A safety operation early warning method and system for a tunnel boring machine main machine dismounting device

Abstract

The application discloses a tunnel boring machine main machine dismounting device safety operation early warning method and system, the method comprises the following steps: preprocessing the data collected by each sensor, and obtaining a safety state coefficient by combining the analytic hierarchy process and the entropy weight method; taking the safety state coefficient as input, establishing different types of prediction models to predict the safety state coefficient of the next stage; meanwhile, a plurality of combined prediction models based on the least square error and the least square method are established according to the mutual combination of different types of prediction models, and the optimal combined prediction model with the best prediction effect is selected as a safety warning model; the safety state coefficient of the dismounting device in the later stage is predicted based on the safety warning model, and the safety state coefficient is compared with the operation warning range and the risk level of the dismounting device to determine whether the dismounting device continues to operate. The application can effectively warn the safety condition of the dismounting device operation, and can ensure the operation safety of the dismounting device.

Description

Technical Field

[0001] This invention relates to the field of tunnel boring machine technology, and in particular to a method and system for early warning of safe operation of the main unit disassembly and assembly device of a tunnel boring machine. Background Technology

[0002] The TBM (Tunnel Boring Machine) main unit assembly and disassembly unit is a large lifting device. During operation, it requires slow lifting and hoisting, resulting in a long working time. Furthermore, the working mode needs to be flexibly adjusted according to the actual site conditions, and the required operating parameters vary depending on the time period. The limited space inside the tunnel and the large tonnage of the main unit make operation difficult and dangerous, posing significant safety hazards. To ensure the safety of the assembly and disassembly unit, effective early warning systems for safety conditions during operation are necessary, based on the structural characteristics and technological requirements of the TBM assembly and disassembly unit. Summary of the Invention

[0003] This invention addresses the problems of existing tunnel boring machine (TBM) main unit disassembly and assembly devices, which are difficult to operate, highly dangerous, and pose significant safety hazards. It proposes a safety operation early warning method and system for TBM main unit disassembly and assembly devices, which can effectively provide early warning of the safety status during disassembly and assembly operations, thus ensuring the safety of the disassembly and assembly devices.

[0004] To achieve the above objectives, the present invention adopts the following technical solution:

[0005] This invention discloses a safety operation early warning method for a tunnel boring machine (TBM) main unit disassembly and assembly device. The TBM main unit disassembly and assembly device is used for lifting and hoisting the TBM main unit. The device is connected to the TBM main unit via two sets of steel wire ropes. The device is equipped with an off-center load sensor, a tension sensor group, a displacement sensor, and a pressure sensor group, which are used to collect off-center load angle data, tension data of the two sets of steel wire ropes, lifting height data, and pressure data of the hydraulic cylinder group during the lifting and hoisting process of the TBM main unit. The method includes:

[0006] Step 1: Preprocess the off-center load angle data, tension data of the two sets of wire ropes, lifting height data, and pressure data of the hydraulic cylinder group during the hoisting and lifting process of the tunnel boring machine main unit;

[0007] Step 2: The preprocessed data above is fused using a combination of the analytic hierarchy process and the entropy weight method to obtain the safety status coefficient. The safety status coefficient is then compared with the operational alarm level and risk level of the disassembly and assembly device to determine the safety status of the disassembly and assembly device at this time.

[0008] Step 3: Using the safety status coefficient obtained from the fusion as input, establish different types of prediction models to predict the safety status coefficient of the disassembly and assembly device in the next stage; at the same time, establish multiple combined prediction models based on the minimum error square and least squares method according to the mutual combination of different types of prediction models; by comparing the safety status coefficient predicted by the combined prediction model with the actual value, select the optimal combined prediction model with the best prediction effect as the safety early warning model.

[0009] Step 4: Based on the safety early warning model, predict the safety status coefficient of the disassembly and assembly device in the later stage, and compare the safety status coefficient with the operation alarm level and risk level of the disassembly and assembly device to determine whether the disassembly and assembly device should continue to operate. If the predicted safety status coefficient exceeds the safety range, issue an alarm and stop the operation of the disassembly and assembly device.

[0010] Further, step 1 includes: normalizing the off-center load angle data, tension data of the two sets of wire ropes, lifting height data, and pressure data of the hydraulic cylinder group during the hoisting process of the tunnel boring machine main unit.

