A large slope tunneling machine adaptive safety control system and method

Abstract

The application discloses a large-gradient tunneling machine self-adaptive safety control system and method, and relates to the technical field of mine engineering mechanical electrical control, which is composed of a central intelligent controller, a multi-dimensional sensing module and an execution control module. The multi-dimensional sensing module collects machine body posture, oil tank liquid level, track and rear support pressure and hydraulic pump vibration temperature signals in real time through sensors. The execution control module realizes accurate regulation and control of tension, walking torque and hydraulic displacement through an electro-hydraulic proportional valve group, an electric proportional driver and a hydraulic displacement adjusting unit. A large-gradient nonlinear dynamics model is established based on posture and pressure data to solve optimal tension and support force, and pre-tightening and pre-jacking are completed by a PWM signal driven execution mechanism. At the same time, adaptive extended Kalman filtering is used to denoise liquid level data, a dynamic liquid surface equation is fitted and a vertical safety distance is calculated, displacement limitation or shutdown protection is implemented according to a grading threshold, and the safety and reliability of large-gradient tunneling operation are improved.

Description

Technical Field

[0001] This invention belongs to the field of electrical control technology for mining machinery, specifically relating to an adaptive safety control system and method for a steep-slope tunneling machine. Background Technology

[0002] With the development of intelligent coal mining and deep underground space, tunneling machines need to operate and travel in inclined shafts with greater slopes.

[0003] The existing electrical control system of tunneling machines has the following defects. First, the anti-slip control is lagging. Traditional methods mostly use open-loop control or a single tilt switch brake. Under steep slopes, the weight component of the tunneling machine body is huge. The existing electrical control system cannot dynamically adjust the track tension and rear support force according to the real-time attitude, resulting in a large starting impact and track slack and suspension, which can lead to slippage accidents. Secondly, the hydraulic anti-cavitation function fails. When the tunneling machine is tilted, the oil flows to the lower end of the oil tank. The traditional level switch, which alarms based on a fixed height, completely fails in the tilted state. This can easily lead to problems such as false alarms of full oil at the high end and false judgments of safety when the low end oil suction port is exposed, which directly causes the hydraulic pump to be damaged by cavitation. Finally, the existing electrical control system also suffers from lagging operation and maintenance, and lacks real-time assessment of the pump group's health status, mostly relying on alarms after a fault occurs, thus failing to achieve predictive maintenance.

[0004] To address the aforementioned problems, this invention proposes an adaptive safety control system and method for tunneling machines with steep gradients. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an adaptive safety control system and method for tunneling machines with steep gradients, which solves the problems of control lag, liquid level misjudgment, and maintenance lag that exist in existing technologies during tunneling machine operation and travel.

[0006] The objective of this invention can be achieved through the following technical solutions: An adaptive safety control system for a tunnel boring machine with a steep gradient, the system comprising: The central intelligent controller is deployed in the main electrical control cabinet of the tunneling machine; The multi-dimensional sensing module, which communicates with the central intelligent controller, includes: Dual-axis tilt sensors detect the attitude data of the tunneling machine body, including pitch and roll angles; An array-type liquid level sensor consists of liquid level detection points distributed at different preset positions in the hydraulic oil tank, which collects liquid level data at multiple points in the oil tank; The tension pressure transmitter detects the pressure signals of the track tensioning cylinder and the rear support cylinder. Vibration and temperature sensors detect vibration and temperature signals from the hydraulic pump. The execution control module, which communicates with the central intelligent controller, includes: Electro-hydraulic proportional valve assembly controls the tension adjustment of the track tensioning cylinder; An electric proportional drive controls the speed and torque output of the travel motor; Hydraulic displacement adjustment unit, used to adjust the displacement of the hydraulic pump; The central intelligent controller interacts with the multi-dimensional sensing module and the execution control module via CAN bus and RS485 bus, respectively. The central intelligent controller has a built-in control algorithm kernel that generates control commands based on real-time data collected by the multi-dimensional sensing module and inputs the control commands to the execution control module.

