Underground powerhouse excavation apparatus and control method thereof

Abstract

The application provides an underground powerhouse excavation equipment and a control method thereof. The method comprises the following steps: collecting a main pump pressure signal of the excavation equipment; periodically checking the main pump pressure signal; when the main pump pressure signal meets the periodic checking condition, extracting first working condition characteristic information of the excavation equipment and a main pump pressure period according to the main pump pressure signal; obtaining a working audio signal of the excavation equipment; performing working condition recognition on the working audio signal after periodic correction of the working audio signal based on the main pump pressure period, to obtain second working condition characteristic information of the excavation equipment; performing working condition prediction on the excavation equipment according to the first working condition characteristic information and the second working condition characteristic information; and adjusting the engine speed of the excavation equipment based on the working condition prediction result. The working condition prediction recognition can be performed according to the working audio of the excavation equipment, the speed of the excavation equipment is adjusted based on the preset working condition, and the phased energy-saving control of the excavation equipment is realized.

Description

Technical Field

[0001] The present application relates to the technical field of underground plant excavation, and more particularly, to an underground plant excavation device and a control method thereof. Background Art

[0002] With the rapid development of urban infrastructure, demand for underground powerhouses, a new type of underground engineering that integrates space utilization, safety protection, and functional integration, is constantly increasing. Excavators, as key machinery in underground powerhouse construction, must withstand demanding operating conditions such as complex geotechnical environments, high load fluctuations, and frequent operating mode switching. To achieve safe, efficient, energy-efficient, and environmentally friendly construction goals, energy management and intelligent control of hydraulic excavators have become a research hotspot in recent years.

[0003] Existing hydraulic excavators generally adopt the form of engine-hydraulic pump coordinated drive. The actuator action modes and load power requirements corresponding to different operating stages vary significantly, resulting in the energy consumption characteristics of the equipment showing obvious stages and dynamics. In order to achieve staged energy-saving control, it is usually necessary to reasonably set the engine speed according to the current operating conditions of the equipment and stabilize the working point by adjusting the hydraulic pump displacement in real time. However, the hydraulic system response has an inherent lag. There is an inevitable lag misidentification when identifying the excavator working stage based on the motion information of the actuator or the performance parameters of the hydraulic system such as the main pump pressure, which will further affect the compliance of the whole machine operation and the response speed of the staged energy-saving control, greatly restricting the application and development of excavators. Summary of the Invention

[0004] The present application provides an underground plant excavation device and a control method thereof, which can predict and identify the working condition based on the working audio of the excavation equipment, and adjust the speed of the excavation equipment based on the preset working condition, thereby realizing phased energy-saving control of the excavation equipment.

[0005] In a first aspect, the present application provides a control method for underground plant excavation equipment. The method can be executed by a network device, or can be executed by a chip configured in the network device, and the present application does not limit this.

[0006] Specifically, the method includes:

[0007] When the excavation equipment is excavating the underground powerhouse, the main pump pressure signal of the excavation equipment is collected;

[0008] performing periodic verification on the main pump pressure signal, and when the main pump pressure signal satisfies a periodic verification condition, extracting first operating condition characteristic information of the excavating equipment and a main pump pressure cycle according to the main pump pressure signal;

[0009] Acquiring an operating audio signal of the excavating equipment, performing period correction on the operating audio signal based on the main pump pressure cycle, and then identifying an operating condition to obtain second operating condition characteristic information of the excavating equipment;

[0010] An operating condition prediction is performed on the excavating equipment according to the first operating condition characteristic information and the second operating condition characteristic information, and an engine speed of the excavating equipment is adjusted based on a result of the operating condition prediction.

[0011] In conjunction with the first aspect, in certain implementations of the first aspect, periodically checking the main pump pressure signal specifically includes:

[0012] Acquiring the main pump pressure signal, and performing normalization processing on the main pump pressure signal to obtain a normalized pressure signal;

[0013] Setting a periodic verification interval, calculating multiple autocorrelation function values ​​of the normalized pressure signal within the periodic verification interval, constructing an autocorrelation function model of the normalized pressure signal, and using the maximum function peak of the autocorrelation function model as the periodic verification strength;

[0014] When the periodic verification intensity is higher than a preset periodic verification threshold, it is determined that the main pump pressure signal meets the periodic verification condition.

[0015] In conjunction with the first aspect, in certain implementations of the first aspect, extracting the first operating condition characteristic information and the main pump pressure cycle of the excavating equipment according to the main pump pressure signal specifically includes:

[0016] Obtaining an autocorrelation function model of the main pump pressure signal, and determining a time delay corresponding to a maximum function peak based on the autocorrelation function model, and using the time delay as a main pump pressure cycle;

[0017] The main pump pressure signal is divided into periods according to the main pump pressure period, and an average period signal of the main pump pressure signal is obtained, the average period signal is divided into time windows, an operating condition identification vector is extracted from the main pump pressure data within any time window, and feature clustering is performed using a clustering model to obtain a corresponding operating condition identification result;

[0018] The operating condition identification results corresponding to each time window in the average periodic signal are combined into the first operating condition characteristic information according to the time sequence.

[0019] In conjunction with the first aspect, in certain implementations of the first aspect, performing period correction on the operation audio signal based on the main pump pressure cycle and then performing operating condition identification to obtain the second operating condition characteristic information of the excavating equipment specifically includes:

[0020] Dividing the operation audio signal into periods based on the main pump pressure period to obtain a plurality of audio signal segments;

[0021] After aligning the audio signal segments, period correction is performed based on the audio signal values ​​through point-to-point weighted averaging to obtain the working condition identification audio.

[0022] Dividing the working condition recognition audio into multiple time windows, extracting audio features from any time window and using them for cluster analysis to obtain working condition recognition results corresponding to each time window;

[0023] The working condition recognition results corresponding to each time window in the working condition recognition audio are combined into the second working condition feature information of the mining device according to the time sequence.