[0011] Furthermore, in step 2, the safety state coefficient is obtained as follows:

[0012]

[0013] Where ω i The comprehensive weight of index i is obtained by combining the analytic hierarchy process (AHP) and the entropy weight method, α i β represents the weight of index i obtained by the analytic hierarchy process. i The weight of index i obtained by the entropy weight method, wherein the index corresponds to the data type collected;

[0014] Multiply the comprehensive weight of each indicator by the corresponding preprocessed data of the same type, and then add them together to obtain the safety status coefficient of the main unit disassembly and assembly device of the tunnel boring machine that is currently in operation.

[0015] Furthermore, in step 3, the types of prediction models established include: GM, ARIMA, and LSTM.

[0016] Furthermore, in step 3, the combined prediction model is established as follows:

[0017]

[0018] Where f t f represents the predicted value of the combined prediction model at time t; n represents the total number of prediction models in the combined prediction model; it k represents the predicted safety state coefficient of the i-th prediction model at time t; iThe weights assigned to the i-th prediction model, and

[0019] The weights assigned to each type of prediction model in the combined prediction model are determined as follows:

[0020]

[0021]

[0022]

[0023] Where J is the sum of squared errors between the predicted and actual values ​​of the combined prediction model, and K n It is the weight coefficient vector of the combined prediction model, E (n) R is the error information matrix between the predicted and actual values ​​of the combined prediction model. n It is an n×1 column matrix.

[0024] Furthermore, in step 3, the optimal combined prediction model is obtained as follows:

[0025] Based on the predicted safety state coefficient of each combined prediction model at the current time and the actual safety state coefficient at that time, calculate the four indicators corresponding to each combined prediction model: correlation coefficient, mean absolute error, mean relative error, and root mean square error. Compare the above four indicators corresponding to each combined prediction model to obtain the optimal combined prediction model.

[0026] A safety operation early warning system for the disassembly and assembly device of a tunnel boring machine includes:

[0027] The data preprocessing module is used to preprocess the off-center load angle data, tension data of the two sets of wire ropes, lifting height data, and pressure data of the hydraulic cylinder group during the hoisting and lifting process of the tunnel boring machine main unit.

[0028] The first data fusion module is used to fuse the preprocessed data using a combination of the analytic hierarchy process and the entropy weight method to obtain a safety status coefficient. Based on the safety status coefficient, the module compares the operational alarm level and risk level of the disassembly and assembly device to obtain the safety status of the disassembly and assembly device at this time.

[0029] The second data fusion module is used to take the fused safety status coefficient as input, establish different types of prediction models, and predict the safety status coefficient of the disassembly and assembly device in the next stage. At the same time, it establishes multiple combined prediction models based on the minimum square error and least squares method according to the mutual combination of different types of prediction models. By comparing the safety status coefficient predicted by the combined prediction model with the actual value, the optimal combined prediction model with the best prediction effect is selected as the safety early warning model.

[0030] The early warning module is used to predict the safety status coefficient of the disassembly and assembly device based on the safety early warning model, and to compare the safety status coefficient with the operation alarm level and risk level of the disassembly and assembly device to determine whether the disassembly and assembly device should continue to operate. If the predicted safety status coefficient exceeds the safety range, an alarm is issued and the operation of the disassembly and assembly device is stopped.

[0031] Furthermore, the data preprocessing module is specifically used to normalize the off-center load angle data, the tension data of the two sets of wire ropes, the lifting height data, and the pressure data of the hydraulic cylinder group during the hoisting process of the tunnel boring machine main unit.

[0032] Furthermore, in the first data fusion module, the security status coefficient is obtained in the following manner:

[0033]

[0034] Where ω i The comprehensive weight of index i is obtained by combining the analytic hierarchy process (AHP) and the entropy weight method, α i β represents the weight of index i obtained by the analytic hierarchy process. i The weight of index i obtained by the entropy weight method, wherein the index corresponds to the data type collected;

[0035] Multiply the comprehensive weight of each indicator by the corresponding preprocessed data of the same type, and then add them together to obtain the safety status coefficient of the main unit disassembly and assembly device of the tunnel boring machine that is currently in operation.

[0036] Furthermore, the types of prediction models established in the second data fusion module include: GM, ARIMA, and LSTM.