[0007] As a further aspect of the present invention, the dual-axis tilt sensor transmits attitude data to the central intelligent controller via a CAN bus; The array-type liquid level sensor transmits multi-point liquid level data to the central intelligent controller via RS485 bus; The tension pressure transmitter and vibration temperature sensor transmit pressure, vibration, and temperature signals to the central intelligent controller via RS485 bus, respectively. The electro-hydraulic proportional valve assembly and electro-proportional actuator receive control commands from the central intelligent controller via the CAN bus. The central intelligent controller also communicates with the remote operation and maintenance terminal via Ethernet.

[0008] An adaptive safety control method for a steep gradient tunneling machine, employing the system described above, is characterized in that the method includes: S1 collects real-time attitude data of the tunneling machine, pressure signals of the track tensioning cylinder and rear support cylinder, multi-point liquid level data in the oil tank, vibration signals and temperature signals of the hydraulic pump through the multi-dimensional sensing module. S2, based on attitude data and pressure signals, determines the pitch angle θ1, tilt angle θ2, current tension force F1, and current support force F2 of the tunneling machine, and establishes a large-slope nonlinear dynamic model with θ1, θ2, F1, and F2 as input variables and attitude data change rate as output. S3, based on a large-gradient nonlinear dynamic model, takes the attitude data change rate approaching zero and the track slip rate approaching zero as optimization objectives within the future monitoring cycle, and takes the maximum allowable tension force of the track tensioning cylinder and the maximum allowable support force of the rear support cylinder as constraints. The NMPC nonlinear model predictive control algorithm is used to optimize the rolling time domain and solve for the target tension force and target support force in the current monitoring cycle. S4, the central intelligent controller sends control commands to the electro-hydraulic proportional valve group, which respectively control the track tensioning cylinder and the rear support cylinder to perform pre-tensioning and pre-lifting actions.

[0009] As a further aspect of the present invention, the method further includes: S5, the central intelligent controller obtains the actual tension force of the track tensioning cylinder and the actual support force of the rear support cylinder through the tensioning pressure transmitter; Verify the actual tension force and actual support force against the target tension force and target support force; If the verification is successful, a start command for the travel motor is output to the electro-proportional driver to control the output torque of the travel motor; If the verification fails, return to step S1 and readjust the control commands until the verification is successful. S6, the central intelligent controller processes the multi-point liquid level data collected by the array liquid level sensor, uses the AEKF adaptive extended Kalman filter algorithm for filtering and noise reduction, uses the least squares method to fit the noise-reduced multi-point liquid level data, constructs the dynamic liquid surface equation of the oil, and calculates the Euclidean normal distance from the oil suction port coordinates to the dynamic liquid surface equation in real time, which is denoted as the vertical safety distance. S7, the central intelligent controller determines whether the vertical safety distance is less than the preset safety threshold; If the displacement is less than the safety threshold, the hydraulic pump displacement is limited by the hydraulic displacement adjustment unit. If the current hydraulic pump displacement is greater than or equal to the safety threshold, maintain the current displacement.

[0010] As a further aspect of the present invention, the method for establishing the nonlinear dynamic model for large slopes in S2 and S3 is as follows: Taking the tunneling machine as the research object, the overall mass, center of gravity position and track ground contact size are obtained as basic parameters. The overall mass, center of gravity position and track ground contact size are all determined by the tunneling machine's own control system and are considered as known values. Using pitch angle θ1 and roll angle θ2 as model inputs, gravity is decomposed into a sliding component along the slope direction and a normal force component perpendicular to the slope direction. Using the current tension force F1 and the current support force F2 as the input to the actuator, the force balance equation along the slope direction and the moment balance equation around the track ground contact boundary of the tunneling machine are established. The acceleration of the tunneling machine along the slope and the pitch attitude angular velocity are used as the attitude data change rate to form the output of the nonlinear dynamic model for large slopes. The central intelligent controller combines a large-slope nonlinear dynamic model and uses the NMPC nonlinear model predictive control algorithm in real time to perform rolling time-domain optimization, and solves the target tension force and target support force in the current monitoring cycle. Using the pitch angle θ1, roll angle θ2, current tension force F1, and current support force F2 as the current state variables, the preset prediction time domain and control time domain are extracted. With the optimization objectives of the pitch attitude angular velocity approaching zero and the track slip rate approaching zero within the future monitoring cycle, and with the maximum allowable tension force of the track tensioning cylinder and the maximum allowable support force of the rear support cylinder as constraints, a cost function is constructed, wherein the monitoring cycle is 5 seconds. The optimal solution of the cost function under constraints is solved in real time within each monitoring cycle, resulting in the optimal control sequence that includes the target tension force and target support force from the current moment to each future monitoring cycle. Extract the target tension force and target support force outputs for the current control cycle from the optimal control sequence, and perform rolling optimization again in the next monitoring cycle.