[0024] In conjunction with the first aspect, in certain implementations of the first aspect, predicting the operating condition of the excavating equipment based on the first operating condition characteristic information and the second operating condition characteristic information specifically includes:

[0025] Based on the mapping label values ​​corresponding to the respective operating conditions, label values ​​of the first operating condition characteristic information and the second operating condition characteristic information are mapped to obtain a first operating condition label sequence and a second operating condition label sequence;

[0026] Performing audio verification based on the first working condition label sequence and the second working condition label sequence to obtain an operation audio recognition degree;

[0027] When the recognition degree of the operation audio is higher than a preset threshold, historical operation audio is obtained, short-term data prediction is performed based on the operation audio recognition degree and the historical operation audio to obtain predicted audio, working condition identification is performed based on the predicted audio, and the working condition identification result is used as the working condition prediction result.

[0028] In combination with the first aspect, in certain implementations of the first aspect, adjusting the engine speed of the excavator based on the results of the working condition prediction specifically includes: querying the corresponding engine target operating point table according to the working condition prediction results, obtaining the target speed corresponding to the predicted working condition, obtaining the current engine speed, calculating the target speed deviation, and selecting an adjustment strategy according to the deviation range.

[0029] In combination with the first aspect, in certain implementations of the first aspect, the main pump pressure signal of the excavating equipment is collected by a main pump pressure sensor at a preset collection frequency.

[0030] In a second aspect, the present application provides an underground powerhouse excavation device, including a speed control unit, wherein the speed control unit includes:

[0031] The pressure signal acquisition module is used to collect the main pump pressure signal of the excavation equipment when the excavation equipment is excavating the underground powerhouse;

[0032] an operating condition identification module, configured to periodically verify the main pump pressure signal, and extract first operating condition characteristic information and a main pump pressure cycle of the excavating equipment based on the main pump pressure signal when the main pump pressure signal satisfies a periodic verification condition;

[0033] The working condition identification module is further configured to obtain an operating audio signal of the excavating equipment, perform period correction on the operating audio signal based on the main pump pressure cycle, and then perform working condition identification to obtain second working condition characteristic information of the excavating equipment;

[0034] The speed control module is used to predict the working condition of the excavating equipment according to the first working condition characteristic information and the second working condition characteristic information, and adjust the engine speed of the excavating equipment based on the result of the working condition prediction.

[0035] In a third aspect, the present application provides a computer terminal device, which includes a memory and a processor, wherein the memory stores a code, and the processor is configured to obtain the code and execute the above-mentioned control method for underground plant excavation equipment.

[0036] In a fourth aspect, the present application provides a computer-readable storage medium storing at least one computer program, which is loaded and executed by a processor to implement the operations performed by the above-mentioned control method for underground plant excavation equipment.

[0037] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects:

[0038] In an underground factory building excavation equipment and its control method provided by the present application, first, when the excavation equipment is excavating the underground factory building, the main pump pressure signal of the excavation equipment is collected; the main pump pressure signal is periodically checked, and when the main pump pressure signal meets the periodic verification condition, the first working condition characteristic information and the main pump pressure cycle of the excavation equipment are extracted according to the main pump pressure signal; the operating audio signal of the excavation equipment is obtained, and the operating condition is identified after periodic correction of the operating audio signal based on the main pump pressure cycle to obtain the second working condition characteristic information of the excavation equipment; the working condition of the excavation equipment is predicted according to the first working condition characteristic information and the second working condition characteristic information, and the engine speed of the excavation equipment is adjusted based on the result of the working condition prediction.

[0039] Therefore, it can be seen that this application performs working condition identification by collecting the working audio signals generated by the excavation equipment at different working stages. Due to the periodic characteristics of excavation operations, and the influence of noise interference and individual equipment differences on working audio signals, direct identification is prone to stage misalignment or judgment error. This solution performs period correction processing on the audio signal based on the main pump pressure cycle, extracting steady-state working condition features from audio clips of multiple cycles, making the audio recognition process periodically consistent and laying a solid foundation for subsequent working condition identification and prediction. The first working condition feature information extracted from the main pump pressure signal and the second working condition feature information extracted from the working audio signal respectively reflect the operating status of the equipment from different dimensions. The two complementary information constitutes the multimodal working condition identification input. The short-term working condition prediction is further realized based on sequence modeling, providing a feedforward signal for engine control and making a power configuration response in advance. After obtaining the working condition prediction result at the next moment, the system dynamically sets the target speed based on the load demand, target power curve, and engine fuel characteristic curve of the corresponding stage. This realizes phased energy-saving control and intelligent energy management of underground powerhouse excavation equipment, improving fuel economy and responsiveness.

[0040] In summary, the present application can predict and identify the working condition based on the working audio of the excavation equipment, and adjust the speed of the excavation equipment based on the preset working condition, thereby realizing phased energy-saving control of the excavation equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 is an exemplary flow chart of a method for controlling underground powerhouse excavation equipment according to some embodiments of the present application;

[0042] Figure 2 is a schematic structural diagram of a speed control unit according to some embodiments of the present application;

[0043] Figure 3 It is a structural diagram of a computer terminal device for implementing a control method for underground plant excavation equipment according to some embodiments of the present application. DETAILED DESCRIPTION

[0044] The application collects the main pump pressure signal of the excavating equipment; periodically checks the main pump pressure signal; when the main pump pressure signal meets the periodic checking condition, extracts the first working condition characteristic information of the excavating equipment and the main pump pressure period according to the main pump pressure signal; acquires the working audio signal of the excavating equipment, periodically corrects the working audio signal based on the main pump pressure period, and then performs working condition recognition to obtain the second working condition characteristic information of the excavating equipment; performs working condition prediction on the excavating equipment according to the first working condition characteristic information and the second working condition characteristic information, adjusts the engine speed of the excavating equipment based on the result of the working condition prediction, can perform working condition prediction and recognition according to the working audio of the excavating equipment, and adjusts the speed of the excavating equipment based on the preset working condition, thereby realizing the phased energy-saving control of the excavating equipment.

