Shield control method and system, cloud control platform and computer readable storage medium

Abstract

The present disclosure relates to a shield control method and system, a cloud control platform and a computer readable storage medium. The method comprises: the cloud control platform receiving real-time data collected by an edge controller in a shield tunneling process; the cloud control platform training a shield intelligent model according to the real-time data to determine optimal tunneling parameters; and the cloud control platform sharing the optimal tunneling parameters to the edge controller in real time to realize automatic adjustment of parameters and closed-loop control in the shield tunneling process. The present disclosure can improve the control efficiency and quality of shield equipment, while reducing complexity and realizing automatic decision-making function.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the field of tunnel shield, in particular to a shield control method and system, a cloud control platform and a computer readable storage medium. BACKGROUND

[0002] The shield machine has a complex structure, a large number of coupling, collaborative control and protection logics between multiple systems, a long-distance and decentralized distribution of sensors and actuators, and high requirements for data real-time performance and reliability. SUMMARY

[0003] The inventors have found that in the related art, earth pressure shield control still relies on the construction experience of a shield driver, is seriously affected by human factors, has low control accuracy and tunneling efficiency, and improper control can cause ground collapse or uplift, even personnel casualties; the IT+PLC architecture of the related art realizes control over shield equipment, the programmable logic controller (PLC) is limited by proprietary connections, dedicated software and license costs, has weak support for advanced networks and security functions, lacks support for IT (Information Technology)-centered programming languages and protocols (such as HTTPS (Hypertext Transfer Protocol Secure) and MQTT (Message Queuing Telemetry Transport)), programming and maintenance are relatively cumbersome, which is not conducive to centralized and unified control of the shield multi-device control system, cannot achieve data connection and sharing, and cannot meet the demand for intelligent control of the shield.

[0004] In view of at least one of the above technical problems, the present disclosure provides a shield control method and system, a cloud control platform and a computer readable storage medium, which can improve the control efficiency and quality of the shield equipment, while reducing complexity and realizing automatic decision-making functions.

[0005] According to one aspect of the present disclosure, a shield control method is provided, comprising:

[0006] The cloud control platform receives real-time data collected by the edge controller during shield tunneling;

[0007] The cloud control platform trains a shield intelligent model according to the real-time data, and determines optimal tunneling parameters;

[0008] The cloud control platform shares the optimal tunneling parameters to the edge controller in real time, to realize automatic parameter adjustment and closed-loop control during shield tunneling.

[0009] In some embodiments of the present disclosure, the tunneling parameters include at least one of a cabin pressure, a cutterhead thrust, a propulsion speed, a recommended pressure, a cutterhead rotation speed, a grouting amount, a total thrust, an auger rotation speed, and a cutterhead torque.

[0010] In some embodiments of the present disclosure, the cloud control platform trains a shield intelligent model according to the real-time data, and determining the optimal tunneling parameters includes:

[0011] A support vector machine-based sliding mode control is adopted, which combines the support vector machine and the sliding mode control, and adjusts the parameters in the sliding mode control system online by using the support vector machine.

[0012] In some embodiments of the present disclosure, the shield intelligent model includes a sliding mode controller and a support vector machine-based sliding mode controller.

[0013] The cloud control platform trains a shield intelligent model according to the real-time data, and determining the optimal tunneling parameters includes:

[0014] In the first stage, a sliding mode controller output is used to control the shield tunneling.

[0015] In the second stage, a support vector machine-based sliding mode controller acquires new sample data by interacting with the shield system, obtains a new training sample set by effectively screening and determining the new sample data, learns online by using the support vector machine-based sliding mode controller to correct the control algorithm in real time, obtains the optimal parameters of the shield tunneling process, and transmits the optimized parameters to an edge controller in real time to realize real-time correction of the shield actuator.

[0016] In some embodiments of the present disclosure, the cloud control platform trains a shield intelligent model according to the real-time data, and determining the optimal tunneling parameters further includes:

[0017] In the first stage, a support vector machine-based sliding mode controller is in a learning stage, and data is effectively screened to be added to a support vector machine training sample set, and the support vector machine-based sliding mode controller obtains the structure and preliminary parameters of the controller by learning.

[0018] When the support vector machine-based sliding mode controller has an approximation error for the sliding mode controller that is less than a set threshold value, the shield tunneling system automatically switches to the support vector machine-based sliding mode controller control by a predetermined learning time.

[0019] In some embodiments of the present disclosure, the cloud control platform supports a support vector machine sliding mode controller to obtain new sample data by interacting with the shield system, to obtain a new training sample set by effectively screening and determining the new sample data, to obtain optimal parameters of the shield tunneling process by online learning and real-time correction of the control algorithm of the support vector machine sliding mode controller, and to realize real-time correction of the shield actuator by transmitting the optimized parameters to the edge controller in real time, including:

[0020] receiving new shield data and constructing a new data pair;

[0021] judging the effectiveness of the data after a set period, and if the data is effective, adding the data to the training sample set;

[0022] using an incremental support vector machine algorithm for online training;

[0023] calculating the output of the support vector machine sliding mode controller;

[0024] calculating the control disturbance value according to the normal distribution with the output of the support vector machine sliding mode controller as the mean value;

[0025] using the output of the support vector machine sliding mode controller plus the disturbance value as the sealing cabin pressure reference value to control the soil cabin pressure object.

[0026] According to another aspect of the present disclosure, a cloud control platform is provided, comprising:

[0027] a data receiving module configured to receive real-time data collected by an edge controller during shield tunneling;

[0028] a parameter determination module configured to train a shield intelligent model according to the real-time data and determine optimal tunneling parameters;

[0029] a parameter sharing module configured to share the optimal tunneling parameters to the edge controller in real time to realize the functions of automatic parameter adjustment and closed-loop control during shield tunneling.

[0030] In some embodiments of the present disclosure, the tunneling parameters include at least one of the sealing cabin pressure, the cutter head thrust, the propulsion speed, the recommended pressure, the cutter head speed, the grouting amount, the total thrust, the screw conveyor speed, and the cutter head torque.

[0031] In some embodiments of the present disclosure, the parameter determination module is configured to use support vector machine-based sliding mode control to combine support vector machines and sliding mode control and adjust the parameters in the sliding mode control system online using support vector machines.

[0032] In some embodiments of the present disclosure, the parameter determination module includes a sliding mode controller and a support vector machine sliding mode controller;

[0033] a sliding mode controller configured to output a shield tunneling control in a first stage;

[0034] the support vector machine sliding mode controller is configured to, in a second stage, obtain new sample data by interacting with the shield system, obtain a new training sample set by effectively screening and determining the new sample data, learn an online control algorithm in real time by using the support vector machine sliding mode controller, obtain optimal parameters of the shield tunneling process, and transmit the optimal parameters to the edge controller in real time to realize real-time correction of the shield actuator.