[0037] Furthermore, in the second data fusion module, a combined prediction model is established as follows:

[0038]

[0039] Where f t f represents the predicted value of the combined prediction model at time t; n represents the total number of prediction models in the combined prediction model; it k represents the predicted safety state coefficient of the i-th prediction model at time t; i The weights assigned to the i-th prediction model, and

[0040] The weights assigned to each type of prediction model in the combined prediction model are determined as follows:

[0041]

[0042]

[0043]

[0044] Where J is the sum of squared errors between the predicted and actual values ​​of the combined prediction model, and K n It is the weight coefficient vector of the combined prediction model, E (n) R is the error information matrix between the predicted and actual values ​​of the combined prediction model. n It is an n×1 column matrix.

[0045] Furthermore, in the second data fusion module, the optimal combined prediction model is obtained in the following manner:

[0046] Based on the predicted safety state coefficient of each combined prediction model at the current time and the actual safety state coefficient at that time, calculate the four indicators corresponding to each combined prediction model: correlation coefficient, mean absolute error, mean relative error, and root mean square error. Compare the above four indicators corresponding to each combined prediction model to obtain the optimal combined prediction model.

[0047] Compared with the prior art, the present invention has the following beneficial effects:

[0048] This invention addresses the problems of high operational difficulty and danger, posing significant safety hazards, in existing tunnel boring machine (TBM) main unit assembly / disassembly devices. It proposes a safety operation early warning method and system for these devices. The method combines the Analytic Hierarchy Process (AHP) and the entropy weight method to derive a safety state coefficient. This coefficient is then compared with the operational alert level and risk grade of the assembly / disassembly device to determine its current safety status. The fused safety state coefficient is used as input to establish different types of prediction models to predict the safety state coefficient for the next stage of the assembly / disassembly device operation. Simultaneously, multiple combined prediction models based on minimum squared error and least squares methods are established by combining different types of prediction models. The optimal combined prediction model, with the best prediction performance, is selected as the safety early warning model by comparing the predicted safety state coefficient with the actual value. This invention offers high prediction accuracy and effectively provides early warnings of the safety status during assembly / disassembly device operation, ensuring operational safety. Attached Figure Description

[0049] Figure 1 This is a flowchart of a safety operation early warning method for the disassembly and assembly device of a tunnel boring machine according to an embodiment of the present invention;

[0050] Figure 2 This is one of the structural schematic diagrams of the tunnel boring machine main unit disassembly and assembly device according to an embodiment of the present invention;

[0051] Figure 3This is the second schematic diagram of the structure of the tunnel boring machine main unit disassembly and assembly device according to an embodiment of the present invention;

[0052] Figure 4 This is a schematic diagram of a two-level information fusion security early warning model constructed according to an embodiment of the present invention;

[0053] Figure 5 This is a prediction graph of the optimal weighted combination prediction model in an embodiment of the present invention;

[0054] Figure 6 This is a schematic diagram of the architecture of a safe operation early warning system for the disassembly and assembly device of a tunnel boring machine according to an embodiment of the present invention;

[0055] In the diagram: 1 is the off-center load sensor, 2 is the trolley, 3 is the crossbeam, 4 is the first tension sensor, 5 is the second tension sensor, 6 is the displacement sensor, 7 is the wire rope, 8 is the four-stage hydraulic cylinder, 9 is the tunnel boring machine main unit, 10 is the first pressure sensor, 11 is the second pressure sensor, 12 is the base, 13 is the bottom traveling mechanism, and 14 is the bottom traveling track. Detailed Implementation

[0056] The present invention will be further explained below with reference to the accompanying drawings and specific embodiments:

