Cantilevered tunneling machine pick wear online monitoring method, system, device and medium

Abstract

The present application belongs to the technical field of pick wear monitoring, and aims to solve the problem of difficult detection of pick wear state and increased regular maintenance cost. A kind of cantilever type roadheader pick wear online monitoring method, system, equipment and medium are provided, comprising the following steps: real-time acquisition of cutting oil cylinder pressure signal, cutting reducer vibration signal and cutting head temperature signal;Feature extraction is carried out on the above sensor data, and the feature weight value is determined based on the correlation between sensors;Based on the correlation between the extracted features and the sensors, the wear condition of the pick is analyzed, and the initial pick wear state is obtained;The above features are corrected based on the double-layer feature extraction model, and the corrected pick wear state is updated;The wear condition of the pick is monitored based on the corrected pick wear state.The present application can real-time master the wear condition of the pick based on the pick online monitoring technology of multi-index information fusion, and carry out intelligent diagnosis and on-demand maintenance on the pick based on the state.

Description

Technical Field

[0001] This invention belongs to the field of cutting tooth wear monitoring technology, specifically relating to an online monitoring method, system, equipment, and medium for the wear of cutting teeth in a cantilever tunneling machine. Background Technology

[0002] As the rock-breaking cutter of a cantilever roadheader, the cutting tooth comes into direct contact with coal and rock under extremely harsh and complex working conditions. It operates under high stress and strong impact for extended periods, bearing complex alternating impact loads. Its power consumption accounts for over 80% of the total machine power. The cutting efficiency, cutting resistance, specific energy consumption, and cutting stability of the roadheader are all closely related to the cutting tooth. During roadheader operation, downtime due to cutting tooth failure accounts for 20% of total operating time, and economic losses account for 30% of total economic losses. Therefore, the reliability, safety, and economy of cutting tooth operation directly determine the cutting performance, service life, and operating costs of the roadheader, thus affecting the tunneling efficiency and economic benefits of coal enterprises.

[0003] During the cutting process, the cutting picks come into direct contact with coal and rock. Because they are encased in coal and rock, it's impossible to directly obtain their wear status using visual or other sensing methods. Furthermore, the small size of the picks prevents the installation of internal sensors. Therefore, mature online monitoring technology for cutting picks is currently lacking. At present, the wear status of the picks is inspected manually, and maintenance methods include either post-failure repair or periodic maintenance. Post-failure repair is a passive and unplanned maintenance activity; that is, when the cutting pick system fails, the failed pick is replaced. Because the timing of failure cannot be predicted and there is a lack of pre-maintenance preparation, it often results in long downtimes for the tunneling machine and serious system failures. Periodic maintenance is a proactive and planned maintenance activity; that is, the picks are inspected regularly, and failed picks are replaced based on their wear status. However, because the maintenance cycle is determined by manual experience, maintenance activities are often too frequent, leading to increased maintenance costs and wasted maintenance resources. Summary of the Invention

[0004] In order to solve at least one of the above-mentioned technical problems in the prior art, the present invention provides a method, system, equipment and medium for online monitoring of wear of cutting teeth of cantilever tunneling machines.

[0005] This invention is achieved using the following technical solution: an online monitoring method for wear of cutting teeth in a cantilever tunneling machine, comprising the following steps:

[0006] S1: Real-time acquisition of sensor data affecting the wear state of the cutting teeth, including but not limited to the pressure signal of the cutting cylinder, the vibration signal of the cutting reducer, and the temperature signal of the cutting head;

[0007] S2: Extract features from the above sensor data and determine feature weight values ​​based on the correlation between the sensors;

[0008] S3: Analyze the wear of the cutting teeth based on the extracted features and the correlation between the sensors to obtain the initial wear state of the cutting teeth;

[0009] S4: Based on the two-layer feature extraction model, the above features are corrected and updated to obtain the corrected cutter wear state; the wear condition of the cutter is monitored based on the corrected cutter wear state.

[0010] Preferably, in S2, the extracted features include:

[0011] The first feature is f1, the second feature is f2, and the third feature is f3, and the expressions for the above features are as follows:

[0012]

[0013] In the formula, P is the pressure sensor data, ε1 is the pressure preset value and ε1 is greater than 0, k1 is the pressure adjustment parameter, and tanh is a hyperbolic tangent function;

[0014]

[0015] In the formula, A is the acceleration sensor data, ε2 is the preset acceleration value and ε2 is greater than 0, k2 is the acceleration adjustment parameter, and sigmoid is an S-shaped function;

[0016]

[0017] In the formula, T is the temperature sensor data, ε3 is the preset temperature value and ε3 is greater than 0, k3 is the temperature adjustment parameter, and relu is a rectified linear unit function.

[0018] Preferably, in S2, the feature weight values ​​obtained based on the correlation between the sensors include:

[0019] The first feature weight value w1, the second feature weight value w2, and the third feature weight value w3 are given by the following expressions:

[0020]

[0021] In the formula, corr(X,Y) represents the correlation between sensor X and sensor Y, P is the pressure sensor, T is the temperature sensor, and A is the acceleration sensor.