[0035] In some embodiments of the present disclosure, the support vector machine sliding mode controller is further configured to, in the first stage, be in a learning stage, add support vector machine training samples to a set by effective data screening, and obtain a structure and preliminary parameters of the controller by learning.

[0036] In some embodiments of the present disclosure, the support vector machine sliding mode controller is configured to receive new shield data, construct a new data pair, judge data effectiveness in a set period, add a training sample set if the data is effective, perform online training using an incremental support vector machine algorithm, calculate an output of the support vector machine sliding mode controller, calculate a control disturbance value according to a normal distribution with the output of the support vector machine sliding mode controller as a mean value, and control a soil cabin pressure object with the output of the support vector machine sliding mode controller plus the disturbance value as a sealing cabin pressure reference value.

[0037] According to another aspect of the present disclosure, a cloud control platform is provided, comprising:

[0038] a memory for storing instructions;

[0039] a processor for executing the instructions, so that the cloud control platform performs operations for implementing the shield control method according to any one of the above embodiments.

[0040] According to another aspect of the present disclosure, a shield control system is provided, comprising an information technology device and an operation technology device, wherein the information technology device comprises a cloud control platform according to any one of the above embodiments, and the operation technology device comprises an edge controller.

[0041] In some embodiments of the present disclosure, the edge controller is configured to collect input and output signals of shield control devices and process data downward, and realize data sharing between the data center and the cloud control platform upward.

[0042] In some embodiments of the present disclosure, the edge controller is configured to set a predetermined filtering condition for the collected data, wherein the filtering condition is to store data in a normal tunneling process or to store tunneling data with an amplitude less than a predetermined set value.

[0043] In some embodiments of the present disclosure, the information technology device further comprises at least one of a cloud gateway, a data center, a mobile terminal and a ground monitoring device.

[0044] In some embodiments of the present disclosure, the operation technology device further comprises at least one of a distributed input / output controller, an upper computer, a guidance system, an intelligent data acquisition device and a video monitoring system.

[0045] According to another aspect of the present disclosure, a computer readable storage medium is provided, wherein the computer readable storage medium stores computer instructions, and the instructions are executed by a processor to implement the shield control method according to any one of the above embodiments.

[0046] The present disclosure can improve the control efficiency and quality of the shield device, while reducing complexity and realizing automatic decision-making functions. BRIEF DESCRIPTION OF DRAWINGS

[0047] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure or the prior art, the drawings needed in the description of the embodiments or the prior art will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present disclosure, and other drawings can be obtained by those skilled in the art without creative labor.

[0048] Figure 1 A schematic diagram of some embodiments of the shield control system of the present disclosure.

[0049] Figure 2 A schematic diagram of some other embodiments of the shield control system of the present disclosure.

[0050] Figure 3 A schematic diagram of some embodiments of the shield control method of the present disclosure.

[0051] Figure 4 A structure diagram of SVM in some embodiments of the present disclosure.

[0052] Figure 5 A schematic diagram of some other embodiments of the shield control method of the present disclosure.

[0053] Figure 6 A schematic diagram of some embodiments of the cloud control platform of the present disclosure.

[0054] Figure 7 A structure schematic diagram of some other embodiments of the cloud control platform of the present disclosure. DETAILED DESCRIPTION

[0055] The technical solutions in the embodiments of the present disclosure will be described clearly and completely below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only part of the embodiments of the present disclosure, rather than all the embodiments. The description of the at least one exemplary embodiment is actually only illustrative, but not as any limitation on the present disclosure and its application or use. Based on the embodiments in the present disclosure, all other embodiments obtained by those of ordinary skill in the art without creative work are within the scope of protection of the present disclosure.

[0056] Unless otherwise specified, the relative arrangement, numerical expressions and values of the components and steps set forth in these embodiments do not limit the scope of the present disclosure.

[0057] It should be understood that the size of each part shown in the drawings is not drawn in accordance with the actual proportional relationship for the convenience of description.

[0058] The technology, methods and devices known to those of ordinary skill in the relevant art can not be discussed in detail, but should be considered as part of the authorized description when appropriate.

[0059] In all examples shown and discussed herein, any specific values should be interpreted as merely exemplary and not as a limitation. Thus, other examples of the exemplary embodiments can have different values.

[0060] It should be noted that similar reference numbers and letters represent similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further discussed in subsequent drawings.

[0061] Figure 1 A schematic diagram of some embodiments of the shield control system of the present disclosure. The shield control system of the present disclosure includes information technology (IT) and OT (Operation Technology) devices, wherein the information technology devices include a cloud control platform 10, and the OT devices include an edge controller 5, wherein:

[0062] The edge controller 5 is configured to collect input and output signals of shield control devices and process data downward, and to realize data sharing between the data center and the cloud control platform upward.

[0063] The cloud control platform 10 is configured to receive real-time data collected by the edge controller 5 during the shield tunneling process, train a shield intelligent model according to the real-time data, determine optimal tunneling parameters, and share the optimal tunneling parameters to the edge controller 5 in real time to realize the parameter automatic adjustment function and closed-loop control during the shield tunneling process.

[0064] In some embodiments of the present disclosure, the edge controller 5 can also be configured to set a predetermined filtering condition for the collected data, wherein the filtering condition is to store the data during the normal tunneling process or to store the tunneling data with an amplitude less than a predetermined set value.

[0065] The inventors have found through research that the shield control network in the related art will develop in the direction of integration and fusion, and the mechanical and electrical systems, video monitoring systems, etc. will be organically connected; for shield construction, the connection of equipment is the foundation, data collection and analysis are the key means, and using the information obtained by analysis to make the best decision to optimize production and operation is the ultimate goal.

[0066] The present disclosure proposes a shield intelligent control system based on an edge controller, which integrates PLC controllers, PCs, gateways, motion control, I / O data collection, fieldbus protocols, machine vision, device networking, etc. in one, aiming to better coordinate data processing and application deployment, realize edge autonomy in the case of time delay sensitivity and limited bandwidth, and effectively reduce the delay between data production and decision-making through the deep integration of edge computing platforms and cloud platforms, forming a full-level open architecture of edge + cloud, completing shield intelligent control on the edge side, driving the big data processing of non-real-time and long-period data of shield complex working conditions on the cloud side, model training, algorithm updating, etc., realizing remote monitoring, diagnosis, analysis, guidance, and data sharing of intelligent optimization control strategies of shield tunneling equipment, forming cloud-edge-end collaboration, and meeting the application requirements of intelligentization, flexible deployment, safety and reliability, etc.