[0045] In order to better understand the above technical solutions, the above technical solutions will be described in detail below in combination with the drawings in the specification and specific embodiments. Reference Figure 1 The figure is an exemplary flow chart of a control method for an underground powerhouse excavating equipment according to some embodiments of the application, which mainly includes the following steps:

[0046] In step S101, when the excavating equipment is excavating the underground powerhouse, the main pump pressure signal of the excavating equipment is collected.

[0047] It should be noted that the main pump pressure signal in the present application is the pressure signal value at the outlet of the main hydraulic pump. Different working condition stages correspond to different actuator starts. Since the load requirements of the actuators on the hydraulic system are different, the collection of the main pump pressure signal can be used for subsequent working condition type identification of the excavating equipment.

[0048] Optionally, in some embodiments, the main pump pressure signal of the excavating equipment can be collected by a main pump pressure sensor at a preset collection frequency. In specific implementation, the pressure collection reserved measurement point of the main pump pressure sensor is located on the main pump outlet pipeline. In some other embodiments, other devices or equipment capable of collecting the main pump pressure signal can also be used, which is not limited in the present application.

[0049] Optionally, in some embodiments, a San Yi SY300 excavator is used as the excavating equipment, and a 4-channel pressure sensor distribution system is used to collect the main pump pressure data every second, which is then transmitted to the equipment control PLC in real time and used for working condition identification.

[0050] In step S102, the main pump pressure signal is periodically checked, and when the main pump pressure signal meets the periodic checking condition, the first working condition characteristic information of the excavating equipment and the main pump pressure period are extracted according to the main pump pressure signal.

[0051] It should be noted that during the excavator's phased earthwork operation, the entire operating cycle can be divided into five typical operating stages: excavation, lifting and rotation, unloading, empty bucket return, and excavation preparation. In each operating stage, the actuators involved in the action are different, and the load conditions vary significantly. This results in clear and distinguishable change characteristics in the pressure signal waveform output by the main pump at each stage. In addition, since the excavator completes loading, transportation, and dumping operations in a fixed order through the coordinated linkage of the actuators, the overall operation process has a strong periodic regularity, which in turn causes the main pump pressure signal to exhibit similar periodic fluctuation characteristics in the time series. That is, after completing a complete operating cycle, the main pump pressure signal forms a set of pressure change waveforms with a repeating pattern on the time axis. Therefore, the main pump pressure signal must be periodically verified to identify the main period of the signal and its stability, and to define the current operating stage of the equipment.

[0052] Optionally, in some embodiments, periodically checking the main pump pressure signal specifically includes:

[0053] Acquiring the main pump pressure signal, and performing normalization processing on the main pump pressure signal to obtain a normalized pressure signal;

[0054] Setting a periodic verification interval, calculating multiple autocorrelation function values ​​of the normalized pressure signal within the periodic verification interval, constructing an autocorrelation function model of the normalized pressure signal, and using the maximum function peak of the autocorrelation function model as the periodic verification strength;

[0055] When the periodic verification intensity is higher than a preset periodic verification threshold, it is determined that the main pump pressure signal meets the periodic verification condition.

[0056] In specific implementation, the normalization processing includes de-averaging and amplitude normalization processing of the original main pump pressure signal, which is used to eliminate the interference of DC offset and signal amplitude differences under different equipment working conditions on periodic judgment, and set a period verification interval. The period verification interval is a delay window set for the main period range that may exist in the normalized pressure signal. For example, in the range of [10S, 30S], the autocorrelation function values ​​of the normalized pressure signal at different delays are calculated in turn to construct an autocorrelation function model of the normalized pressure signal; the autocorrelation function model is obtained by fitting the autocorrelation function values ​​of the signal at different time delays, and the autocorrelation function model is used to characterize the self-similarity of the signal at different time delays, thereby extracting the periodic structure characteristics of the pressure waveform.

[0057] The maximum function peak value within the period verification interval in the autocorrelation function model is used as the target period peak value, and its corresponding function value is defined as the period verification strength; the period verification strength is used to measure whether the main pump pressure signal has a stable and significant periodic pattern in the current time period. When the period verification strength is higher than the preset period verification threshold, it is judged that the main pump pressure signal meets the period verification condition, indicating that the signal has good periodic characteristics and can be used for subsequent working condition feature extraction and filtering identification operations based on period analysis; if it is lower than the threshold, it is considered that the signal is in a non-periodic or abnormal disturbance state, and subsequent periodic analysis and processing are not performed to avoid identification errors or control misadjustments caused by non-periodic states.

[0058] Preferably, in some embodiments, when the main pump pressure signal does not meet the periodic verification conditions, the engine speed of the excavation equipment is adjusted according to the non-periodic speed control scheme. The non-periodic speed control scheme includes setting the default operating parameters for the non-periodic state, performing machine learning based on characteristic indicators such as the real-time average value, slope change trend and fluctuation amplitude of the current main pump pressure signal, determining the engine operating mode, and outputting the target engine speed instruction.

[0059] Optionally, in some embodiments, extracting the first operating condition characteristic information and the main pump pressure cycle of the excavating equipment according to the main pump pressure signal specifically includes:

[0060] Obtaining an autocorrelation function model of the main pump pressure signal, and determining a time delay corresponding to a maximum function peak based on the autocorrelation function model, and using the time delay as a main pump pressure cycle;

[0061] The main pump pressure signal is divided into periods according to the main pump pressure period, and an average period signal of the main pump pressure signal is obtained, the average period signal is divided into time windows, an operating condition identification vector is extracted from the main pump pressure data within any time window, and feature clustering is performed using a clustering model to obtain a corresponding operating condition identification result;

[0062] The operating condition identification results corresponding to each time window in the average periodic signal are combined into the first operating condition characteristic information according to the time sequence.