[0057] A safety operation early warning method for a tunnel boring machine (TBM) main unit disassembly and assembly device, wherein the TBM main unit disassembly and assembly device (hereinafter referred to as the disassembly and assembly device for convenience) is used for lifting and hoisting the TBM main unit 9, and the TBM main unit disassembly and assembly device is connected to the TBM main unit 9 by two sets of steel wire ropes 7; the TBM main unit disassembly and assembly device is as follows: Figure 1 , Figure 2 As shown, the tunnel boring machine (TBM) main unit assembly / disassembly device is equipped with an off-center load sensor 1, a tension sensor group (including a first tension sensor 4 and a second tension sensor 5), a displacement sensor 6, and a pressure sensor group (including two first pressure sensors 10 and two second pressure sensors 11). The off-center load sensor 1 is used to collect off-center load angle data during the lifting and hoisting process of the TBM main unit 9; the displacement sensor 6 is used to collect lifting height data during the lifting and hoisting process of the TBM main unit 9; the first pressure sensors 10 and second pressure sensors 11 are respectively used to collect pressure data of the hydraulic cylinder group (including four four-stage hydraulic cylinders 8) during the lifting and hoisting process of the TBM main unit 9; the first tension sensor 4 and second tension sensor 5 are located on the hook and are used to collect tension data of the two sets of wire ropes 7 during the lifting and hoisting process of the TBM main unit 9; the method includes:

[0058] Step 1: Preprocess the off-center load angle data, tension data of the two sets of wire ropes 7, lifting height data, and hydraulic cylinder pressure data during the lifting and hoisting process of the tunnel boring machine main unit 9;

[0059] Step 2: The above eight sets of preprocessed data (including 1 set of off-center load angle data, 1 set of lifting height data, 4 sets of pressure data, and 2 sets of tension data) are fused together using the combination of the analytic hierarchy process and the entropy weight method to obtain the safety status coefficient. The safety status coefficient is then compared with the operational alarm level and risk level of the disassembly and assembly device to obtain the safety status of the disassembly and assembly device at this time.

[0060] Step 3: Using the safety status coefficient obtained from the fusion as input, establish different types of prediction models to predict the safety status coefficient of the disassembly and assembly device in the next stage; at the same time, establish multiple combined prediction models based on the minimum error square and least squares method according to the mutual combination of different types of prediction models; by comparing the safety status coefficient predicted by the combined prediction model with the actual value, select the optimal combined prediction model with the best prediction effect as the safety early warning model.

[0061] Step 4: Based on the safety early warning model, predict the safety status coefficient of the disassembly and assembly device in the later stage, and compare the safety status coefficient with the operation alarm level and risk level of the disassembly and assembly device to determine whether the disassembly and assembly device should continue to operate. If the predicted safety status coefficient exceeds the safety range, issue an alarm and stop the operation of the disassembly and assembly device.

[0062] Specifically, such as Figure 3 As shown, the safety operation early warning method is divided into three monitoring stages. The first stage is when the tunnel boring machine (TBM) main unit disassembly and assembly device reaches the designated work position, hooks the hook onto the lifting point at the head of the TBM main unit 9, and then begins to lift the TBM main unit 9. At this time, the tension of the wire rope 7, the pressure of the four-stage hydraulic cylinder 8, the lifting height of the TBM main unit 9, and the off-center load angle of the TBM main unit 9 are all at their lowest values. As the four-stage hydraulic cylinder 8 is lifted, the TBM main unit disassembly and assembly device moves forward, and the wire rope 7 will be stretched. This stage continues until the wire rope 7 has tension and the pressure of the four-stage hydraulic cylinder 8 changes.

[0063] The second stage: The head of the tunnel boring machine (TBM) main unit 9 will be gradually lifted. During this time, the tension of the wire rope 7 and the pressure of the four-stage hydraulic cylinder 8 will gradually increase, and the lifting height of the TBM main unit 9 will increase uniformly. There will also be a vertical off-center load angle between the TBM main unit 9 and the TBM main unit assembly / disassembly device. The sensors will collect data on tension, pressure, height, and off-center load angle, which will be converted from data by an A / D converter and transmitted to the controller, and then to a remote PC. This data will be used to establish a two-level information fusion safety early warning model, such as... Figure 4 As shown.

[0064] First-level fusion: The four types of raw data are first normalized.

[0065]

[0066] In the formula x ij It is the j-th value of index i in the original data, x min It is the minimum value of index i when the tunnel boring machine's main unit disassembly and assembly device is unloaded, x max It is the maximum value of the i index.

[0067] Then, the normalized index data at a given time point are combined with the weights determined by the corresponding fusion algorithm. Specifically, a fusion algorithm combining the analytic hierarchy process (AHP) and the entropy weight method can yield more scientific and accurate index weight values.