[0067] Figure 2 A schematic diagram of another embodiment of the shield control system of the present disclosure. The shield control system of the present disclosure includes information technology equipment and operation technology equipment, wherein the information technology equipment includes at least one of a cloud control platform 10, a cloud gateway 7, a data center 8, a mobile terminal 9, and a ground monitoring device 11, and the operation technology equipment includes at least one of an edge controller 5, a distributed input / output controller 1, an upper computer 6, a guidance system 4, an intelligent data collection device 2, and a video monitoring system 3, wherein:

[0068] The cloud control platform 10 can be an intelligent cloud control platform.

[0069] The mobile terminal 9 can be an application APP of a mobile terminal.

[0070] In some embodiments of the present disclosure, as shown in Figure 2 The distributed I / O controller 1 is connected to the edge controller 5 for device data acquisition and control; the edge controller 5 is connected to the intelligent cloud control platform 10 through the optical fiber for data transmission; the intelligent cloud control platform 10 schedules data of the data center 8, generates control parameters through deep learning training of the control model, and recommends the control parameters to the edge controller 5; the host computer 6, the guidance system 4, etc. are connected to the edge controller for shield state display and data monitoring; the intelligent data acquisition device 2 includes intelligent Internet of Things sensors and intelligent video cameras which access the edge controller 5 through wireless control devices to access unstructured data; and the video monitoring system 3 is connected to the edge controller 5.

[0071] The shield control system based on the edge controller utilizes the latest IT communication and Internet of Things technologies, while retaining the advantages of PLC / PAC (Programmable Automation Controller) in the OT aspect. In the process of intelligent shield construction, the edge controller can collect I / O (Input / Output) signals of shield control devices and process data downward, and communicate with other control devices; and can share data with IT data centers and cloud servers upward.

[0072] The edge controller of the present disclosure can provide a complete connection solution for shield informatization, including acting as a TCP / IP (Transmission Control Protocol / Internet Protocol), OPC (Object Linking and Embedding for Process Control), or MQTT (Message Queuing Telemetry Transport) server, while the servers in the related art PLC communication application are external devices.

[0073] Meanwhile, the edge controller of the present disclosure can store data, pre-process data, respond to local requests, and forward standardized data to central storage, thus reducing the demand for central networks and servers.

[0074] Compared with flat file storage, the present disclosure also improves the flexibility and response capability at the process level. The store-and-forward technology can also establish fault tolerance in the case of network stability problems.

[0075] In addition, the edge controller of the present disclosure embeds higher security standards, requires user authentication and supports multiple access levels. However, because they are network-oriented, they include standard protection from the network, such as an internal firewall to prevent unauthorized access, Ethernet interfaces and SSL (Secure Sockets Layer) / TLS (Transport Layer Security) encryption and authentication to isolate trusted and untrusted traffic.

[0076] The present disclosure utilizes edge computing technology to realize downward execution of mechanical equipment data collection, detection, control and other tasks, upward connection to a cloud platform, deep integration of IT and OT of a shield control system under high-speed series connection, opening up multiple joints of intelligent manufacturing, and realization of shield intelligentization and flexibility.

[0077] (1) The present disclosure takes the edge controller as the core, utilizes edge control technology, deeply integrates IPC (Industrial Personal Computer) and PLC, and greatly improves the operation efficiency of the shield. The PLC+IPC scheme of the related art is connected through a network cable, and the shield server peripheral is on a separate computer. Many uncertain factors can cause connection interruption, such as loosening of connectors or electromagnetic interference on site. The edge controller of the present disclosure integrates the IPC and the PLC into one controller, even in one CPU (central processing unit), while maintaining independent operation by using a split-core system technology. The hardware stability of the entire system is greatly improved.

[0078] (2) The shield intelligent control system based on the edge controller of the present disclosure has sufficient interfaces to connect various heterogeneous execution components on the shield, supports multiple protocols, and in addition to conventional sensors, supports wired, WIFI, 5G and other transmission modes to connect image, sound, video, signal and other intelligent Internet of Things sensing devices that release a large amount of information data of the shield, so as to integrate different manufacturers and different execution components in the production site, open up network connection, and solve the problem of multi-manufacturer integration.

[0079] (3) The edge controller of the present disclosure can meet various functions such as shield process control, I / O acquisition, wireless transmission, remote operation and maintenance, machine vision, database and cloud computing based on the IEC (International Electrotechnical Commission) 61131-3 international standard programming method, reduce the difficulty, time cost and data cost of programming and maintenance, and realize interchangeability and reusability. At the same time, the collection, storage, analysis and calculation of real-time data and IIoT (Industrial Internet of Things) sensing signals are completed, and IT and OT are seamlessly integrated to meet the needs of reducing manual work and improving equipment intelligence.

[0080] (4) The edge controller of the present disclosure can run a custom application and run a database server locally to realize storage, combination, sharing and protection of process data and fast processing, and reduce network latency. In the era of big data, cloud analysis, machine learning and IoT, as the data traffic becomes larger and larger, the edge controller can standardize the additional work of data, improve the efficiency of network transmission, reduce the running burden of the central server, and avoid the complexity and inefficiency of the shield control system.

[0081] The present disclosure uses the edge controller as a bridge to develop and utilize a large number of brownfield equipment of the shield, releases the value of shield data through protocol conversion to standard protocols and information models; the edge computing node of the present disclosure effectively filters shield data, better cooperates data processing and application deployment, reduces system latency, improves data transmission efficiency, and realizes edge autonomy; the big data center of the present disclosure obtains shield equipment information and tunneling data through the edge controller, realizes storage, combination, sharing and protection of data, and provides sufficient and reliable data sources for data analysis and decision-making; the present disclosure introduces advanced algorithms such as big data and machine learning to provide intelligent perception and decision-making capabilities for shield tunneling, realizes autonomous learning and upgrading algorithms through real-time training of the shield model by the central server, automatically adjusts the running code of the shield tunneling system, and thus makes decisions and implements shield closed-loop control.

[0082] The shield control method and cloud control platform of the present disclosure are described below through specific embodiments.

[0083] The inventor finds through research that the shield tunneling process is an extremely complex process and is disturbed by various internal and external environments, so there are many factors affecting the earth pressure in the sealed cabin. The sizes of the shield thrust, the advancing speed and the cutter head rotating speed can directly affect the earth intake of the sealed cabin, and the rotating speed of the screw conveyor is related to the earth discharge, so the site construction adopts the way of controlling the amount of earth discharge to indirectly control the earth pressure in the sealed cabin, so as to keep the earth pressure in the sealed cabin and the water and earth pressure of the excavation face in dynamic balance. According to different geological conditions, the earth pressure in the sealed cabin is set, and the real-time collected earth pressure in the sealed cabin is compared with the set value, and the absolute value of the difference between the two is minimized to adjust the rotating speed of the screw conveyor and the advancing speed, so as to ensure the balance of the earth pressure in the sealed cabin.