[0022] Preferably, in S3, the formula for the initial wear state of the cutting teeth is:

[0023] Cutting tooth wear condition = (w1*f1 + w2*f2 + w3*f3) / (w1 + w2 + w3)

[0024] Preferably, in S4, the step of modifying the features includes:

[0025] Set the first feature f1, the second feature f2, and the third feature f3 as the input sequence;

[0026] A two-layer feature extraction model is used to extract features from the input sequence and output the corrected tooth wear state.

[0027] The modified formula for the wear state of the cutting teeth is:

[0028]

[0029] In the formula, y1, y2, and y3 represent the wear state sequence of the cutting teeth calculated by the two-layer feature extraction model.

[0030] The present invention also provides an online monitoring system for wear of cutting teeth of a cantilever tunneling machine, including a sensor module, a data acquisition module, a data analysis module and a monitoring module;

[0031] The sensor module includes a pressure sensor, an acceleration sensor, and a temperature sensor. The pressure sensor is installed on the oil port of the multi-way directional valve of the hydraulic system of the cantilever tunneling machine, which is used to indirectly monitor the force on the cutting teeth. The acceleration sensor is installed on the cutting reducer of the cantilever tunneling machine, which is used to monitor the vibration of the cutting teeth. The temperature sensor is installed inside the cutting head of the cantilever tunneling machine, which is used to monitor the temperature change of the cutting teeth.

[0032] The data acquisition module is used to receive data sent by the sensor module and transmit it to the data analysis module;

[0033] The data analysis module is used to extract features from sensor data, determine the correlation between sensors, and analyze the wear of the cutting teeth based on the sensor data and the correlation between sensors to obtain the initial wear state of the cutting teeth; at the same time, the extracted features are optimized through a two-layer feature extraction model to update and obtain the corrected wear state of the cutting teeth.

[0034] The monitoring module is used to acquire the cutting tooth status information sent by the data analysis module in real time and monitor the wear of the cutting tooth.

[0035] Preferably, the data analysis module includes a preprocessing module, an extraction module, and a state feature determination module;

[0036] The preprocessing module is used to preprocess the sensor data to obtain intermediate data;

[0037] The extraction module is used to extract state features from intermediate data to describe the wear state of the cutting teeth;

[0038] The status characteristic determination module is used to determine the characteristics of the pressure sensor, acceleration sensor, and temperature sensor, and to comprehensively evaluate the wear condition of the cutting tooth based on the above characteristics, thereby determining the status information of the cutting tooth.

[0039] Preferably, in the input sequence X = {x1, x2, x3} of the two-layer feature extraction model, x1, x2, and x3 are the first feature f1, the second feature f2, and the third feature f3, respectively.

[0040] The formula for feature extraction in the two-layer feature extraction model is as follows:

[0041] h 1,t ,c 1,t =LSTM1(x t ,h 1,t-1 ,c 1,t-1 )

[0042] h 2,t ,c 2,t =LSTM2(h 1,t ,h 2,t-1 ,c 2,t-1 )

[0043] In the formula, h 1,t and c 1,t h represents the hidden state and cell state of the first layer of the two-layer feature extraction model at time step t, respectively. 2,t and c 2,t These represent the hidden state and cell state of the second layer of the two-layer feature extraction model at time step t, respectively; h 1,t-1 and c 1,t-1 These represent the hidden state and cell state of the first layer of the two-layer feature extraction model at time step t-1, respectively; h 2,t-1 and c 2,t-1 These represent the hidden state and cell state of the second layer of the two-layer feature extraction model at time step t-1, respectively; x t X = {x1, x2, x3} represents the input sequence of the first layer of the two-layer feature extraction model at time step t; LSTM1 represents the first layer of the Long Short-Term Memory network model; LSTM2 represents the second layer of the Long Short-Term Memory network model.

[0044] The output of the two-layer feature extraction model is:

[0045] y t =g(W hy h 2,t +b y )

[0046] In the formula, g is the activation function, and Why and b y These are the weight matrix and bias terms of the output layer, y t As a correction of the wear state of the cutting teeth.

[0047] Compared with the prior art, the beneficial effects of the present invention are:

[0048] This invention, based on multi-index information fusion, provides online monitoring technology for cutting teeth, enabling real-time monitoring of tooth wear and allowing for condition-based intelligent diagnosis and condition-based maintenance. This ensures cutting performance and guarantees the safe, efficient, and economical operation of the tunneling machine.

[0049] Through the sensor module and data acquisition module, the system monitors the stress, vibration, and temperature changes of the cutting teeth in real time, enabling real-time monitoring of the wear condition of the cutting teeth. The data analysis module analyzes the sensor data to accurately assess the wear condition of the cutting teeth, providing precise cutting tooth status information, which helps to take timely maintenance measures and extend the service life of the tunneling machine's cutting head. By acquiring the cutting tooth status information in real time through the monitoring module, the wear condition of the cutting teeth can be detected in a timely manner, avoiding safety hazards caused by excessive wear of the cutting teeth and improving the operational safety of the tunneling machine.