[0068]

[0069] In the above formula, ω i The comprehensive weight of index i is obtained by combining the analytic hierarchy process (AHP) and the entropy weight method, α i β represents the weights obtained through the analytic hierarchy process. i The weights obtained through the entropy weighting method correspond to the data types (groups) collected. Specifically, there are a total of 8 indicators.

[0070] Multiplying the comprehensive weight of each indicator by the preprocessed corresponding type (group) data and then summing them yields the safety status coefficient of the tunnel boring machine main unit assembly / disassembly device at that moment. The value of the safety status coefficient ranges from 0 to 1. This safety status coefficient can be compared with the alarm level and risk level of the tunnel boring machine main unit assembly / disassembly device to determine whether the operation is safe. Specifically, in this embodiment, the alarm level and risk level classification criteria are shown in Table 1.

[0071] Table 1 Alertness and Risk Level Classification Standards

[0072]

[0073]

[0074] Second-level fusion: Using the security state coefficient obtained from the first-level fusion as input, prediction models such as GM, ARIMA, and LSTM are established to predict the security state coefficient in subsequent stages. Let the predicted value of the i-th individual prediction model at time t be f. it The i-th single-item prediction model is assigned a weight of k. i ,and n represents the total number of individual prediction models in the combined prediction model, then the predicted value f of the combined prediction model at time t is... t for:

[0075]

[0076] The purpose of using a combined prediction model is to reduce the sum of squared errors between predicted and actual values. Therefore, the problem of weight allocation for the optimal combined prediction model is transformed into a mathematical programming problem using the least squares method to minimize the sum of squared errors, i.e.:

[0077]

[0078] Where J is the sum of squared errors between the predicted and actual values ​​of the combined prediction model, and K n It is the weight coefficient vector of the combined prediction model, E (n) This is the error information matrix between the predicted and actual values ​​of the combined prediction model. Solving the above equation using the Lagrange multiplier method yields:

[0079]

[0080] Where R n Given an n×1 column matrix, combining the above two equations, the minimum sum of squared errors of the combined prediction model can be obtained as:

[0081]

[0082] Based on the above method, multiple (optimal weighted) combined prediction models were obtained.

[0083] Based on the predicted safety state coefficient of each combined prediction model at the current time and the actual safety state coefficient at that time, calculate the four indicators corresponding to each combined prediction model: correlation coefficient (R), mean absolute error (MAE), mean relative error (MAPE), and root mean square error (RMSE). Compare the above four indicators corresponding to each combined prediction model to obtain the optimal combined prediction model, and use the optimal combined prediction model as the safety early warning model.

[0084] Specifically, the four indicators and their expressions are as follows.

[0085] 1) Correlation coefficient (R), the expression is:

[0086]

[0087] Where y t This represents the actual safety state coefficient at time t. This represents the average value of the actual safety status coefficient. This represents the average value of the predicted safety status coefficient.

[0088] 2) Mean Absolute Error (MAE), the expression is:

[0089]

[0090] 3) Mean relative error (MAPE), the expression is:

[0091]

[0092] 4) Root Mean Square Error (RMSE), expressed as:

[0093]

[0094] Based on the safety early warning model, the safety status coefficient of the disassembly and assembly device is predicted. The safety status coefficient is compared with the operational alarm level and risk level of the disassembly and assembly device to determine whether the disassembly and assembly device should continue to operate. If the predicted safety status coefficient exceeds the safety range, an alarm is issued and the disassembly and assembly device is stopped from operating.

[0095] As one possible implementation, when n is 3, three individual prediction models (GM, ARIMA, and LSTM) are established to predict the safety state coefficient in the later stages. The prediction result of the optimal weighted combination model is as follows: Figure 5 As shown in the figure, the four indicators of the GM-LSTM and GM-ARIMA-LSTM combined prediction models are quite similar and better than the other two combined prediction models. Although the GM-LSTM combined prediction model is better than the GM-ARIMA-LSTM combined prediction model in all indicators, in the later prediction data group (sequence numbers 25-30), the relative error between the predicted value and the actual value of the GM-ARIMA-LSTM combined prediction model is smaller than that of GM-LSTM, and its predicted value is closer to the actual value. It has the best fitting effect in the later stage and the best prediction accuracy among all combined prediction models. Therefore, the GM-ARIMA-LSTM combined prediction model is selected as the second-level fusion safety early warning model from the optimal weighted combined prediction models. It is combined with the first-level fusion to construct a two-level information fusion-based safety early warning model for disassembly and assembly operations. If the second stage of operation is safe throughout, the third stage will proceed.