[0084] In the shield tunneling process, the thrust, the advancing speed and the rotating speed of the screw conveyor have important influence on the change of the earth pressure in the sealed cabin. According to the earth intake and discharge balance theory of the sealed cabin in the shield tunneling process, the continuous equation of the complex operations such as the earth entering the soil cabin after being excavated from the soil body and being discharged by the screw conveyor and the earth being compressed is established in the shield construction process.

[0085] The earth intake Q of the sealed cabin i As shown in formula (1):

[0086] Q i = πR 2 v = S v (1)

[0087] In formula (1), R is the radius of the cutter head; v is the shield tunneling speed; S is the cutter head area of the cutting surface.

[0088] The earth discharge Q0 of the screw conveyor is as shown in formula (2):

[0089]

[0090] In formula (2), η is the earth discharge efficiency, A is the effective contact area of the screw conveyor; T is the rotating screw pitch; n s is the rotating speed of the screw conveyor, r1 is the radius of the screw conveyor, and r2 is the radius of the screw conveyor shaft.

[0091] And the flow continuity equation in the sealed cabin of the earth pressure balance shield is as shown in formula (3):

[0092]

[0093] In formula (3), ce is the leakage coefficient outside the sealed cabin; p is the earth pressure in the sealed cabin; p0 is the leakage earth pressure outside the sealed cabin; V e is the volume of the sealed cabin; β e is the effective compression coefficient of the earth material, liquid and gas in the sealed cabin.

[0094] In the process of shield construction, the propulsion cylinder provides propulsion power, and the shield itself is taken as the research object. Its mechanical equilibrium equation is shown in equation (4):

[0095] F-(f+P)=ma(4)

[0096] In equation (4), F is the total thrust of the shield, f is the total resistance when propelling, P is the earth pressure on the cutter head panel, m is the mass of the shield, and a is the acceleration of the shield propulsion.

[0097] Considering that the excavation speed is very slow in the actual shield process, it can be approximated as uniform motion, i.e. a = 0, so the above equation becomes equation (5):

[0098] F=(f+P)(5)

[0099] For a panel-type shield, under normal shield conditions, the earth pressure on the cutter head panel and the earth pressure in the sealed cabin have a relationship as shown in equation (6):

[0100]

[0101] In equation (6), D is the diameter of the cutter head, δ is the pressure addition value on the panel, and λ is the opening rate.

[0102] The above equations can be arranged to obtain the relationship model between the earth pressure in the sealed cabin and the thrust, propulsion speed, and screw conveyor speed as shown in equations (7) and (8):

[0103]

[0104]

[0105] In addition, during the construction process of the cutter head, the original earth pressure on the excavation face is the initial earth pressure. In fact, the earth pressure on the excavation face is constantly changing. For the earth pressure on the cutter head panel, it is shown in equation (9):

[0106]

[0107] In equation (9), F(p') is the state factor of the earth pressure along the p' axis; F(q') is the state factor of the earth pressure along the q' axis; a and b are constants v d is the cutter head speed; r is the radius of the calculation point on the cutter head; K is the lateral earth pressure coefficient; H c is the distance from the ground to the shield axis; θ is the relevant measurement angle; γ is the unit weight of the soil.

[0108] According to the relationship between the earth pressure on the cutter head panel and the earth pressure in the sealed cabin in the above equation, the relationship between the earth pressure in the sealed cabin and the cutter head speed is obtained as shown in equation (10):

[0109]

[0110] The cutterhead torque is one of the key parameters to ensure the normal and safe advancement of the shield, which is determined by the contact pressure between the shield and the excavated soil. An increase in the cutterhead torque will also cause a corresponding increase in the sealing chamber pressure. The empirical formula for the sealing chamber pressure and the cutterhead torque is as formula (11):

[0111] M = K c pv d (11)

[0112] In formula (11), M is the cutterhead torque, K c is a parameter associated with the cutterhead.

[0113] The sealing chamber pressure of the shield has a nonlinear dependence on the total thrust of the shield, the advancement speed, the screw conveyor speed, the cutterhead speed, and the cutterhead torque. However, the shield tunneling process is a complex, nonlinear, multivariable, and strongly coupled industrial object. It is difficult to establish an accurate and reliable soil pressure prediction model based on the mathematical relationship derived from mechanism analysis. Therefore, the present disclosure uses statistical theory to establish a nonlinear prediction model for the soil pressure chamber. The nonlinear prediction model of the present disclosure is a whole prediction model for the shield machine. To analyze the sealing chamber pressure, the output of the model can be understood as the sealing chamber pressure. If the cutterhead thrust is analyzed, the output of the model can also be understood as the cutterhead thrust.

[0114] Figure 3 FIG. 1 is a schematic diagram of some embodiments of the shield control method of the present disclosure. Preferably, the embodiments can be executed by the shield control system of the present disclosure or the cloud control platform of the present disclosure. Figure 3 The method of the embodiments can include at least one of steps 31 to 33, wherein:

[0115] Step 31, the cloud control platform receives real-time data collected by the edge controller during the shield tunneling process.

[0116] Step 32, the cloud control platform trains the shield intelligent model according to the real-time data to determine the optimal tunneling parameters.

[0117] In some embodiments of the present disclosure, the tunneling parameters can include at least one of the sealing chamber pressure, the cutterhead thrust, the advancement speed, the recommended pressure, the cutterhead speed, the grouting amount, the total thrust, the screw conveyor speed, and the cutterhead torque.

[0118] In some embodiments of the present disclosure, step 32 can include using support vector machine-based sliding mode control, combining support vector machines and sliding mode control, and adjusting the parameters in the sliding mode control system online using support vector machines.

[0119] In some embodiments of this disclosure, SVM (Support Vector Machine) is a novel learning machine based on statistical learning theory and the principle of structural risk minimization. Its training algorithm does not suffer from local minima, exhibits strong generalization ability, can automatically design model complexity, and avoids the curse of dimensionality. RBF (Radial Basis Function) is applicable to various situations, including sample size and dimensionality, and has a wide convergence region. Therefore, this disclosure selects an RBF kernel function that satisfies Mercer's theorem, as shown in formula (12):

[0120]

[0121] In formula (12), the center is the support vector and σ is the width of the Gaussian function.

[0122] Figure 4 This is a structural diagram of the SVM in some embodiments of this disclosure. For example... Figure 4 As shown, for any input in the training samples, the SVM system will produce a corresponding output value. Choosing model parameters: The kernel function largely determines the structure of the feature space and the type and complexity of the support vector regression machine. Changes in the kernel parameters indirectly alter the low-dimensional to high-dimensional feature mapping, and also affect the dimensionality and complexity of the feature subspace.