[0063] In a specific implementation, an autocorrelation function model of the main pump pressure signal is obtained. By detecting the maximum peak value of the model within a positive delay range, a time delay corresponding to the maximum function peak value is determined. The time delay is the main pump pressure period of the main pump pressure signal, which is used to reflect the time scale corresponding to the completion of an operation cycle of the excavating equipment. Based on the main pump pressure period, the main pump pressure signal is period-aligned and period-averaged to obtain an average period signal with periodic representativeness. Furthermore, a time window partitioning scheme is constructed based on the average period signal to divide a complete main cycle into multiple time windows of equal or adaptive lengths. Within each time window, feature quantities of the main pump pressure signal are extracted and a working condition identification vector is constructed. The feature quantities may include the mean, variance, maximum value, slope of change, and peak and trough positions of the pressure value. The working condition identification vector is input into a preset clustering model for multi-class clustering analysis. The clustering model may adopt a K-means clustering model. After identifying the working condition category corresponding to each time window, the working condition identification results corresponding to each time window in the average period signal are sequentially combined according to time sequence to form first working condition feature information reflecting the distribution of working stages within the main cycle.

[0064] The following is a specific example of using the K-means clustering model in this application to perform feature clustering on the working condition identification vector to obtain the corresponding working condition identification result:

[0065] In this example, to identify the operating conditions of the excavation equipment during each operating phase, the main pump pressure signal was collected for 15 complete excavation cycles. Through periodic verification, cycle averaging, and time window partitioning, feature vectors corresponding to five typical operating phases (excavation, hoisting and swinging, unloading, empty bucket return, and excavation preparation) were extracted. 200 samples were extracted for each operating phase, resulting in a total of 1000 feature vectors.

[0066] In order to eliminate the differences in dimensions and numerical ranges between different feature dimensions and improve the clustering effect, the eigenvalues ​​of each dimension in all feature vectors are normalized and standardized to the interval [−1, +1] through linear transformation.

[0067] Using these 1000 sample vectors as input data, we apply the K-means clustering algorithm to perform unsupervised feature clustering. The hyperparameters of the clustering model are set as follows:

[0068] The number of clusters, K, was set to 5, corresponding to five known operating phases. Initial cluster centers were initialized using the K-means++ method. The maximum number of iterations was set to 300, and the convergence tolerance was set to 10^-4. Euclidean distance was used as the distance metric between samples. After clustering, the model automatically divided the 1,000 feature vectors into five categories. Based on the prior correspondence between sample distribution and operating phases, the center vector of each cluster was similarly matched with the mean vector of the operating phase in the annotated sample set to establish a mapping table between cluster categories and actual operating phases.

[0069] To validate the classification effectiveness of the clustering model, 1,000 samples were randomly divided into a training set (800) and a test set (200), maintaining a constant number of 160 and 40 samples per class, respectively. Each vector in the test set was input into the clustering model, and its distance to the cluster center was calculated and assigned a label. The model's recognition accuracy was calculated by comparing the labels of the actual working conditions in the test set. The results showed that the clustering accuracy reached 91.7%, with particularly significant recognition performance during the excavation, unloading, and empty bucket return phases. This demonstrates that the K-means model can effectively extract the structural characteristics of the working conditions implicit in the pressure signals.

[0070] This embodiment shows that the K-means clustering model adopted in this application can effectively divide the excavation operation stages based on the feature vector extracted from the main pump pressure signal without relying on supervised label training. It has the advantages of high computational efficiency, strong adaptability and simple implementation, and is suitable for deployment in on-site edge controllers to realize real-time working condition identification and response control.

[0071] In step S103, an operating audio signal of the excavating equipment is acquired, and an operating condition is identified after period correction is performed on the operating audio signal based on the main pump pressure cycle to obtain second operating condition characteristic information of the excavating equipment.

[0072] It should be noted that this application performs periodic correction on the operating audio signal based on the main pump pressure cycle, which can significantly suppress occasional background noise, operator interference noise, motor screams and other non-periodic components, making the working condition characteristics clearer, the robustness enhanced, and the sound signal more representative and recognizable. The digitization process of the sound signal in this application is the basic link for the mining equipment status identification and working condition analysis in this article. During the digitization operation of the mining equipment sound signal, in order to ensure that the collected sound data has a high signal-to-noise ratio and sufficient time-frequency accuracy, the analog signal needs to be pre-filtered before sampling. This preprocessing link mainly has the following two functions:

[0073] 1) Anti-aliasing filtering: Since actual signals may contain some high-frequency noise components exceeding the Nyquist frequency, direct sampling will cause frequency aliasing, affecting the accuracy of subsequent spectrum analysis. Therefore, a low-pass filter is used to limit the bandwidth of the original analog signal before sampling to suppress frequency components higher than half the sampling frequency.

[0074] 2) Suppressing power supply interference and construction site background noise: Considering the presence of a large number of electromagnetic interference sources (such as engines, cables, transformers, etc.) at the construction site, it is necessary to configure a power supply filtering circuit (such as a 50 / 60 Hz notch filter) and a bandpass filter at the front end of the analog signal acquisition to filter out strong power frequency signal interference and environmental background noise, thereby improving the signal-to-noise ratio of the effective signal.

[0075] The typical acoustic signal frequency range of excavation equipment is mainly concentrated between 1 kHz and 8 kHz, which corresponds to the structural noise and friction noise generated during the operation of actuators such as hydraulic pumps, gear mechanisms, and rotary systems. According to the Nyquist sampling theorem, in order to reconstruct the original analog signal without distortion, the sampling frequency must be higher than twice the highest frequency of the signal. Therefore, this application uses a high-performance microphone array with a sampling frequency of 20 kHz for acoustic signal acquisition, which is far higher than the minimum requirement, ensuring that all effective acoustic information can be captured.