[0096] The third stage: The tunnel boring machine (TBM) main unit 9 is completely off the ground, and the safety status prediction is stopped. Special attention needs to be paid to the pressure, tension, and off-center load angle. The TBM main unit 9 will inevitably sway after being off the ground, and the off-center load angle may increase. At the same time, the tension and pressure are at their maximum. It is necessary to prevent these three values ​​from exceeding the specified values. In addition, the lifting height of the TBM main unit 9 should not exceed the maximum lifting height. Lifting should be stopped when the specified height is reached.

[0097] Based on the above embodiments, such as Figure 6 As shown, a safety operation early warning system for the disassembly and assembly of a tunnel boring machine host is also proposed, including:

[0098] The data preprocessing module is used to preprocess the off-center load angle data, tension data of the two sets of wire ropes 7, lifting height data, and pressure data of the hydraulic cylinder group during the lifting and hoisting process of the tunnel boring machine main unit 9.

[0099] The first data fusion module is used to fuse the preprocessed data using a combination of the analytic hierarchy process and the entropy weight method to obtain a safety status coefficient. Based on the safety status coefficient, the module compares the operational alarm level and risk level of the disassembly and assembly device to obtain the safety status of the disassembly and assembly device at this time.

[0100] The second data fusion module is used to take the fused safety status coefficient as input, establish different types of prediction models, and predict the safety status coefficient of the disassembly and assembly device in the next stage. At the same time, it establishes multiple combined prediction models based on the minimum square error and least squares method according to the mutual combination of different types of prediction models. By comparing the safety status coefficient predicted by the combined prediction model with the actual value, the optimal combined prediction model with the best prediction effect is selected as the safety early warning model.

[0101] The early warning module is used to predict the safety status coefficient of the disassembly and assembly device based on the safety early warning model, and to compare the safety status coefficient with the operation alarm level and risk level of the disassembly and assembly device to determine whether the disassembly and assembly device should continue to operate. If the predicted safety status coefficient exceeds the safety range, an alarm is issued and the operation of the disassembly and assembly device is stopped.

[0102] Furthermore, the data preprocessing module is specifically used to normalize the off-center load angle data, the tension data of the two sets of wire ropes 7, the lifting height data, and the pressure data of the hydraulic cylinder group during the lifting and hoisting process of the tunnel boring machine main unit 9.

[0103] Furthermore, in the first data fusion module, the security status coefficient is obtained in the following manner:

[0104]

[0105] Where ω i The comprehensive weight of index i is obtained by combining the analytic hierarchy process (AHP) and the entropy weight method, α i β represents the weight of index i obtained by the analytic hierarchy process. i The weight of index i obtained by the entropy weight method, wherein the index corresponds to the data type collected;

[0106] Multiply the comprehensive weight of each indicator by the corresponding preprocessed data of the same type, and then add them together to obtain the safety status coefficient of the main unit disassembly and assembly device of the tunnel boring machine that is currently in operation.

[0107] Furthermore, the types of prediction models established in the second data fusion module include: GM, ARIMA, and LSTM.

[0108] Furthermore, in the second data fusion module, a combined prediction model is established as follows:

[0109]

[0110] Where f t f represents the predicted value of the combined prediction model at time t; n represents the total number of prediction models in the combined prediction model; it k represents the predicted safety state coefficient of the i-th prediction model at time t; i The weights assigned to the i-th prediction model, and

[0111] The weights assigned to each type of prediction model in the combined prediction model are determined as follows:

[0112]

[0113]

[0114]

[0115] Where J is the sum of squared errors between the predicted and actual values ​​of the combined prediction model, and K n It is the weight coefficient vector of the combined prediction model, E (n) R is the error information matrix between the predicted and actual values ​​of the combined prediction model. n It is an n×1 column matrix.