[0123] As the most commonly used parameter selection method, cross-validation's basic idea is to divide the initial samples into groups: one group is used for training to build the model, and the other is used for testing and evaluating the model. K-CV cross-validation involves dividing the data into k equal parts, performing a test on each part, and using the remaining k-1 parts for training. This cross-validation is repeated k times, and the root mean square error (RMSE) of the k cross-validations is then averaged as the best parameters for the regression machine. The value of k is generally an integer greater than or equal to 2, but in practice, it is mostly chosen starting from 3, and only when the original data is small is 2 attempted. This K-CV cross-validation method effectively avoids overlearning and underlearning, increasing the reliability of the results. The penalty factor C is used to adjust the ratio between the confidence interval and the empirical risk; the smaller the value, the greater the risk, and vice versa.

[0124] The present disclosure can utilize cross-validation method for cyclic search, reduce step size setting in reasonable parameter range, and improve accuracy. In support vector regression machine, cross-validation returns mean square error MSE, if there are multiple sets of parameters C and gamma corresponding to the highest MSE, the set of C and g with the smallest penalty parameter is selected, because higher C will cause over-learning, affecting the generalization ability of SVM. The SVM model has two very important parameters C and gamma. C is the penalty coefficient, that is, the tolerance of error. Gamma is a parameter of the RBF function selected as the kernel. The larger the gamma, the higher the dimension of the mapping, the better the training result, and the fewer the support vectors, but it is more likely to cause overfitting, that is, low generalization ability. The smaller the gamma value, the more support vectors, which will definitely affect the training and prediction speed, and also affect the accuracy, that is, the underfitting phenomenon.

[0125] In some embodiments of the present disclosure, sliding mode control is a nonlinear control that can overcome the instability of the system, has strong robustness and adaptability to external disturbances and parameter perturbations, has simple algorithm and fast response time, and the system structure changes with time. The switching characteristic forces the system to vibrate up and down along the state trajectory with small amplitude and high frequency.

[0126] Generally, in the system x∈R n , there is a hyper-surface s(x)=s(x1,x2,…,x n )=0 that divides the system state space into two parts s>0 and s<0. When the state point reaches the sliding surface, the system state trajectory will be maintained on the sliding surface, and the system state satisfies formula (13)

[0127]

[0128] Then formula (14) is obtained

[0129]

[0130] The control amount u(x) is solved from the above equation, which is regarded as the equivalent control amount applied by the system on the switching surface s(x)=0.

[0131] The shield intelligent control method of the present disclosure adopts sliding mode control based on support vector machine, combines support vector machine and sliding mode control, adjusts the parameters in the sliding mode control system online by using support vector machine, overcomes the limitation of pre-setting the approaching law parameters in the conventional sliding mode control, and improves the system control quality.

[0132] The current sealing cabin earth pressure signal is collected in real time by the pressure sensor, the optimal earth pressure value range of the current ring is automatically calculated according to the shield machine equipment information and the tunneling geological survey data, and the optimal earth pressure value range is taken as the following target input controller, and the sliding mode variable is obtained according to the sealing cabin pressure value deviation.

[0133] In some embodiments of the present disclosure, the shield intelligent model comprises a sliding mode controller and a support vector machine sliding mode controller.

[0134] In some embodiments of the present disclosure, the sliding mode controller is abbreviated as SMC controller, and the support vector machine sliding mode controller is abbreviated as SVM-SMC controller.

[0135] In some embodiments of the present disclosure, step 32 can comprise at least one of step 321 and step 322, the self-learning control method is realized by an edge controller-based control system, and the optimization process comprises two stages of step 321 and step 322, wherein:

[0136] In step 321, in the first stage, the sliding mode controller output is used to control the shield tunneling.

[0137] In some embodiments of the present disclosure, step 321 can further comprise that in the first stage, the support vector machine sliding mode controller is in a learning stage, the support vector machine training sample set is added after effective data screening, the support vector machine sliding mode controller obtains the structure and preliminary parameters of the controller through learning, and when the approximation error of the support vector machine sliding mode controller to the sliding mode controller is less than a set threshold value through a predetermined learning time, the shield tunneling system is automatically switched to the support vector machine sliding mode controller control.

[0138] In some embodiments of the present disclosure, step 321 can comprise that the sliding mode controller is used to control the shield tunneling system object in the early stage of shield tunneling, at this time, the SVM-SMC controller is in a learning stage, the sliding mode controller input and output form the training sample (s i ,p t ) of the SVM-SMC controller, and the training sample set D is added; the input and output of the sliding mode controller are sliding mode variable s i and controller reference sealing cabin earth pressure value p t; the training sample set D is added to the support vector machine learning controller after a set evaluation period and effective judgment is made on the data (the sample added for training should meet the set threshold condition) ; the SVM-SMC controller learns the training sample data according to the objective function by using the support vector machine learning algorithm to obtain the model and preliminary parameters of the SVM-SMC controller, wherein the preliminary parameters refer to selecting appropriate controller parameters (proportion, integral time and differential time) for an already designed and installed control system to improve the steady state and dynamic characteristics of the system and to make the transition process of the system meet the most satisfactory quality index requirement.

[0139] In some embodiments of the present disclosure, the threshold condition for determining the learning data is defined from the perspective of improving the stability of the shield tunneling control system and stabilizing the cabin pressure, as shown in formula (15) :

[0140]

[0141] In formula (15), κ is a weight coefficient, ξ is a defined threshold, E{·} represents a statistical average value, p r is the current theoretical pressure of the cabin, F r is the cutterhead thrust under the theoretical pressure, which is used as the optimization index, and the data of the training sample should meet the above condition. At the beginning, the system is in static adjustment, and κ is set to a small value at the early stage, at which time the controller is biased to stabilize the cabin pressure and obtain the preliminary parameters of the structure of the controller, κ can be increased when dynamic adjustment is performed, so that the mechanical fatigue load of the shield tunneling system is reduced when the soil cabin pressure is stabilized.

[0142] Step 322, in the second stage, the support vector machine sliding mode controller obtains new sample data by interacting with the shield system, obtains a new training sample set by effectively screening and determining the new sample data, and uses the support vector machine sliding mode controller to learn and correct the control algorithm in real time to obtain the optimal parameters of the shield tunneling process, and the optimized parameters are transmitted to the edge controller in real time to realize real-time correction of the shield actuator.