[0076] This system uses an analog / digital converter to digitize analog signals, which includes the following three main steps:

[0077] Sampling: The analog sound signal in continuous time is periodically sampled according to the set sampling frequency to obtain a series of discrete signal points with time intervals;

[0078] Quantization: Mapping the analog amplitude of each sampling point to a finite number of discrete numerical levels, approximating the continuous amplitude to a finite set of discrete values. Quantization accuracy directly affects the ability to restore signal details. This paper uses a 16-bit quantization depth, which offers superior dynamic range and resolution compared to 8-bit and 12-bit schemes. This allows for better preservation of detailed variations in the acoustic signals of mining equipment, meeting the requirements of subsequent complex spectrum modeling and feature extraction.

[0079] Coding: The discrete signal values ​​obtained by sampling and quantization are converted into a standard binary coding format for storage and transmission. In this study, the acoustic signal coding data is uniformly encoded in linear PCM format to ensure lossless expression of the original amplitude changes of the signal, which is conducive to subsequent FFT analysis and time-frequency fusion processing.

[0080] To further improve the collection accuracy and anti-interference ability, the application selects a high-sensitivity microphone chip based on MEMS micro-electro-mechanical system technology, and cooperates with a customized low-noise preamplifier circuit to form a microphone module. The module not only has excellent response frequency band coverage ability and dynamic range control ability, but also has the advantages of small structure, strong anti-vibration ability, and adaptation to harsh site environment, and is especially suitable for deployment in the sound collection scene of heavy excavating equipment. At the same time, the sound signal collection module supports multi-channel synchronous sampling, combined with array layout and time sequence control, can be expanded into a multi-point spatial sound field analysis and sound source positioning system, thereby providing a data basis for operation state evaluation and multi-modal working condition recognition of excavating equipment.

[0081] In some specific embodiments of the application, when the excavating equipment is performing underground plant excavation work, MEMS microphones or acoustic arrays deployed at key structural areas of the equipment, such as engine compartment, hydraulic station, boom base, etc. Real-time collection of work audio signals, sampling frequency is set to 20kHz, quantization accuracy is 16bit.

[0082] Optionally, in some embodiments, the working condition recognition is performed after the work audio signal is corrected based on the main pump pressure cycle, and the second working condition feature information of the excavating equipment includes:

[0083] Based on the main pump pressure cycle, the work audio signal is divided into multiple audio signal segments;

[0084] After aligning each audio signal segment, the cycle correction is realized by point-to-point weighted average based on the audio signal value, and the working condition recognition audio is obtained;

[0085] Based on the working condition recognition audio, multiple time windows are divided, and for any one time window, the audio feature is extracted and used for clustering analysis, and the working condition recognition result corresponding to each time window is obtained;

[0086] According to the time sequence, the working condition recognition results corresponding to each time window in the working condition recognition audio are combined to form the second working condition feature information of the excavating equipment.

[0087] In specific implementation, the main pump pressure cycle is obtained, and the original operation audio signal is segmented on the time axis using this as the cycle length to obtain multiple audio signal segments corresponding to the cycles, each segment representing the complete sound data of the equipment in a certain operation cycle, and the multiple periodic audio signal segments are time-aligned. Specifically, an envelope alignment method can be adopted, which is not limited in this application. On the basis of signal alignment, a point-to-point weighted average is performed on multiple signal segments at the same time point to obtain a period-corrected audio sequence, which is used as the working condition identification audio. The average weight in the point-to-point weighted averaging process can be determined based on the inverse of the deviation ratio between the audio signal value and the average value, and the working condition identification audio is then divided into time windows according to the set length to form multiple continuous non-overlapping time window segments. Among them, in order to facilitate the comparative analysis of working conditions of multi-dimensional data, the time window length of the working condition identification audio can be consistent with the time window length when the average periodic signal of the main pump pressure signal is divided into time windows, and the frequency domain and statistical feature parameters are extracted for the audio signal in each time window. The features may include but are not limited to: spectral centroid, bandwidth, power spectral density, MFCC coefficient, spectral flatness, spectral slope, etc., to construct a working condition feature vector, and then input the audio feature vector of each time window into a trained audio working condition clustering model, such as a K-means clustering model for unsupervised cluster analysis, to distinguish the audio patterns corresponding to different working condition stages, output the working condition identification result corresponding to each time window, and form the second working condition feature information of the mining equipment according to the corresponding time window sequence.

[0088] In this embodiment, a K-means clustering model is used to extract and cluster the audio feature vectors of each time window to obtain the second working condition feature information of the equipment. In actual operation, the system performs period correction and sliding window division on the audio of each operation cycle, extracts the 6-dimensional feature vector of each time window, inputs the K-means model for cluster prediction, and automatically identifies the working condition category corresponding to the time period through the cluster center label mapping table. In some other embodiments, the audio feature vectors of each time window can also be input into a trained neural network model for feature classification to obtain the working condition classification results corresponding to each time window. This application is not limited to this. Specifically, in some specific embodiments of the present application, based on the period information T=16.3s obtained by autocorrelation analysis of the main pump pressure signal, the audio signal is divided into 120 segments according to the period. The working condition audio in each time window is identified, and the first six order data of the Mel frequency cepstral coefficient are extracted to form a 6-dimensional feature vector. A total of 15 operation cycle audio data are collected to generate 1800 valid time window feature vector samples to form a sample space. To train the clustering model, 600 samples are first manually labeled. The working condition category of samples was used as a priori benchmark to identify five typical working conditions: excavation, lifting and rotation, unloading, empty bucket return, and preparation. The corresponding cluster number K=5 was used. Z-score normalization was performed on all sample feature vectors to make the mean of each dimension zero and the standard deviation 1 to eliminate dimensional differences. A maximum of 300 iterations were performed or until the error converged, so that the center point change was less than 10^−4. The Euclidean distance was calculated to assign samples to the nearest center, and the cluster center was updated. The labeled samples in the training set were cross-compared with the clustering results to establish a mapping relationship between cluster labels and actual working conditions. Each cluster center corresponded to a working condition identification result. The identification results of each time window were spliced ​​in chronological order to form an information sequence of complete second working condition characteristic information.

[0089] In step S104, an operating condition prediction is performed on the excavating equipment according to the first operating condition characteristic information and the second operating condition characteristic information, and the engine speed of the excavating equipment is adjusted based on the result of the operating condition prediction.