[0116] Furthermore, in the second data fusion module, the optimal combined prediction model is obtained in the following manner:

[0117] Based on the predicted safety state coefficient of each combined prediction model at the current time and the actual safety state coefficient at that time, calculate the four indicators corresponding to each combined prediction model: correlation coefficient, mean absolute error, mean relative error, and root mean square error. Compare the above four indicators corresponding to each combined prediction model to obtain the optimal combined prediction model.

[0118] In summary, this invention addresses the problems of high operational difficulty and risk associated with existing tunnel boring machine (TBM) main unit assembly / disassembly devices, which pose significant safety hazards. It proposes a safety operation early warning method and system for TBM main unit assembly / disassembly devices. This method combines the Analytic Hierarchy Process (AHP) and the entropy weight method to derive a safety state coefficient. This safety state coefficient is then compared with the operational alert level and risk grade of the assembly / disassembly device to determine its current safety status. The fused safety state coefficient is then used as input to establish different types of prediction models to predict the safety state coefficient of the assembly / disassembly device in the next stage. Simultaneously, multiple combined prediction models based on minimum squared error and least squares methods are established by combining different types of prediction models. By comparing the safety state coefficients predicted by these combined prediction models with the actual values, the optimal combined prediction model with the best prediction effect is selected as the safety early warning model. This invention has high prediction accuracy and can effectively provide early warnings of the safety status during assembly / disassembly device operation, ensuring the operational safety of the assembly / disassembly device.

[0119] The above description is only a preferred embodiment of the present invention. It should be noted that those skilled in the art can make several improvements and modifications without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A safety operation early warning method for a tunnel boring machine (TBM) main unit disassembly and assembly device, wherein the TBM main unit disassembly and assembly device is used for lifting and hoisting the TBM main unit, and the TBM main unit disassembly and assembly device is connected to the TBM main unit via two sets of steel wire ropes, characterized in that... The tunnel boring machine (TBM) main unit disassembly and assembly device is equipped with an off-center load sensor, a tension sensor group, a displacement sensor, and a pressure sensor group, which are used to collect off-center load angle data, tension data of the two sets of wire ropes, lifting height data, and pressure data of the hydraulic cylinder group during the lifting and hoisting process of the TBM main unit; the method includes: Step 1: Preprocess the off-center load angle data, tension data of the two sets of wire ropes, lifting height data, and pressure data of the hydraulic cylinder group during the hoisting and lifting process of the tunnel boring machine main unit; Step 2: The preprocessed data above is fused using a combination of the analytic hierarchy process and the entropy weight method to obtain the safety status coefficient. The safety status coefficient is then compared with the operational alarm level and risk level of the disassembly and assembly device to determine the safety status of the disassembly and assembly device at this time. Step 3: Using the safety status coefficient obtained from the fusion as input, establish different types of prediction models to predict the safety status coefficient of the disassembly and assembly device in the next stage; at the same time, establish multiple combined prediction models based on the minimum error square and least squares method according to the mutual combination of different types of prediction models; by comparing the safety status coefficient predicted by the combined prediction model with the actual value, select the optimal combined prediction model with the best prediction effect as the safety early warning model. Step 4: Based on the safety early warning model, predict the safety status coefficient of the disassembly and assembly device in the later stage, and compare the safety status coefficient with the operation alarm level and risk level of the disassembly and assembly device to determine whether the disassembly and assembly device should continue to operate. If the predicted safety status coefficient exceeds the safety range, issue an alarm and stop the operation of the disassembly and assembly device.

2. The method for early warning of safe operation of the main unit disassembly and assembly device of a tunnel boring machine according to claim 1, characterized in that, Step 1 includes: normalizing the off-center load angle data, tension data of the two sets of wire ropes, lifting height data, and pressure data of the hydraulic cylinder group during the hoisting process of the tunnel boring machine main unit.

3. The method for early warning of safe operation of the tunnel boring machine main unit disassembly and assembly device according to claim 1, characterized in that, In step 2, the safety state coefficient is obtained as follows: in for The comprehensive weights of the indicators are obtained through a combination of the analytic hierarchy process (AHP) and the entropy weight method. Obtained by the Analytic Hierarchy Process The weight of the indicator Obtained by the entropy weight method The weight of the indicator, which corresponds to the data type collected; Multiply the comprehensive weight of each indicator by the corresponding preprocessed data of the same type, and then add them together to obtain the safety status coefficient of the main unit disassembly and assembly device of the tunnel boring machine that is currently in operation.