[0050] By using a two-layer feature extraction model, the complex relationships between input features can be better captured, and more representative feature representations can be extracted, which helps to improve the accuracy of predicting the wear state of the cutting teeth. Attached Figure Description

[0051] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0052] Figure 1 This is a schematic diagram of a cantilever tunneling machine cutter tooth wear monitoring system module provided in an embodiment of the present invention;

[0053] Figure 2 This is a schematic diagram of the method for monitoring the wear of cutting teeth of a cantilever tunneling machine provided in an embodiment of the present invention;

[0054] Figure 3 This is a schematic diagram of the sensor installation position in the cantilever tunneling machine cutter tooth wear monitoring system provided in an embodiment of the present invention;

[0055] Figure 4 This is a schematic diagram of the pressure sensor installation position in the cantilever tunneling machine cutter wear monitoring system provided in an embodiment of the present invention;

[0056] Figure 5This is a schematic diagram of the installation position of the acceleration sensor in the cantilever tunneling machine cutter wear monitoring system provided in an embodiment of the present invention;

[0057] Figure 6 This is a schematic diagram of the installation position of the temperature sensor in the cantilever tunneling machine cutter wear monitoring system provided in an embodiment of the present invention.

[0058] In the diagram: 1-Cutting lifting and cutting rotation; 2-Cutting reducer; 3-Cutting head; 4-Pressure sensor; 5-Acceleration sensor; 6-Temperature sensor. Detailed Implementation

[0059] The technical solutions of the embodiments of the present invention will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other implementation methods obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0060] It should be noted that the structures, proportions, sizes, etc., shown in the accompanying drawings of this specification are only for the purpose of assisting those skilled in the art in understanding and reading the content disclosed in the specification, and are not intended to limit the conditions under which the present invention can be implemented. Therefore, they have no substantial technical significance. Any modifications to the structure, changes in the proportional relationships, or adjustments to the size, without affecting the effects and objectives that the present invention can produce, should fall within the scope of the technical content disclosed in the present invention. It should be noted that in this specification, relational terms such as "first" and "second" are only used to distinguish one entity from several other entities, and do not necessarily require or imply any actual relationship or order between these entities.

[0061] This invention provides an embodiment:

[0062] like Figures 1 to 6 As shown, an online monitoring system for wear of cutting teeth in a cantilever tunneling machine includes a sensor module, a data acquisition module, a data analysis module, and a monitoring module.

[0063] The sensor module includes a pressure sensor, an acceleration sensor, and a temperature sensor. The pressure sensor is installed on the oil port of the multi-way directional valve of the hydraulic system of the cantilever tunneling machine, which is used to indirectly monitor the force on the cutting teeth. The acceleration sensor is installed on the cutting reducer of the cantilever tunneling machine, which is used to monitor the vibration of the cutting teeth. The temperature sensor is installed inside the cutting head of the cantilever tunneling machine, which is used to monitor the temperature change of the cutting teeth.

[0064] The data acquisition module is used to receive data sent by the sensor module and transmit it to the data analysis module;

[0065] The data analysis module is used to extract features from sensor data, determine the correlation between sensors, and analyze the wear of the cutting teeth based on the sensor data and the correlation between sensors to obtain the initial wear state of the cutting teeth; at the same time, the extracted features are optimized through a two-layer feature extraction model to update and obtain the corrected wear state of the cutting teeth.

[0066] The monitoring module is used to acquire the cutting tooth status information sent by the data analysis module in real time and monitor the wear of the cutting teeth. The monitoring module can provide real-time cutting tooth status information to the equipment operator through the display screen, alarm, remote communication and other means so that maintenance measures can be taken in a timely manner.

[0067] In the past, the failure of cutting teeth could be directly determined by visual inspection. Cutting tooth failure is categorized into soft failure (wear-related) and hard failure (impact-related). Soft failure occurs when the wear height of the cutting tooth reaches the soft failure threshold, which can be determined by measuring the tooth height with calipers. Hard failure occurs when the cutting tooth chipps or breaks, which can be determined by observing its shape. Timely replacement of severely worn cutting teeth ensures the normal operation of the tunneling machine and extends the service life of the cutting head. Since the cutting teeth are encased in coal and rock during operation, direct monitoring of wear using visual or other sensing methods is impossible. Therefore, using physical quantities that indirectly reflect the wear state of the cutting teeth for multi-indicator information fusion to comprehensively evaluate the wear state has become a key technology for online cutting tooth monitoring.

[0068] The pressure sensor can be installed on the oil port of the multi-way directional valve in the hydraulic system of the cantilever tunneling machine, specifically on the cutting lifting and slewing links, to indirectly monitor the stress on the cutting teeth. During the cutting process, the degree of wear on the cutting teeth can be inferred by monitoring the pressure changes in the cutting lifting and slewing cylinders. Higher cylinder pressure indicates harder coal and rock being cut, leading to faster tooth wear. As tooth wear intensifies, the cylinder pressure will also change accordingly.

[0069] Accelerometers can be installed on the cutting reducer of a cantilever roadheader to monitor the vibration of the cutting teeth. During the cutting process, by monitoring the vibration changes of the cutting reducer, the degree of wear on the cutting teeth can be inferred. A larger vibration amplitude indicates harder coal and rock being cut, and faster tooth wear. As the wear of the cutting teeth intensifies, the vibration of the cutting reducer will also change accordingly.

[0070] Temperature sensors can be installed inside the cutting head of a cantilever roadheader to monitor temperature changes in the cutting teeth. During the cutting process, monitoring the temperature changes of the cutting head allows for the assessment of the degree of wear on the cutting teeth. Higher temperatures indicate harder coal and rock being cut, leading to faster tooth wear. As tooth wear intensifies, the temperature of the cutting head will also change accordingly.