[0090] It should be noted that when a time window is used to extract the main pump pressure signal within a time range as the signal source, when the signal amplitude varies widely, such as during phase transitions, the feature vector cannot promptly reflect the change in the working phase, resulting in a high rate of hysteresis misidentification and difficulty in correction. The hydraulic system response has inherent hysteresis. Identifying the excavator's operating phase based on actuator motion information or hydraulic system performance parameters (such as main pump pressure) is inevitably subject to hysteresis misidentification, which further affects the operational compliance of the entire machine and the response speed and effectiveness of the staged energy-saving control. Compared to physical signals such as hydraulic pressure, operating audio, as an external manifestation, usually changes synchronously with the actual operating state of the actuator, or even before the hydraulic response. This application introduces an audio verification mechanism to dynamically evaluate the consistency of multi-source recognition results. When the operating audio recognition degree of the identified operating audio is high, it combines historical operating audio for short-term predictive modeling and early identification of future operating phases. This not only effectively solves the hysteresis misidentification problem in main pump pressure identification, but also significantly improves the foresight of the entire machine's operating condition judgment and the response speed of energy-saving control, thereby achieving efficient and intelligent collaborative operation control.

[0091] Optionally, in some embodiments, performing operating condition prediction on the mining equipment according to the first operating condition characteristic information and the second operating condition characteristic information specifically includes:

[0092] Based on the mapping label values ​​corresponding to the respective operating conditions, label values ​​of the first operating condition characteristic information and the second operating condition characteristic information are mapped to obtain a first operating condition label sequence and a second operating condition label sequence;

[0093] Performing audio verification based on the first working condition label sequence and the second working condition label sequence to obtain an operation audio recognition degree;

[0094] When the recognition degree of the operation audio is higher than a preset threshold, historical operation audio is obtained, short-term data prediction is performed based on the operation audio recognition degree and the historical operation audio to obtain predicted audio, working condition identification is performed based on the predicted audio, and the working condition identification result is used as the working condition prediction result.

[0095] In specific implementation, a set of working condition category labels is preset, including multiple working stages such as excavation, lifting and rotation, unloading, empty bucket return and excavation preparation, which correspond to different label values ​​(such as 0 to 4). Based on the mapping relationship, the first working condition feature information and the second working condition feature information are respectively labeled and mapped to obtain the first working condition label sequence and the second working condition label sequence. The above two label sequences are compared moment by moment to obtain the Pearson correlation coefficient between the first working condition label sequence and the second working condition label sequence as the working audio recognition degree. The working audio recognition degree is used to quantify the consistency between the working condition results extracted by different modalities, and then when the working audio recognition When the degree is higher than the preset threshold, the audio feature sequence of the audio clips of the last three operation cycles at the current time point is extracted. The Mel-frequency cepstral coefficient (MFCC) can be extracted as the audio feature sequence and input into the constructed time series prediction model, such as the LSTM time series prediction model or the moving average autoregressive model, to predict the audio feature sequence within a small time window in the future to obtain the predicted audio. The predicted historical sample width can be mapped based on the operation audio recognition degree, and then the predicted audio is subjected to feature extraction and cluster analysis, which is consistent with the real-time audio processing process to obtain the working condition label result within the prediction time window, that is, the working condition prediction result for the future at the current moment.

[0096] In some preferred embodiments of this application, to predict future short-term operating audio of mining equipment, thereby improving the foresight of working condition identification and the intelligence level of engine control strategy, a moving average autoregressive model can be used to model and predict audio features. The audio features are preferably Mel-frequency cepstral coefficients, and the specific implementation method is as follows:

[0097] First, we mapped the audio recognition thresholds for the task, creating a 10-second prediction period. We also set the MFCC sliding window parameters to 25ms frame length and 10ms frame shift. Within this prediction period, we continuously extracted audio frames from the period-corrected audio. For each frame, we extracted the first 13 Mel-frequency cepstral coefficients, forming a 13-dimensional MFCC vector. We then arranged the MFCC vectors for all frames to generate a time-series MFCC feature matrix.

[0098] In some embodiments, other audio features may also be used, such as spectral centroid, frequency band energy, etc.; in this preferred embodiment, MFCC is used as the main modeling feature.

[0099] Select a certain order of MFCC coefficients (e.g., the first-order coefficients) to form a one-dimensional time series, which serves as the feature channel for the prediction object. For example, if 1000 frames of data are collected during a 10-second prediction period, a data sequence of length 1000 is obtained. A time series plot of this MFCC sequence can then be plotted (with frame numbers on the horizontal axis and coefficient values ​​on the vertical axis) to observe its changing trends. If a non-stationary trend is observed (e.g., variance increases over time), exponential smoothing is performed on the sequence to stabilize its variance and eliminate the effects of time-dependent shifts. Furthermore, the autocorrelation function (ACF) and partial autocorrelation function (PACF) plots of the MFCC time series can be plotted to preliminarily determine the appropriate model order range. For example, if the ACF plot significantly decays to near 0 after the fourth order, the autoregressive model can be preliminarily determined to be of order 4; if the PACF plot truncates after the second order, the moving average model can be presumed to be of order 2; and the ARMA model can be preliminarily determined to be of order (4, 2).

[0100] In specific implementations, the least squares method can be used to estimate the parameters of the ARMA model, and the t-test or Ljung-Box test can be used to assess parameter significance. In this example, a test level of 0.05 is used to determine whether each parameter is statistically significant. Subsequently, the Schwarz Bayesian Criterion (BIC) or the Akaike Information Criterion (AIC) is used to compare multiple order combinations, and the optimal order combination (p, q) is selected to construct the final ARMA model.

[0101] By feeding the constructed MFCC time series into the trained ARMA model, the MFCC coefficients for several future frames can be predicted, yielding a predicted audio feature sequence. Specifically, the MFCC vectors for the next 20 to 50 frames can be predicted. Furthermore, the predicted MFCC sequence can be input into the classification model previously described to determine the second operating condition feature information for operating condition identification. This yields the operating condition label for the future time window, enabling short-term audio prediction and early identification of the operating condition.