[0011] As a further embodiment of the present invention, in S4 and S5, the control command is a pulse width modulation (PWM) signal. The central intelligent controller controls the valve core opening of the electro-hydraulic proportional valve group by adjusting the duty cycle of the PWM signal, thereby adjusting the actual tensioning force and actual support force of the track tensioning cylinder and the rear support cylinder. Before outputting torque to the walking motor, the central intelligent controller continuously receives feedback from the tension pressure transmitter on the actual tension force and actual support force during the monitoring cycle. When the actual tension force reaches the target tension force and the actual support force reaches the target support force, the central intelligent controller outputs a walking motor start command to the electro-proportional drive.

[0012] As a further aspect of the present invention, in S6, the AEKF adaptive extended Kalman filter algorithm performs real-time filtering processing on multi-point liquid level data to eliminate measurement noise caused by tunneling machine vibration and oil sloshing, and outputs the optimal liquid level estimate value for each liquid level detection point. The least squares method fits the dynamic liquid level equation z=ax+by+c based on the spatial coordinates of each liquid level detection point and the corresponding optimal liquid level estimate, where a, b, and c are all fitting coefficients.

[0013] As a further aspect of the present invention, in S7, the preset security threshold includes a first preset threshold and a second preset threshold, wherein the first preset threshold is greater than the second preset threshold; When the vertical safety distance is less than the first preset threshold and greater than or equal to the second preset threshold, the central intelligent controller outputs a displacement limit command to the hydraulic displacement adjustment unit via the CAN bus to limit the current displacement of the hydraulic pump to within the preset displacement upper limit. When the vertical safety distance is less than the second preset threshold, the central intelligent controller outputs a forced shutdown command to the hydraulic displacement adjustment unit and simultaneously outputs an alarm signal.

[0014] As a further aspect of the present invention, the method further includes a device health management step: The central intelligent controller receives vibration and temperature signals collected by the vibration and temperature sensors in real time, and compares the vibration and temperature signals with preset vibration health thresholds and temperature health thresholds, respectively. When the vibration signal exceeds the vibration health threshold or the temperature signal exceeds the temperature health threshold, the central intelligent controller generates predictive maintenance warning information and sends it to the remote operation and maintenance terminal via Ethernet.