[0071] Pressure sensors, temperature sensors, and acceleration sensors should be installed on the tunneling machine at locations where feedback signals from the cutting teeth can be collected. The specific installation locations can be determined based on the actual situation, and this embodiment of the invention does not impose specific limitations on this. Ultimately, the wear state of the cutting teeth is indirectly and comprehensively evaluated using deep learning technology based on multi-index information fusion based on the signals fed back by the above three sensors.

[0072] The data analysis module includes a preprocessing module, an extraction module, and a state feature determination module;

[0073] The preprocessing module is used to preprocess sensor data to obtain intermediate data; it improves data quality and accuracy. For example, it can perform operations such as filtering, noise reduction, and data correction to remove noise and eliminate interference, making subsequent data analysis more reliable and accurate.

[0074] The extraction module is used to extract state features describing the wear state of the cutting teeth from the preprocessed intermediate data. These features can be statistical features, frequency domain features, time domain features, etc., and can be extracted using signal processing methods, time series analysis, spectrum analysis, and other techniques. For example, features such as the average, maximum, and minimum values ​​of the force on the cutting teeth, and the peak value of the vibration frequency can be extracted.

[0075] Taking pressure sensor data processing as an example: when the cutting teeth are severely worn, the pressure sensor data may show a larger fluctuation range or a faster rate of change. By processing the data, this characteristic can be better captured, thereby timely detection and prediction of the wear of the cutting teeth.

[0076] Assuming we use a standard linear normalization method, which scales the pressure sensor data P to the range of 0 to 1, the calculation formula is:

[0077]

[0078] Substitute the data above into the formula to calculate:

[0079]

[0080] After ordinary linear normalization, the normalized pressure sensor data is 0.6.

[0081] Now, let's assume that no normalization has been performed and the raw pressure sensor data is used directly for analysis. In this case, if the data range is very large, comparison and analysis between the data may become difficult because the numerical differences between different sensor data are significant, making it hard to intuitively compare the trends and characteristics of the data.

[0082] Therefore, by performing nonlinear normalization processing on pressure sensor data, such as logarithmic function transformation, the data can be mapped to a more suitable range, better reflecting the changing characteristics and trends of the data, and improving the accuracy and reliability of data processing. By combining nonlinear transformation and ordinary normalization processing, the following improvements are achieved in the field of cantilever tunneling machine cutter wear monitoring technology:

[0083] Better capture of data characteristics: Nonlinear transformations can better capture data trends and fluctuations, while ordinary normalization can scale data to a suitable range. Combining these two methods allows for a more comprehensive analysis of data characteristics. Improved data processing accuracy: By combining these two processing methods, the changes in data can be reflected more accurately, improving the accuracy and reliability of data processing. More comprehensive monitoring: Combining nonlinear and linear normalization processing allows for more comprehensive monitoring of the wear condition of cantilever tunneling machine cutting teeth, improving the comprehensiveness and effectiveness of wear monitoring.

[0084] Pressure sensor data is extracted from intermediate data, and a nonlinear transformation is performed on the pressure sensor data to capture its changing and fluctuating trends, thereby obtaining pressure sensor characteristics. The calculation expression for the pressure sensor characteristics is as follows:

[0085]

[0086] Where P represents the pressure sensor data, P normalized P represents the normalized characteristics of the pressure sensor. min P represents the minimum value of the pressure sensor data. max This indicates the maximum value of the pressure sensor data.

[0087] Pressure sensor data is extracted from intermediate data, and a nonlinear transformation is performed on the pressure sensor data to capture its changing and fluctuating trends, thereby obtaining pressure sensor characteristics. The calculation expression for the pressure sensor characteristics is as follows:

[0088]

[0089] Where P represents the pressure sensor data, P normalized P represents the normalized characteristics of the pressure sensor. min P represents the minimum value of the pressure sensor data. maxThis indicates the maximum value of the pressure sensor data;

[0090] Suppose we have the following data:

[0091] P = 80, P min =50, P max =100;

[0092] Substitute these values ​​into the formula to perform the calculation:

[0093] Therefore, after calculation, the normalized pressure sensor data P was obtained. normalized ≈1.69.

[0094] Processing pressure sensor data in this way has the following advantages:

[0095] Nonlinear Transformation: The nonlinear transformation of the logarithmic function can better capture the changing trends and fluctuation characteristics of data, improving the expressive power of data processing and the accuracy of feature extraction. Strong Adaptability: By introducing data processing formulas, it is applicable to different data distributions and characteristics, improving the flexibility and adaptability of data processing. Improved Accuracy: It can more accurately capture the characteristics and changes of pressure sensor data in the wear monitoring of cantilever tunneling machine cutting teeth, improving monitoring effectiveness and prediction accuracy.