[0102] Preferably, in some embodiments, when the operation audio recognition degree is lower than a preset threshold, an operating condition characteristic prediction is performed based on the first operating condition characteristic information to obtain an operating condition prediction result. In specific implementation, a moving average autoregressive model can be used to predict the operating condition characteristic of the first operating condition characteristic information.

[0103] Preferably, in some embodiments, adjusting the engine speed of the excavator based on the result of the working condition prediction specifically includes: querying the corresponding engine target operating point table according to the working condition prediction result, obtaining the target speed corresponding to the predicted working condition, obtaining the current engine speed, calculating the target speed deviation, and selecting an adjustment strategy according to the deviation range. When the deviation range is higher than the preset deviation threshold, a rigid speed adjustment strategy based on the maximum acceleration principle is adopted; when the deviation range is lower than the preset deviation threshold, a flexible speed adjustment strategy based on predicted power optimization is adopted.

[0104] In addition, in another aspect of the present application, in some embodiments, the present application provides an underground plant excavation device, the system includes a speed control unit, reference Figure 2 , which is a schematic diagram of exemplary hardware and / or software structure of a speed control unit according to some embodiments of the present application. The speed control unit 200 includes: a pressure signal acquisition module 201, a working condition identification module 202, and a speed control module 203, which are described as follows:

[0105] The pressure signal acquisition module 201 is used to collect the main pump pressure signal of the excavation equipment when the excavation equipment is excavating the underground powerhouse;

[0106] an operating condition identification module 202 for periodically verifying the main pump pressure signal and extracting first operating condition characteristic information and a main pump pressure cycle of the excavating equipment based on the main pump pressure signal when the main pump pressure signal satisfies a periodic verification condition;

[0107] The working condition identification module 202 is further configured to obtain an operating audio signal of the excavating equipment, perform period correction on the operating audio signal based on the main pump pressure cycle, and then perform working condition identification to obtain second working condition characteristic information of the excavating equipment;

[0108] The speed control module 203 is configured to predict the working condition of the excavating equipment according to the first working condition characteristic information and the second working condition characteristic information, and adjust the engine speed of the excavating equipment based on the result of the working condition prediction.

[0109] The above describes in detail an example of an underground plant excavation device and a control method thereof provided in an embodiment of the present application. It can be understood that in order to realize the above functions, the corresponding device includes a hardware structure and / or software module corresponding to executing each function.

[0110] Those skilled in the art should easily realize that, in combination with the units and algorithm steps of each example described in the embodiments disclosed herein, the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a certain function in the application is executed in hardware or in a computer software-driven hardware manner depends on the specific application and design constraints of the technical solution. Therefore, professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0111] In addition, the present application also provides a computer terminal device, which includes a memory and a processor, the memory stores code, and the processor is configured to obtain the code and execute the above-mentioned control method for underground plant excavation equipment.

[0112] In some embodiments, reference Figure 3 , which is a schematic diagram of the structure of a computer terminal device for implementing a control method for underground powerhouse excavation equipment according to some embodiments of the present application. The control method for underground powerhouse excavation equipment in the above embodiment can be achieved by Figure 3 The computer terminal device 300 shown in FIG. 1 is implemented as shown in FIG. 1 , and the computer terminal device 300 includes at least one communication bus 301 , a communication interface 302 , a processor 303 and a memory 304 .

[0113] The processor 303 can be a general-purpose central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more processors for controlling the execution of a control method for an underground plant excavation device in the present application.

[0114] The communication bus 301 may include a path for transmitting information between the aforementioned components.

[0115] Memory 304 may be, but is not limited to, a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, a random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disk storage, an optical disk storage (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), a magnetic disk or other magnetic storage device, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer. Memory 304 may be independent and connected to processor 303 via communication bus 301. Memory 304 may also be integrated with processor 303.

[0116] Memory 304 is used to store program code for executing the solution of the present application, and is controlled by processor 303 for execution. Processor 303 is used to execute the program code stored in memory 304. The program code may include one or more software modules. Determination of the first operating condition characteristic information in the above embodiment can be implemented by processor 303 and one or more software modules in the program code in memory 304.

[0117] The communication interface 302 uses any transceiver or other device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area network (WLAN), etc.

[0118] Optionally, the computer terminal device 300 may further include a power supply 305 for providing power to various devices or circuits in the real-time computer terminal device.

[0119] In a specific implementation, as an example, a computer terminal device may include multiple processors, each of which may be a single-core (single-CPU) processor or a multi-core (multi-CPU) processor. A processor herein may refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0120] The aforementioned computer terminal device can be a general-purpose computer terminal device or a dedicated computer terminal device. In a specific implementation, the computer terminal device can be a desktop computer, a portable computer, a network server, a personal digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. The embodiments of this application do not limit the type of computer terminal device.

[0121] In addition, other aspects of the present application further provide a computer-readable storage medium, which stores at least one computer program, and the computer program is loaded and executed by a processor to implement the operations performed by the above-mentioned control method for underground plant excavation equipment.

[0122] In summary, in an underground factory excavation equipment and a control method thereof disclosed in an embodiment of the present application, first, a main pump pressure signal of the excavation equipment is collected; the main pump pressure signal is periodically checked, and when the main pump pressure signal meets the periodic verification condition, the first working condition characteristic information and the main pump pressure cycle of the excavation equipment are extracted according to the main pump pressure signal; the operating audio signal of the excavation equipment is obtained, and the operating condition is identified after periodic correction of the operating audio signal based on the main pump pressure cycle to obtain the second working condition characteristic information of the excavation equipment; the working condition of the excavation equipment is predicted based on the first working condition characteristic information and the second working condition characteristic information, and the engine speed of the excavation equipment is adjusted based on the result of the working condition prediction. The working condition prediction and identification can be performed based on the working audio of the excavation equipment, and the speed of the excavation equipment can be adjusted based on the preset working condition, thereby realizing staged energy-saving control of the excavation equipment.