[0015] The beneficial effects of this invention are: This invention enhances the safety and reliability of tunneling machines under steep slope conditions by integrating multi-dimensional perception and intelligent decision-making. Specifically, through the collaborative use of multiple sensors, including dual-axis tilt angle sensors, arrayed liquid level sensors, pressure sensors, and vibration temperature sensors, it identifies potential problems such as machine instability, hydraulic tilting, track loosening, and pump overheating in real time, preventing air suction, leakage, and mechanical overload. The execution module adjusts tension and displacement based on the algorithm kernel, providing precise response with minimal lag, avoiding the risk of human error. The parallel bus architecture ensures stable command transmission, and the remote maintenance interface supports data uploading to the cloud and fault prediction, thereby reducing unplanned downtime and extending the lifespan of the hydraulic and travel systems. This invention employs NMPC rolling optimization to simultaneously incorporate attitude stability and track slip rate into the target, proactively matching tension and support forces under physical constraints to suppress slope instability and slippage; simultaneously, it utilizes AEKF filtering and liquid level fitting to calculate the safe distance of the oil suction port in real time, actively limiting discharge to prevent cavitation and protecting the hydraulic system; a closed-loop verification mechanism ensures execution accuracy, and the overall control combines predictability, robustness, and engineering practicality; This invention introduces gravity component decomposition by using pitch and roll attitude angles as dynamic inputs, enabling the model to adapt to sudden slope changes and avoiding the instability risk caused by model mismatch in traditional linear control under steep slopes. Simultaneously, it sets the dual objectives of zeroing attitude angular velocity and slip ratio within a 5-second monitoring period and solves for the optimal control sequence under constraints, ensuring real-time optimal matching of tension and support forces. PWM pulse width modulation achieves high-frequency, high-precision response of the hydraulic actuator, shortening the convergence time of attitude adjustment. Once the actual force value fully tracks the target, the travel motor is allowed to output torque, fundamentally eliminating the risk of slippage or overturning and improving operational safety and intelligence under extreme conditions. This invention improves the accuracy of level measurement in tunneling machines under severe conditions of strong vibration and oil sloshing by combining the AEKF adaptive extended Kalman filter algorithm with the least squares method. It can output the optimal level estimate for each detection point in real time and dynamically fit the liquid surface equation. Based on this, according to the comparison results of vertical safety distance and dual thresholds, it outputs displacement limit or forced shutdown commands in stages, avoiding the impact of frequent shutdowns on work efficiency and achieving rapid safety protection under extreme dangerous conditions. In addition, the equipment health management steps generate predictive maintenance warnings in advance by comparing vibration, temperature signals and health thresholds in real time, avoiding sudden failures. Attached Figure Description

[0016] The invention will now be further described with reference to the accompanying drawings.

[0017] Figure 1 This is a schematic diagram of the system architecture described in this invention; Figure 2 This is a schematic diagram of the implementation process of the system described in this invention; Figure 3 This is a schematic diagram illustrating the implementation process of the liquid level control algorithm of the system described in this invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] like Figure 1 , Figure 2 , Figure 3 As shown, this application provides an adaptive safety control system and method for a tunneling machine with a large slope. The system installs a central intelligent controller in the main electrical control cabinet of the tunneling machine, and interacts with the multi-dimensional sensing module and the execution control module through the CAN bus and RS485 bus to realize active safety control of the tunneling machine for anti-slip and anti-overturning under the condition of large slope.

[0020] As an embodiment 1 of this application, it specifically includes: The central intelligent controller is deployed in the main electrical control cabinet of the tunneling machine; The multi-dimensional sensing module, which communicates with the central intelligent controller, includes: Dual-axis tilt sensors detect the attitude data of the tunneling machine body, including pitch and roll angles; An array-type liquid level sensor consists of liquid level detection points distributed at different preset positions in the hydraulic oil tank, which collects liquid level data at multiple points in the oil tank; The tension pressure transmitter detects the pressure signals of the track tensioning cylinder and the rear support cylinder. Vibration and temperature sensors detect vibration and temperature signals from the hydraulic pump. The execution control module, which communicates with the central intelligent controller, includes: Electro-hydraulic proportional valve assembly controls the tension adjustment of the track tensioning cylinder; An electric proportional drive controls the speed and torque output of the travel motor; Hydraulic displacement adjustment unit, used to adjust the displacement of the hydraulic pump; The central intelligent controller interacts with the multi-dimensional sensing module and the execution control module via CAN bus and RS485 bus, respectively. The central intelligent controller has a built-in control algorithm kernel that generates control commands based on real-time data collected by the multi-dimensional sensing module and inputs the control commands to the execution control module. Example 2

[0021] This embodiment is based on Embodiment 1 and provides a basic implementation method for an adaptive safety control system and method for a tunneling machine with a large gradient, specifically including the following: In this embodiment, the central intelligent controller adopts an industrial embedded control motherboard with a built-in control algorithm kernel; the central intelligent controller includes a CAN interface, an RS485 interface, and an Ethernet interface, the Ethernet interface being used for communication with a remote operation and maintenance terminal.