[0096] The state characteristic determination module is used to determine the characteristics of the pressure sensor, acceleration sensor, and temperature sensor, and based on these characteristics, comprehensively evaluate the wear condition of the cutting tooth to determine its state information. For example, by analyzing the changing trends and amplitudes of the pressure sensor data, the stress condition of the cutting tooth can be determined; by analyzing the vibration frequency and peak values ​​of the acceleration sensor data, the vibration condition of the cutting tooth can be evaluated; and by analyzing the changing trends and peak values ​​of the temperature sensor data, the temperature changes of the cutting tooth can be monitored. The data analysis module can use machine learning algorithms and statistical analysis methods for data processing and wear condition prediction.

[0097] The sensor data is processed by a preprocessing module, and the feature extraction module extracts the wear condition characteristics of the cutting teeth from the intermediate data. The feature determination module then uses these characteristics to determine the wear condition information of the cutting teeth. This improves the accuracy and reliability of the system's assessment of the cutting teeth's wear condition, helping users better understand the usage of the cutting teeth and make the most appropriate maintenance decisions.

[0098] The condition information of the cutting tooth is determined based on the characteristics of pressure, acceleration, and temperature sensors. Based on these sensor characteristics, the wear state of the cutting tooth is comprehensively evaluated. A comprehensive evaluation model can be established, weighted and combined to obtain information such as the overall wear degree and remaining lifespan of the cutting tooth. For example, machine learning algorithms and fuzzy logic methods can be used to achieve this comprehensive evaluation.

[0099] Through the above steps, the condition characteristic determination module can determine the characteristics of the pressure sensor, acceleration sensor, and temperature sensor based on the condition characteristics, and comprehensively evaluate the wear condition information of the cutting tooth based on these characteristics. This condition characteristic determination module can provide accurate and reliable cutting tooth condition information, helping users to understand the usage status of the cutting tooth in a timely manner and make maintenance and replacement decisions.

[0100] A second aspect of the present invention provides a method for online monitoring of wear on cutting teeth of a cantilever tunneling machine, comprising the following steps:

[0101] S1: Real-time acquisition of sensor data affecting the wear state of the cutting teeth, including but not limited to the pressure signal of the cutting cylinder, the vibration signal of the cutting reducer, and the temperature signal of the cutting head;

[0102] S2: Extract features from the above sensor data and determine feature weight values ​​based on the correlation between the sensors;

[0103] S3: Analyze the wear of the cutting teeth based on the extracted features and the correlation between the sensors to obtain the initial wear state of the cutting teeth;

[0104] S4: Based on the two-layer feature extraction model, the above features are corrected and updated to obtain the corrected cutter wear state; the wear condition of the cutter is monitored based on the corrected cutter wear state.

[0105] In S2, the extracted features include:

[0106] The first feature is f1, the second feature is f2, and the third feature is f3, and the expressions for the above features are as follows:

[0107]

[0108] In the formula, P is the pressure sensor data, ε1 is the pressure preset value and ε1 is greater than 0, k1 is the pressure adjustment parameter, and tanh is a hyperbolic tangent function;

[0109]

[0110] In the formula, A is the acceleration sensor data, ε2 is the preset acceleration value and ε2 is greater than 0, k2 is the acceleration adjustment parameter, and sigmoid is an S-shaped function;

[0111]

[0112] In the formula, T is the temperature sensor data, ε3 is the preset temperature value and ε3 is greater than 0, k3 is the temperature adjustment parameter, and relu is a rectified linear unit function.

[0113] Design principles:

[0114] For the first feature f1, logarithmic operations are used to convert the range of pressure sensor data to a more suitable scale, and the data with a large range of variation is compressed by the tanh function to prevent the excessive range of variation from having too much impact on the calculation of wear state.

[0115] For the second feature f2, the square root operation is used to convert the range of change of the accelerometer data into a more suitable scale, and the sigmoid function is used to compress the data with a large range of change, so that the feature with a large range of change will not occupy too much weight in the calculation.

[0116] For the third feature f3, the range of temperature sensor data is converted to a more suitable scale using exponential operations, and the data with a large range of variation is compressed using the ReLU function, so that features with a large range of variation do not occupy too much weight in the calculation.

[0117] Based on the above design principles, we can more accurately consider the different characteristics of sensor data and assign corresponding weights according to the importance and contribution of the characteristics, thereby calculating a more accurate wear state of the cutting teeth.

[0118] The expressions for the first feature weight value w1, the second feature weight value w2, and the third feature weight value w3, and their correlation with the sensor, can be selected and determined based on the following design principles and strategies:

[0119] Feature contribution-based: Weight values ​​are determined based on the contribution of different sensors to the wear state. If the data from a particular sensor plays a more significant role in predicting the wear state, its corresponding feature weight value can be increased accordingly.

[0120] Based on the correlation between features: Considering the inherent correlation between sensors, the correlation between different sensors is reflected by setting weight values. If the data from different sensors have a high correlation in wear condition prediction, then their corresponding feature weight values ​​can be increased accordingly.

[0121] The specific expressions for the first feature weight value w1, the second feature weight value w2, and the third feature weight value w3, and their correlation with the sensor, can be determined based on the design principles and strategies described above. These expressions can be adjusted and optimized according to actual application scenarios and data analysis results to achieve more accurate and effective wear condition prediction.

[0122] The expressions for the above feature weights are as follows:

[0123]

[0124] In the formula, corr(X,Y) represents the correlation between sensor X and sensor Y, which reflects the intrinsic relationship between sensors by calculating the correlation between each sensor; P is a pressure sensor, T is a temperature sensor, and A is an acceleration sensor.