[0143] In some embodiments of the present disclosure, step 322 can include: when the learning reaches a certain degree and the approximation error of the SVM-SMC controller to the conventional sliding mode controller is less than a set threshold, the shield tunneling machine training model automatically switches to the SVM-SMC controller, enters a self-learning phase, the SVM-SMC controller obtains new sample data by interacting with the shield system, and obtains a new training sample set (the training sample meets the set threshold condition, is filtered according to the threshold condition, and is added to the training sample set) through effective screening and judgment of the new sample data, uses the SVM-SMC controller to learn online to correct the control algorithm in real time, obtains the optimal parameters of the shield tunneling process, and recommends the optimized tunneling parameters to the edge controller in real time, realizes real-time correction of the shield actuator, and realizes data closed-loop control.

[0144] Figure 5 The figure is a schematic diagram of another embodiment of the shield control method of the present disclosure. Preferably, the embodiment can be executed by the shield control system of the present disclosure or the cloud control platform of the present disclosure. Figure 5 The method (for example, step 322) of the embodiment can include at least one of steps 51 to 57, wherein:

[0145] Step 51, receiving new shield data, constructing new data pair (s i ,p t ).

[0146] Step 52, judging the data validity through a set period, adding the valid data to the training sample set D, or discarding the data back to step 51.

[0147] Step 53, using the incremental SVM algorithm for online training.

[0148] Step 54, calculating the SVM-SMC controller output p t+1 .

[0149] Step 55, taking p t+1 as the mean value, calculating the control disturbance value according to the normal distribution.

[0150] Step 56, taking the SVM-SMC controller output plus the disturbance value as the sealing cabin pressure reference value, controlling the soil cabin pressure object.

[0151] Step 57, waiting for data online update, and returning to step 51.

[0152] The present disclosure continuously adds new data to the support vector machine training sample set D, learns the increased sample in each iteration process, adjusts the SVM-SMC controller parameters online, and realizes automatic update of the shield tunneling parameters.

[0153] In step 33, the cloud control platform shares the optimal tunneling parameters to the edge controller in real time, so as to realize the parameter automatic adjustment function and closed-loop control in the shield tunneling process.

[0154] The support vector machine structure model of the present disclosure has the advantages of simple structure, strong generalization ability and global optimal solution. The online support vector machine does not retrain the samples in the entire sample set, but learns the samples added or reduced in each training process. The model parameters are calculated and adjusted online based on the learning results and parameters of the previous time, so that the calculation difficulty is greatly reduced. While realizing a small time cost, the adaptability of the model is increased, and the model parameters and structure can change with the change of the environment.

[0155] The present disclosure provides a shield intelligent control method, which combines the sliding mode control and the online support vector algorithm, realizes the controller self-learning function by using the online support vector machine algorithm, continuously iteratively optimizes, does not need to identify the object model online, is more suitable for a nonlinear time-varying system, can correct the model parameters online through learning, makes the established model closer to the real situation, and the system has stronger robustness and adaptability, has good dynamic quality, and ensures that the shield control system normally, efficiently and stably operates.

[0156] The present disclosure establishes a shield intelligent model on a cloud control platform, arranges intelligent algorithms, realizes real-time collection of parameters in the shield tunneling process, finds optimal tunneling parameters through model training, shares the optimized data parameters to an edge controller in real time, realizes the parameter automatic adjustment function and closed-loop control in the shield tunneling process. Shield algorithm learning is divided into two parts. In STEP1, the initial learning stage, a conventional sliding mode controller (SMC) is used to output shield tunneling control. After effective data screening, the support vector machine training sample set is added. The support vector machine sliding mode controller (SVM-SMC) obtains the structure and preliminary parameters of the controller through learning. After a certain learning time, when the approximation error of SVM-SMC to SMC is less than the set threshold value, the shield tunneling system automatically switches to SVM-SMC control. In STEP2, the self-learning stage, the SVM-SMC controller obtains new sample data through data interaction with the shield system. Through effective screening and judgment of the new sample data, a new training sample set is obtained. The SVM-SMC controller learns and corrects the control algorithm in real time, obtains the optimal parameters in the shield tunneling process, and transmits the optimized parameters to the edge controller in real time, so as to realize real-time correction of the shield actuator.

[0157] The present disclosure uses edge computing technology to realize downward execution of mechanical equipment data collection, detection, control and other tasks, upward connection of a cloud platform, deep integration of IT and OT of a shield control system under high-speed series connection, opening up of multiple joints of intelligent manufacturing, and realization of shield intelligentization and flexibility.

[0158] The inventors have noticed that with the introduction of smart shields, big data and the industrial internet, network devices with IP addresses in the manufacturing industry are rapidly and widely covering smart factories, and when the production process and information are combined, both OT and IT personnel face the problem of needing to solve shield data access. Thus, the present disclosure provides a new shield tunneling operation mode. OT is considered the backbone of modern smart factories, while IT is essential for all smart enterprises, and the OT team not only focuses on monitoring a single device, but also cooperates with IT to collect and analyze data from all interconnected devices. The cross-functional efforts of the present disclosure provide a more comprehensive perspective for the enterprise, identify inefficient aspects, and achieve the goal of maximizing ROI, such as achieving real-time control and edge computing of shield tunneling, real-time monitoring of shield tunneling status, providing flexible reports, efficient fault alarms, precise energy management control, and through a cloud control platform, data analysis, management and optimization of remote shield assets, etc., providing data support for shield fault early warning, health management, life prediction, etc.

[0159] The edge controller is used as a bridge to develop and utilize a large number of brownfield devices of the shield, and the shield data value is released through protocol conversion to a standard protocol and information model, wherein the shield brownfield devices refer to shield machine vibration amplitude detection, humidity, environmental temperature change, belt machine residue solidification, automatic light adjustment and various frequency electromagnetic interference.

[0160] The present disclosure installs various vendor intelligent devices, such as intelligent cameras for detecting belt machine solidification and vibration sensors for detecting tunneling impact force, and collects data through the powerful protocol conversion and interface capabilities of the edge controller to develop and utilize more shield data.

[0161] The edge computing node of the present disclosure effectively filters shield data, better cooperates with data processing and application deployment, reduces system latency, improves data transmission efficiency, and achieves edge autonomy.

[0162] The edge controller of the present disclosure is a subsystem of the entire control system, and in order to reduce data traffic and reduce network latency, certain filtering conditions are set for data collection in the edge controller, such as storing data during normal tunneling or selecting tunneling data with an amplitude less than a certain set value for storage, and the rest of the data is ignored. For raw data, it is a kind of rough selection.

[0163] The data screening of the edge node of the present disclosure is to collect and store normal tunneling data by setting specific conditions, and provide the data to the data center. The shield learning algorithm uses a certain amount of data that meets the set threshold condition to automatically train and establish a shield machine model, collects stable system data, and the data screening is more detailed, and the shield machine model is also constantly updated. For example, new collected data can be selected to join the training set and remove the same amount of trained data, or according to a certain time, how much data is newly added to remove how much data in the trained data to update the model.