[0123] The above description is merely an embodiment of the present application. Common knowledge such as the specific technical solutions or features of the solutions is not described in detail herein. It should be noted that those skilled in the art may make various modifications and improvements without departing from the technical solution of the present application, and these modifications and improvements should also be considered within the scope of protection of the present application. These modifications and improvements will not affect the effectiveness of the implementation of the present application or the practical application of the patent.

Claims

1. A control method for underground powerhouse excavation equipment, characterized in that: include: When the excavation equipment is excavating the underground powerhouse, the main pump pressure signal of the excavation equipment is collected; performing periodic verification on the main pump pressure signal, and when the main pump pressure signal satisfies a periodic verification condition, extracting first operating condition characteristic information of the excavating equipment and a main pump pressure cycle according to the main pump pressure signal; Acquiring an operating audio signal of the excavating equipment, performing period correction on the operating audio signal based on the main pump pressure cycle, and then identifying an operating condition to obtain second operating condition characteristic information of the excavating equipment; performing an operating condition prediction on the excavating equipment according to the first operating condition characteristic information and the second operating condition characteristic information, and adjusting an engine speed of the excavating equipment based on the operating condition prediction result; The step of extracting the first operating condition characteristic information and the main pump pressure cycle of the excavating equipment according to the main pump pressure signal specifically includes: Obtaining an autocorrelation function model of the main pump pressure signal, and determining a time delay corresponding to a maximum function peak based on the autocorrelation function model, and using the time delay as a main pump pressure cycle; The main pump pressure signal is divided into periods according to the main pump pressure period, and an average period signal of the main pump pressure signal is obtained, the average period signal is divided into time windows, an operating condition identification vector is extracted from the main pump pressure data within any time window, and feature clustering is performed using a clustering model to obtain a corresponding operating condition identification result; Combining the operating condition identification results corresponding to each time window in the average periodic signal into the first operating condition characteristic information according to the time sequence; The operating audio signal is periodically corrected based on the main pump pressure cycle, and then the operating condition is identified to obtain the second operating condition characteristic information of the excavating equipment, specifically including: Dividing the operation audio signal into periods based on the main pump pressure period to obtain a plurality of audio signal segments; After aligning the audio signal segments, period correction is performed based on the audio signal values ​​through point-to-point weighted averaging to obtain the working condition identification audio. Dividing the working condition recognition audio into multiple time windows, extracting audio features from any time window and using them for cluster analysis to obtain working condition recognition results corresponding to each time window; The working condition recognition results corresponding to each time window in the working condition recognition audio are combined into the second working condition feature information of the mining device according to the time sequence.

2. The method according to claim 1, wherein The periodic verification of the main pump pressure signal specifically includes: Acquiring the main pump pressure signal, and performing normalization processing on the main pump pressure signal to obtain a normalized pressure signal; Setting a periodic verification interval, calculating multiple autocorrelation function values ​​of the normalized pressure signal within the periodic verification interval, constructing an autocorrelation function model of the normalized pressure signal, and using the maximum function peak of the autocorrelation function model as the periodic verification strength; When the periodic verification intensity is higher than a preset periodic verification threshold, it is determined that the main pump pressure signal meets the periodic verification condition.

3. The method according to claim 1, wherein Predicting the working condition of the excavating equipment according to the first working condition characteristic information and the second working condition characteristic information specifically includes: Based on the mapping label values ​​corresponding to the respective operating conditions, label values ​​of the first operating condition characteristic information and the second operating condition characteristic information are mapped to obtain a first operating condition label sequence and a second operating condition label sequence; Performing audio verification based on the first working condition label sequence and the second working condition label sequence to obtain an operation audio recognition degree; When the recognition degree of the operation audio is higher than a preset threshold, historical operation audio is obtained, short-term data prediction is performed based on the operation audio recognition degree and the historical operation audio to obtain predicted audio, working condition identification is performed based on the predicted audio, and the working condition identification result is used as the working condition prediction result.

4. The method according to claim 1, wherein Adjusting the engine speed of the excavator based on the working condition prediction result specifically includes: querying the corresponding engine target operating point table according to the working condition prediction result, obtaining the target speed corresponding to the predicted working condition, obtaining the current engine speed, calculating the target speed deviation, and selecting an adjustment strategy according to the deviation range.

5. The method according to claim 1, wherein The main pump pressure signal of the excavating equipment is collected by the main pump pressure sensor at a preset collection frequency.

6. An underground powerhouse excavation device, comprising a rotation speed control unit, wherein the rotation speed control unit is configured to execute the control method for underground powerhouse excavation device according to any one of claims 1 to 5, characterized in that: The speed control unit includes: The pressure signal acquisition module is used to collect the main pump pressure signal of the excavation equipment when the excavation equipment is excavating the underground powerhouse; an operating condition identification module, configured to periodically verify the main pump pressure signal, and extract first operating condition characteristic information and a main pump pressure cycle of the excavating equipment based on the main pump pressure signal when the main pump pressure signal meets a periodic verification condition; The working condition identification module is further configured to obtain an operating audio signal of the excavating equipment, perform period correction on the operating audio signal based on the main pump pressure cycle, and then perform working condition identification to obtain second working condition characteristic information of the excavating equipment; The speed control module is used to predict the working condition of the excavating equipment according to the first working condition characteristic information and the second working condition characteristic information, and adjust the engine speed of the excavating equipment based on the working condition prediction result.

7. A computer terminal device, characterized in that: The computer terminal device includes a memory and a processor, the memory stores codes, and the processor is configured to obtain the codes and execute the control method for underground plant excavation equipment according to any one of claims 1 to 5.

8. A computer-readable storage medium storing at least one computer program, characterized in that: The computer program is loaded and executed by a processor to implement the operations performed by the control method for underground plant excavation equipment according to any one of claims 1 to 5.

Images