[0125] Design rationale:

[0126] The design rationale for the first feature weight value w1 is as follows: considering the correlation between pressure sensor P and temperature sensor T and acceleration sensor A, as well as the correlation between temperature sensor T and acceleration sensor A, the weight value corresponding to the sensor with higher correlation is increased by summing and normalizing the two correlations to emphasize their contribution to the wear state.

[0127] The design rationale for the second feature weight value w2 is as follows: considering the correlation between the accelerometer A and the pressure sensor P and the temperature sensor T, as well as the correlation between the temperature sensor T and the pressure sensor P, the weight value corresponding to the sensor with higher correlation is increased by summing and normalizing the two correlations to emphasize its contribution to the wear state.

[0128] The design rationale for the third feature weight value w3 is as follows: considering the correlation between the temperature sensor T and the pressure sensor P and the acceleration sensor A, as well as the correlation between the pressure sensor P and the acceleration sensor A, the weight value corresponding to the sensor with higher correlation is increased by summing and normalizing the two correlations to emphasize their contribution to the wear state.

[0129] The specific expressions for the first feature weight value w1, the second feature weight value w2, and the third feature weight value w3 can be adjusted and determined based on the actual application scenario and the results of data analysis.

[0130] In S3, the formula for the initial wear state of the cutting teeth is:

[0131] Cutting tooth wear condition = (w1*f1 + w2*f2 + w3*f3) / (w1 + w2 + w3)

[0132] To improve the accuracy of the cutter wear state, a two-layer feature extraction model is added to correct the first feature f1, the second feature f2, and the third feature f3, thereby updating the cutter wear state formula. The feature correction steps in S4 include:

[0133] Set the first feature f1, the second feature f2, and the third feature f3 as the input sequence X = {x1, x2, x3};

[0134] A two-layer feature extraction model is used to extract features from the input sequence. The calculation formula for the extracted features in the two-layer feature extraction model is as follows:

[0135] h 1,t ,c 1,t =LSTM1(x t ,h 1,t-1 ,c 1,t-1 )

[0136] h 2,t ,c 2,t =LSTM2(h 1,t ,h 2,t-1 ,c 2,t-1 )

[0137] In the formula, h 1,t and c 1,t h represents the hidden state and cell state of the first layer of the two-layer feature extraction model at time step t, respectively. 2,t and c 2,t These represent the hidden state and cell state of the second layer of the two-layer feature extraction model at time step t, respectively; h 1,t-1 and c 1,t-1 These represent the hidden state and cell state of the first layer of the two-layer feature extraction model at time step t-1, respectively; h 2,t-1 and c 2,t-1 These represent the hidden state and cell state of the second layer of the two-layer feature extraction model at time step t-1, respectively; x t X = {x1, x2, x3} represents the input sequence of the first layer of the two-layer feature extraction model at time step t; LSTM1 represents the first layer of the Long Short-Term Memory network model; LSTM2 represents the second layer of the Long Short-Term Memory network model.

[0138] t-1 represents the previous time step, and LSTM stands for Long Short-Term Memory Network Model, used to process sequential data, such as time series analysis or natural language processing tasks. LSTM is a special type of recurrent neural network (RNN) designed to avoid long-term dependency problems by controlling the flow of information through gating mechanisms.

[0139] It also outputs the corrected wear status of the cutting teeth; y t =g(Why h 2,t +b y )

[0140] In the formula, g is the activation function, and W hy and b y These are the weight matrix and bias terms of the output layer, y t As a correction of the wear state of the cutting teeth.

[0141] The modified formula for the wear state of the cutting teeth is:

[0142]

[0143] In the formula, y1, y2, and y3 represent the wear state sequence of the cutting teeth calculated by the two-layer feature extraction model.

[0144] By employing the above steps, the wear state of the cutting tooth can be calculated by comprehensively utilizing the preset feature extraction model and the formula for calculating the wear state of the cutting tooth, thereby improving the accuracy and reliability of predicting the wear state of the cutting tooth. The above method effectively combines the feature extraction model and the wear state calculation formula, thus enabling a better understanding and prediction of the wear state of the cutting tooth.

[0145] Combining a pre-defined feature extraction model with a cutter wear state formula offers the following benefits: More accurate prediction: By using a two-layer feature extraction model, the complex relationships between input features can be better captured, and more representative feature representations can be extracted. This helps improve the accuracy of predicting the cutter wear state. More flexible model: Compared to simple linear formulas, the two-layer feature extraction model has stronger expressive power and flexibility. By abstracting and extracting features at multiple levels, it can better adapt to different types of data and feature variations. Better generalization ability: The two-layer feature extraction model can learn more abstract and high-level representations of input features, thereby improving the model's generalization ability and enabling it to perform well even on unseen data.

[0146] A concrete example could be predicting the wear state of a tunnel boring machine's cutting teeth. Using accelerometer data and three features (f1, f2, and f3) as input, a two-layer feature extraction model is used to extract the hidden state h representing the wear state of the cutting teeth. 2,t Then, the hidden state h 2,t Substituting the values ​​into the calculation formula, we obtain the final wear state of the cutting teeth.