[0164] The big data center of the present disclosure obtains shield equipment information and tunneling data through the edge controller, realizes storage, combination, sharing and protection of data, and provides sufficient and reliable data sources for data analysis and decision-making.

[0165] The present disclosure introduces advanced algorithms such as big data and machine learning to provide intelligent perception decision-making capability for shield tunneling. Through real-time training of the shield model by the central server, autonomous learning and upgrading of the algorithm are realized, and the shield tunneling system operation code is automatically adjusted, so that decisions are made and shield closed-loop control is implemented.

[0166] Figure 6 A schematic diagram of some embodiments of the cloud control platform of the present disclosure. As shown in Figure 6 , the cloud control platform (for example Figure 1 or Figure 2 the cloud control platform 10 of an embodiment) can include a data receiving module 61, a parameter determining module 62, and a parameter sharing module 63, wherein:

[0167] The data receiving module 61 is configured to receive real-time data collected by the edge controller during shield tunneling.

[0168] The parameter determining module 62 is configured to train a shield intelligent model according to the real-time data and determine optimal tunneling parameters.

[0169] In some embodiments of the present disclosure, the tunneling parameters include at least one of cabin pressure, cutterhead thrust, propulsion speed, recommended pressure, cutterhead speed, grouting amount, total thrust, screw conveyor speed, and cutterhead torque.

[0170] In some embodiments of the present disclosure, the parameter determining module 62 can be configured to use support vector machine-based sliding mode control, combine support vector machines and sliding mode control, and use support vector machines to adjust parameters in the sliding mode control system online.

[0171] In some embodiments of the present disclosure, the parameter determining module 62 includes a sliding mode controller and a support vector machine sliding mode controller.

[0172] The sliding mode controller is configured to output a shield tunneling control in the first stage.

[0173] The support vector machine sliding mode controller is configured to, in the second stage, obtain new sample data by interacting with the shield system, obtain a new training sample set by effectively screening and judging the new sample data, learn the control algorithm in real time by using the support vector machine sliding mode controller, obtain optimal parameters of the shield tunneling process, and transfer the optimal parameters to the edge controller in real time to realize real-time correction of the shield actuator.

[0174] In some embodiments of the present disclosure, the support vector machine sliding mode controller can also be configured to, in the first stage, be in a learning stage, add support vector machine training samples to the set by effective data screening, and obtain the structure and preliminary parameters of the controller by learning.

[0175] In some embodiments of the present disclosure, the support vector machine sliding mode controller can be configured to receive new shield data, construct a new data pair, judge the data effectiveness in a set period, add the training sample set if the data is effective, perform online training using an incremental support vector machine algorithm, calculate the output of the support vector machine sliding mode controller, calculate the control disturbance value according to the normal distribution with the output of the support vector machine sliding mode controller as the mean value, and control the soil cabin pressure object with the output of the support vector machine sliding mode controller plus the disturbance value as the sealing cabin pressure reference value.

[0176] The parameter sharing module 63 is configured to share the optimal tunneling parameters to the edge controller in real time to realize the automatic adjustment function and closed-loop control of the parameters in the shield tunneling process.

[0177] The present disclosure uses edge control technology to increase the edge side data preprocessing and storage capacity of the shield equipment, reduce system complexity, reduce the demand for a central server, and improve the control efficiency of the shield.

[0178] The present disclosure combines the sliding mode control and the online support vector machine algorithm by deep integration of the edge computing platform and the cloud platform, realizes the automatic decision function of the shield tunneling system, improves the construction quality and safety, and reduces the labor intensity.

[0179] Figure 7 FIG. 1 is a structural schematic diagram of some embodiments of the cloud control platform of the present disclosure. Figure 7 As shown in FIG. 1, the cloud control platform includes a memory 71 and a processor 72.

[0180] The memory 71 is configured to store instructions, and the processor 72 is coupled to the memory 71 and is configured to execute the method related to the above-mentioned embodiments based on the instructions stored in the memory.

[0181] As shown in Figure 7 The cloud control platform further includes a communication interface 73 for information interaction with other devices. Meanwhile, the cloud control platform further includes a bus 74, and the processor 72, the communication interface 73, and the memory 71 complete communication with each other through the bus 74.

[0182] The memory 71 can include a high-speed RAM memory, and can further include a non-volatile memory, for example, at least one disk memory. The memory 71 can also be a memory array. The memory 71 can also be divided into blocks, and the blocks can be combined into a virtual volume according to a certain rule.

[0183] In addition, the processor 72 can be a central processing unit CPU, or can be an application specific integrated circuit ASIC, or one or more integrated circuits configured to implement the embodiments of the present disclosure.

[0184] According to another aspect of the present disclosure, a computer readable storage medium is provided, wherein the computer readable storage medium stores computer instructions, and the instructions are executed by a processor to implement the shield control method according to any one of the above-mentioned embodiments.

[0185] In some embodiments of the present disclosure, the computer readable storage medium can be a non-transitory computer readable storage medium.

[0186] The present disclosure proposes a shield intelligent control system and method based on edge computing technology, which improves the processing efficiency of a large amount of data in shield tunnel construction, reduces the time delay, reduces the system complexity, reduces the maintenance complexity, provides system stability, improves control accuracy and tunneling efficiency.

[0187] Those skilled in the art should understand that the embodiments of the present disclosure can be provided as a method, an apparatus, or a computer program product. Therefore, the present disclosure can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present disclosure can take the form of a computer program product implemented on one or more computer-usable non-transitory storage media (including, but not limited to, disk storage, CD-ROMs, optical storage, etc.) containing computer-usable program code.

[0188] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart Figure 1 one or more functions specified in the flowchart or multiple flows and / or block or multiple blocks. Figure 1 one or more functions specified in the flowchart or multiple flows and / or block or multiple blocks.

[0189] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions which implement the function specified in the flowchart Figure 1 one or more functions specified in the flowchart or multiple flows and / or block or multiple blocks. Figure 1 one or more functions specified in the flowchart or multiple flows and / or block or multiple blocks.

[0190] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart Figure 1 one or more functions specified in the flowchart or multiple flows and / or block or multiple blocks. Figure 1 one or more functions specified in the flowchart or multiple flows and / or block or multiple blocks.

[0191] The cloud control platform, the data receiving module, the parameter determining module, the parameter sharing module, the cloud gateway, the data center, the edge controller, the distributed input and output controller and the host computer described above can be implemented as a general processor, a programmable logic controller (PLC), a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component or any appropriate combination thereof for executing the functions described in the present application.

[0192] So far, the present disclosure has been described in detail. In order to avoid obscuring the concept of the present disclosure, some details known in the art are not described. Those skilled in the art can fully understand how to implement the technical solutions disclosed herein according to the above description.