4. A method for early warning of safe operation of a tunnel boring machine main unit disassembly and assembly device according to claim 1, characterized in that, In step 3, the types of prediction models established include: GM, ARIMA, and LSTM.

5. A method for early warning of safe operation of a tunnel boring machine main unit disassembly and assembly device according to claim 1, characterized in that, In step 3, the combined prediction model is established as follows: in Indicates the first The predicted value of the time-combined prediction model; n represents the total number of prediction models in the combined prediction model; Indicates the first The predicted safety state coefficient of the i-th prediction model at time i; The weights assigned to the i-th prediction model, and ; The weights assigned to each type of prediction model in the combined prediction model are determined as follows: in This is the sum of squared errors between the predicted and actual values ​​of the combined prediction model. It is the weight coefficient vector of the combined prediction model. It is the error information matrix between the predicted values ​​and the actual values ​​of the combined prediction model. for The array.

6. A method for early warning of safe operation of a tunnel boring machine main unit disassembly and assembly device according to claim 1, characterized in that, In step 3, the optimal combination prediction model is obtained as follows: Based on the predicted safety state coefficient of each combined prediction model at the current time and the actual safety state coefficient at that time, calculate the four indicators corresponding to each combined prediction model: correlation coefficient, mean absolute error, mean relative error, and root mean square error. Compare the above four indicators corresponding to each combined prediction model to obtain the optimal combined prediction model.

7. A safety operation early warning system for the disassembly and assembly device of a tunnel boring machine, characterized in that, include: The data preprocessing module is used to preprocess the off-center load angle data, tension data of the two sets of wire ropes, lifting height data, and pressure data of the hydraulic cylinder group during the hoisting and lifting process of the tunnel boring machine main unit. The first data fusion module is used to fuse the preprocessed data using a combination of the analytic hierarchy process and the entropy weight method to obtain a safety status coefficient. Based on the safety status coefficient, the module compares the operational alarm level and risk level of the disassembly and assembly device to obtain the safety status of the disassembly and assembly device at this time. The second data fusion module is used to take the fused safety status coefficient as input, establish different types of prediction models, and predict the safety status coefficient of the disassembly and assembly device in the next stage. Simultaneously, multiple combined prediction models based on the minimum squared error and least squares method are established according to the mutual combination of different types of prediction models. By comparing the safety state coefficient predicted by the combined prediction model with the actual value, the optimal combined prediction model with the best prediction effect is selected as the safety early warning model. The early warning module is used to predict the safety status coefficient of the disassembly and assembly device based on the safety early warning model, and to compare the safety status coefficient with the operation alarm level and risk level of the disassembly and assembly device to determine whether the disassembly and assembly device should continue to operate. If the predicted safety status coefficient exceeds the safety range, an alarm is issued and the operation of the disassembly and assembly device is stopped.

8. A safety operation early warning system for the disassembly and assembly device of a tunnel boring machine according to claim 7, characterized in that, The types of prediction models established in the second data fusion module include: GM, ARIMA, and LSTM.

9. A safety operation early warning system for the disassembly and assembly device of a tunnel boring machine according to claim 7, characterized in that, In the second data fusion module, a combined prediction model is established as follows: in Indicates the first The predicted value of the time-combined prediction model; n represents the total number of prediction models in the combined prediction model; Indicates the first The predicted safety state coefficient of the i-th prediction model at time i; The weights assigned to the i-th prediction model, and ; The weights assigned to each type of prediction model in the combined prediction model are determined as follows: in This is the sum of squared errors between the predicted and actual values ​​of the combined prediction model. It is the weight coefficient vector of the combined prediction model. It is the error information matrix between the predicted values ​​and the actual values ​​of the combined prediction model. for The array.

10. A method for early warning of safe operation of a tunnel boring machine main unit disassembly and assembly device according to claim 7, characterized in that, In the second data fusion module, the optimal combined prediction model is obtained in the following manner: Based on the predicted safety state coefficient of each combined prediction model at the current time and the actual safety state coefficient at that time, calculate the four indicators corresponding to each combined prediction model: correlation coefficient, mean absolute error, mean relative error, and root mean square error. Compare the above four indicators corresponding to each combined prediction model to obtain the optimal combined prediction model.

Images