[0147] In practical applications, a two-layer feature extraction model can be trained by collecting a large amount of accelerometer sensor data and feature data, and this model can be used to predict the wear state of tunnel boring machine (TBM) cutter teeth. By combining the feature extraction model and the wear state formula, the wear state of TBM cutter teeth can be predicted more accurately, improving the efficiency and accuracy of equipment maintenance. For example:

[0148] This formula uses a two-layer LSTM (Long Short-Term Memory) network as the feature extraction model to extract feature representations of the input sequence. Below are the specific expressions for each parameter and the methods for determining their values ​​in examples:

[0149] h 1,t and c 1,t : Represents the hidden state and cell state of the first layer in the two-layer feature extraction model at time step t. In LSTM, the hidden state h and cell state c are obtained through operations within the LSTM unit, and are used as memory units for transmitting information and controlling the information flow, respectively.

[0150] h 2,t and c 2,t : Represents the hidden state and cell state of the second layer in the two-layer feature extraction model at time step t. The hidden state and cell state of the second layer are calculated based on the hidden state and cell state of the first layer, and are used to learn a higher-level feature representation.

[0151] y t : Represents the output result at time step t, passed through the weight matrix W of the output layer. hy and bias term b y For the second hidden state h 2,t It is obtained by performing a linear transformation and activation function g.

[0152] In this embodiment of the invention, hidden states and cell states are important concepts in the LSTM (Long Short-Term Memory) model, used to capture long-term dependencies and memory characteristics in time-series data. In sensor data feature extraction, hidden states and cell states can be understood as states that encode and remember the features in the sensor data and the relationships between them. Specifically, the hidden state h can be seen as the extracted features of the sensor data at the current time step, containing an abstract representation of the current data and reflecting the important information contained in the current data; while the cell state c is the memory of long-term dependencies between data, which is controlled by the gating structure of the LSTM model to control the input and forgetting of information, thereby capturing long-term dependencies between data. Therefore, in sensor data feature extraction, using hidden states and cell states can better capture and encode the features and relationships in sensor data, helping the model to better understand and predict the state and changes of sensor data.

[0153] In specific examples, the following parameters and value retrieval methods can be selected:

[0154] The hidden state size of LSTM1 is d1, and the cell state size is d1, i.e., h 1,t and c 1,t The dimension is d1.

[0155] The hidden state size of LSTM2 is d2, and the cell state size is d2, i.e., h 2,t and c 2,t The dimension is d2.

[0156] The activation function for the output layer can be either the ReLU function or the Sigmoid function, etc.

[0157] The weight matrix W of the output layer hy The dimension is d2×m, where m is the output dimension, which can be set according to the specific task.

[0158] output layer bias term b y The dimension is m.

[0159] In the field of tooth cutting state monitoring technology, appropriate activation functions g and output layer weight matrices W can be selected. hy and bias term b y Let's build a model suitable for predicting the state of a cut-off tooth. Specifically, the activation function g is as follows: In cut-off tooth state monitoring, common activation functions include the Sigmoid function and the Softmax function. Both functions can map the output value to a range, suitable for binary or multi-class classification tasks. Sigmoid function: Mapping output values ​​to the range (0,1) is suitable for binary classification tasks. The Softmax function: The output values ​​are normalized, making it suitable for multi-class classification tasks. The weight matrix W of the output layer... hy and bias term b y In the monitoring of the cutting edge state, the weight matrix and bias terms of the output layer can be defined as follows: Weight matrix W hy The dimension of the output is d2×1, where d2 is the dimension of the second-layer hidden state, and the dimension of the output is 1 because it may be performing a binary classification task (e.g., predicting whether the cutting edge is worn). The bias term b... y The dimension is 1, which is used to adjust the offset of the output result.

[0160] Based on the above, an output layer formula suitable for tooth cutting condition monitoring can be constructed as follows:

[0161] y t =g(W hy h 2,t +by )

[0162] Where g can be either the Sigmoid function or the Softmax function, W hy It is a d2×1 weight matrix, b y It is a bias term. Through this output layer formula, the second-layer hidden state h can be... 2,t It is mapped to a range and used to predict the wear condition of the cutting teeth.

[0163] In practical applications, the parameters in the model can be learned through training data, including the weights and biases in the LSTM, as well as the weight matrix and biases of the output layer. The model parameters are then updated using backpropagation and a loss function, enabling the model to better fit the data and make predictions. Through the above design and parameter selection, a two-layer feature extraction model can be constructed and combined with the output layer to predict the wear state of the cutting teeth.

[0164] This invention also provides a device, comprising: a processor; and a memory for storing processor-executable instructions;

[0165] The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0166] This invention also provides a computer-readable storage medium storing computer program instructions that, when executed by a processor, implement the aforementioned method. This invention can be a method, apparatus, system, and / or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of this invention.