[0193] Those skilled in the art can understand that all or part of the steps of the above-mentioned embodiments can be completed by hardware, or can be instructed by a program to complete the related hardware, and the program can be stored in a non-transitory computer readable storage medium. The storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc.

[0194] The description of the present disclosure is given for the purpose of illustration and description, and is not intended to be exhaustive or to limit the present disclosure to the disclosed form. Many modifications and variations will be apparent to those of ordinary skill in the art. The embodiments are chosen and described in order to best explain the principles of the present disclosure and its practical application, and to enable others skilled in the art to understand the present disclosure in order to design various embodiments with various modifications for specific use cases.

Claims

1. A shield control method, comprising: a cloud control platform receiving real-time data collected by an edge controller during a shield tunneling process; the cloud control platform training a shield intelligent model according to the real-time data to determine optimal tunneling parameters; the cloud control platform sharing the optimal tunneling parameters to the edge controller in real time to realize automatic adjustment of parameters and closed-loop control during the shield tunneling process; wherein the shield intelligent model comprises a sliding mode controller and a support vector machine sliding mode controller; the cloud control platform training the shield intelligent model according to the real-time data to determine the optimal tunneling parameters comprises: in a first stage, the shield tunneling is controlled by the sliding mode controller output; in a second stage, the support vector machine sliding mode controller obtains new sample data by interacting with the shield system, and obtains a new training sample set by effectively screening and judging the new sample data; the support vector machine sliding mode controller learns online to correct the control algorithm in real time to obtain the optimal tunneling parameters of the shield tunneling process, and the optimal tunneling parameters are transmitted to the edge controller in real time to realize real-time correction of the shield actuator, including: receiving new shield data to construct a new data pair; judging the data validity after a set period, if valid, adding to the training sample set; using an incremental support vector machine algorithm for online training; calculating the output of the support vector machine sliding mode controller; calculating the control disturbance value according to the normal distribution with the output of the support vector machine sliding mode controller as the mean value; using the output of the support vector machine sliding mode controller plus the disturbance value as the pressure reference value of the sealed cabin to control the soil cabin pressure object. 2.The shield control method of claim 1, wherein: the tunneling parameters comprise at least one of the sealed cabin pressure, the cutter head thrust, the propulsion speed, the recommended pressure, the cutter head speed, the grouting amount, the total thrust, the screw conveyor speed, and the cutter head torque.

3. The shield control method according to claim 1 or 2, wherein the cloud control platform training the shield intelligent model according to the real-time data to determine the optimal tunneling parameters comprises: using support vector machine-based sliding mode control to combine support vector machine and sliding mode control, and using support vector machine to adjust the parameters in the sliding mode control system online.

4. The shield control method according to claim 1 or 2, wherein the cloud control platform training the shield intelligent model according to the real-time data to determine the optimal tunneling parameters further comprises: in the first stage, the support vector machine sliding mode controller is in the learning stage, and the support vector machine training sample set is added after data effective screening, and the support vector machine sliding mode controller obtains the structure and preliminary parameters of the controller by learning; after a predetermined learning time, when the approximation error of the support vector machine sliding mode controller to the sliding mode controller is less than a set threshold, the shield tunneling system automatically switches to the support vector machine sliding mode controller control. 5.A cloud control platform, comprising: a data receiving module configured to receive real-time data collected by an edge controller during a shield tunneling process; a parameter determination module configured to train a shield intelligent model according to the real-time data to determine optimal tunneling parameters; a parameter sharing module configured to share the optimal tunneling parameters to the edge controller in real time to realize automatic adjustment of parameters and closed-loop control during the shield tunneling process. The parameter determination module comprises a sliding mode controller and a support vector machine sliding mode controller. The sliding mode controller is configured to output shield tunneling control in a first stage. The support vector machine sliding mode controller is configured to obtain new sample data by interacting with the shield system, filter and determine the new sample data to obtain a new training sample set, learn the control algorithm in real time to correct the control algorithm, obtain optimal tunneling parameters of the shield tunneling process, and transmit the optimal tunneling parameters to the edge controller in real time to correct the shield actuator in real time. The support vector machine sliding mode controller is configured to receive new shield data, construct a new data pair, judge the data validity after a set period, add the training sample set if the data is valid, perform online training using an incremental support vector machine algorithm, calculate the output of the support vector machine sliding mode controller, calculate the control disturbance value according to the normal distribution with the output of the support vector machine sliding mode controller as the mean value, and control the soil cabin pressure object with the output of the support vector machine sliding mode controller plus the disturbance value as the cabin pressure reference value.

6. The cloud control platform of claim 5, wherein: The tunneling parameters comprise at least one of cabin pressure, cutter head thrust, propulsion speed, recommended pressure, cutter head speed, grouting amount, total thrust, screw conveyor speed, and cutter head torque.

7. The cloud control platform of claim 5 or 6, wherein: The parameter determination module is configured to use support vector machine-based sliding mode control, combine support vector machines and sliding mode control, and adjust the parameters in the sliding mode control system online using support vector machines.

8. The cloud control platform of claim 5 or 6, wherein: The support vector machine sliding mode controller is further configured to be in a learning stage in the first stage, add support vector machine training sample sets after data filtering, obtain the structure and preliminary parameters of the controller by learning, and automatically switch to support vector machine sliding mode controller control when the approximation error of the support vector machine sliding mode controller to the sliding mode controller is less than a set threshold value after a predetermined learning time.

9. A cloud control platform comprising: a memory for storing instructions; a processor for executing the instructions, so that the cloud control platform implements the shield control method of any one of claims 1-4.

10. A shield control system comprising: Information technology equipment and operation technology equipment, wherein the information technology equipment comprises the cloud control platform of any one of claims 5-9, and the operation technology equipment comprises an edge controller.

11. The shield control system of claim 10, wherein: The edge controller is configured to collect input and output signals of shield control equipment and process data downward, and share data with the data center and the cloud control platform upward.

12. The shield control system of claim 10 or 11, wherein: The edge controller is configured to set a predetermined filtering condition for the collected data, wherein the filtering condition is to store data in a normal tunneling process or to store tunneling data with an amplitude less than a predetermined set value.

13. The shield control system according to claim 10 or 11, wherein: The information technology device further comprises at least one of a cloud gateway, a data center, a mobile terminal and a ground monitoring device; The operation technology device further comprises at least one of a distributed input and output controller, an upper computer, a guidance system, an intelligent data acquisition device and a video monitoring system.

14. A computer readable storage medium, wherein, The computer readable storage medium stores computer instructions, and the instructions are executed by the processor to implement the shield control method in any one of claims 1-4.

Images