first and second hardness sensing branches of the cutting section, respectively, to determine the hardness sensing features of the first and second cutting sections. The hardness perception results of the cutting section are determined by interacting with the first and second cutting section hardness perception features.

5. The adaptive cutting head optimization method for a tunneling machine as described in claim 4, characterized in that, Interacting with the first and second cut surface hardness sensing features, the hardness sensing result of the cut surface is determined, including: A multidimensional approximation analysis is performed on the hardness perception characteristics of the first and second cut surfaces to determine the set of multidimensional approximation analysis similarity coefficients. A perceptual feature adjacency matrix is ​​constructed based on the set of similarity coefficients obtained from the multidimensional approximation analysis. The hardness perception feature of the second cut surface is enhanced by the adjacency matrix of the perception feature to obtain the hardness perception result of the cut surface.

6. The adaptive cutting head optimization method for a tunneling machine as described in claim 1, characterized in that, Based on the adaptive optimization command of the cutting head and the hardness perception result of the cutting surface, the preset cutting head parameters are optimized to determine the optimized cutting head parameters, including: A pre-built parameter optimizer is provided, wherein the parameter optimizer is obtained by training the support vector machine with positive and negative samples based on the training data; Based on the adaptive optimization command of the cutting head, the parameter optimizer is invoked to analyze the hardness perception result of the cutting section and the preset cutting head parameters to obtain the optimized cutting head parameters.

7. An adaptive cutting head optimization system for a tunneling machine, characterized in that, The system is used to execute the adaptive cutting head optimization method for the tunneling machine as described in any one of claims 1-6, and the system includes: The probability analysis module is used to acquire historical tunneling data sets of the target area to perform probability analysis of hardness changes in the cutting section, and to determine the first hardness sensing frequency and the second hardness sensing frequency based on the acquired initial hardness sensing frequency set. The data acquisition module is used to excavate the target area using the target tunneling machine, and to collect the cutting motor current, cutting head vibration acceleration and rotary cylinder pressure according to the first hardness sensing frequency and the second hardness sensing frequency, respectively, to obtain the first hardness sensing data sequence and the second hardness sensing data sequence. The hardness sensing module is used to sense the hardness of the cutting surface based on the first hardness sensing data sequence and the second hardness sensing data sequence, and obtain the hardness sensing result of the cutting surface. When the difference between the hardness sensing result of the cutting surface and the hardness of the preset cutting surface corresponding to the preset cutting head parameters is greater than or equal to the preset difference threshold, the cutting head adaptive optimization instruction is triggered. The optimization module is used to optimize the preset cutting head parameters based on the adaptive optimization command of the cutting head and the hardness perception result of the cutting section, and to determine the optimized cutting head parameters.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the adaptive cutting head optimization method for the tunneling machine as described in any one of claims 1-6.

Images