[0167] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for online monitoring of cutter tooth wear in a cantilever tunneling machine, characterized in that, Includes the following steps: S1: Real-time acquisition of sensor data affecting the wear state of the cutting teeth, including but not limited to the pressure signal of the cutting cylinder, the vibration signal of the cutting reducer, and the temperature signal of the cutting head; S2: Extract features from the above sensor data and determine feature weights based on the correlation between the sensors; the extracted features include: First characteristic Second feature and third feature And the expressions for the above features are as follows: In the formula, For pressure sensor data, The pressure is a preset value, and Greater than , For pressure regulation parameters, It is a hyperbolic tangent function; In the formula, For accelerometer data, The acceleration is a preset value, and Greater than , For acceleration adjustment parameters, For one Type function; In the formula, For temperature sensor data, This is the preset temperature value, and Greater than 0, For temperature control parameters, It is a rectified linear unit function; S3: Analyze the wear of the cutting teeth based on the extracted features and the correlation between the sensors to obtain the initial wear state of the cutting teeth; S4: Based on the two-layer feature extraction model, the above features are corrected and updated to obtain the corrected cutter wear state; the wear of the cutter is monitored based on the corrected cutter wear state.

2. The online monitoring method for wear of cutting teeth in a cantilever tunneling machine according to claim 1, characterized in that: In S2, the feature weight values ​​obtained based on the correlation between the sensors include: First feature weight value Second feature weight value and the third feature weight value And the expressions for the above feature weight values ​​are as follows: In the formula, Indicates sensor and sensors The correlation between them For pressure sensors, It is a temperature sensor. It is an accelerometer.

3. The online monitoring method for wear of cutting teeth in a cantilever tunneling machine according to claim 2, characterized in that: In S3, the formula for the initial wear state of the cutting teeth is: 。 4. The online monitoring method for wear of cutting teeth in a cantilever tunneling machine according to claim 3, characterized in that: In S4, the steps for modifying the features include: The first feature Second feature and third feature Set as the input sequence; A two-layer feature extraction model is used to extract features from the input sequence and output the corrected tooth wear state. The modified formula for the wear state of the cutting teeth is: In the formula, , , These represent the wear state sequences of the cutting teeth calculated using a two-layer feature extraction model.

5. An online monitoring system for wear of cutting teeth in a cantilever tunneling machine, used to implement the online monitoring method for wear of cutting teeth in a cantilever tunneling machine as described in any one of claims 1 to 4, characterized in that: It includes a sensor module, a data acquisition module, a data analysis module, and a monitoring module; The sensor module includes a pressure sensor, an acceleration sensor, and a temperature sensor. The pressure sensor is installed on the oil port of the multi-way directional valve of the hydraulic system of the cantilever tunneling machine, which is used to indirectly monitor the force on the cutting teeth. The acceleration sensor is installed on the cutting reducer of the cantilever tunneling machine, which is used to monitor the vibration of the cutting teeth. The temperature sensor is installed inside the cutting head of the cantilever tunneling machine, which is used to monitor the temperature change of the cutting teeth. The data acquisition module is used to receive data sent by the sensor module and transmit it to the data analysis module; The data analysis module is used to extract features from sensor data, determine the correlation between sensors, and analyze the wear of the cutting teeth based on the sensor data and the correlation between sensors to obtain the initial wear state of the cutting teeth; at the same time, the extracted features are optimized through a two-layer feature extraction model to update and obtain the corrected wear state of the cutting teeth. The monitoring module is used to acquire the cutting tooth status information sent by the data analysis module in real time and monitor the wear of the cutting tooth.

6. The online monitoring system for wear of cutting teeth in a cantilever tunneling machine according to claim 5, characterized in that: The data analysis module includes a preprocessing module, an extraction module, and a state feature determination module; The preprocessing module is used to preprocess the sensor data to obtain intermediate data; The extraction module is used to extract state features from intermediate data to describe the wear state of the cutting teeth; The status characteristic determination module is used to determine the characteristics of the pressure sensor, acceleration sensor, and temperature sensor, and to comprehensively evaluate the wear condition of the cutting tooth based on the above characteristics, thereby determining the status information of the cutting tooth.

7. The online monitoring system for wear of cutting teeth in a cantilever tunneling machine according to claim 5, characterized in that: Input sequence of a two-layer feature extraction model middle The first feature Second feature and third feature ; The formula for feature extraction in the two-layer feature extraction model is as follows: In the formula, and These represent the first layer of the two-layer feature extraction model at time step [number missing]. Hidden states and cellular states within the cell. and These represent the second layer of the two-layer feature extraction model at time step [number missing]. Hidden states and cellular states within; and These represent the first layer of the two-layer feature extraction model at time step [number missing]. Hidden states and cellular states within; and These represent the second layer of the two-layer feature extraction model at time step [number missing]. Hidden states and cellular states within; This indicates that the first layer of the two-layer feature extraction model is at time step [missing information]. Input sequence within ; This represents the first layer of a long short-term memory network model; This represents the second layer of the Long Short-Term Memory (LSTM) network model; The output of the two-layer feature extraction model is: In the formula, It is an activation function. and These are the weight matrix and bias terms of the output layer. As a correction of the wear state of the cutting teeth.

8. An online monitoring device for wear of cutting teeth in a cantilever tunneling machine, comprising a memory and a processor, characterized in that: The memory contains a computer program, and the processor is configured to run the computer program to perform the cantilever tunneling machine cutter wear monitoring method according to any one of claims 1 to 4.

9. A readable storage medium, characterized in that: The readable storage medium stores a computer program, which includes program code for controlling a process to execute the process, the process including the method for monitoring wear of cutter teeth of a cantilever tunneling machine as described in any one of claims 1 to 4.

Images