[0022] As described in Embodiment 1, the system includes a central intelligent controller, a multi-dimensional sensing module, and an execution control module; The multidimensional sensing module consists of a dual-axis tilt sensor, an array-type liquid level sensor, a tension pressure transmitter, and a vibration temperature sensor. The dual-axis tilt sensor is a dynamic tilt sensor based on the MEMS accelerometer principle. It sends attitude data to the central intelligent controller via the CAN bus at a refresh rate of 100Hz, including 16-bit signed integer values ​​of pitch attitude angle θ1 and roll attitude angle θ2.

[0023] An array-type liquid level sensor is used for multi-point liquid level detection in a hydraulic oil tank. In this embodiment, six submersible liquid level transmitters are installed at six different positions preset on the top cover of the hydraulic oil tank. The spatial coordinates of each detection point are (x1, y2, 0), (x2, y3, 0), ..., (x6, y6, 0), with the z-axis pointing vertically upward.

[0024] Each level transmitter has a range of 0 to 1m, and the output reports the level height value sequentially via RS485 bus using slave addresses 1 to 6, with a reporting cycle of 50ms.

[0025] The tension pressure transmitters are installed in the rodless chambers of the track tensioning cylinder and the rear support cylinder, respectively. The range is 0 to 400 bar, and the output signal is 4 to 20 mA. After being converted into a digital signal, it is transmitted to the central intelligent controller via RS485 bus.

[0026] The vibration and temperature sensor is a triaxial vibration and temperature composite sensor that is attached to the surface of the hydraulic pump casing by a magnetic base. It integrates a piezoelectric accelerometer and a temperature sensor. The collected signals are conditioned and then uploaded every 10ms via RS485 bus in data frame format. Each frame contains the effective values ​​of the X, Y, and Z axis vibrations and the temperature value.

[0027] The execution control module includes an electro-hydraulic proportional valve group, an electro-proportional actuator, and a hydraulic displacement adjustment unit.

[0028] The electro-hydraulic proportional valve group uses a two-position three-way proportional pressure reducing valve with a PWM drive signal frequency of 200Hz. The valve core opening is controlled by adjusting the duty cycle, thereby adjusting the pressure of the track tensioning cylinder.

[0029] The electric proportional drive is used to drive the walking motor and receives speed and torque commands from the central intelligent controller via the CAN bus.

[0030] The hydraulic displacement adjustment unit is a proportional displacement control electromagnet for a variable displacement piston pump, which changes the displacement by controlling the swashplate tilt angle of the pump through a PWM signal.

[0031] In the above, the central intelligent controller receives data from the dual-axis tilt sensor via the CAN bus, and receives data from the array-type liquid level sensor, tension pressure transmitter, and vibration temperature sensor via the RS485 bus. At the same time, it sends control commands to the electro-hydraulic proportional valve group and electro-proportional actuator via the CAN bus, and connects to the remote operation and maintenance terminal via Ethernet.

[0032] In this embodiment, the pitch angle θ1 and tilt angle θ2 of the tunneling machine are collected in real time by the multi-dimensional sensing module, along with the pressure P1 and pressure P2 obtained from the pressure signals of the track tension cylinder and the rear support cylinder, as well as the liquid level data of six liquid level detection points in the oil tank, the vibration signal and temperature signal of the three axes of the hydraulic pump.

[0033] Next, the central intelligent controller multiplies the collected pressures P1 and P2 by the effective working areas A1 and A2 of the corresponding track tension cylinder and rear support cylinder, respectively, to obtain the current tension force F1 = P1 × A1 and the current support force F2 = P2 × A2.

[0034] Taking the tunneling machine as the research object, the machine's mass m and center of gravity coordinates (L) are obtained by the tunneling machine's own control system. G H G The track ground contact length L and width B are known basic parameters.

[0035] Using pitch angle θ1 and roll angle θ2 as model inputs, gravity is decomposed into a sliding component along the slope direction: m×g×sinθ1×cosθ2 and a normal force component perpendicular to the slope direction: m×g×cosθ1×cosθ2.

[0036] Using the current tension force F1 and the current support force F2 as the input to the actuator, the force balance equation along the slope direction and the moment balance equation around the leading edge of the track ground contact boundary of the tunneling machine are established. This yields a large-slope nonlinear dynamic model with θ1, θ2, F1, and F2 as input variables and attitude data change rates, namely pitch angular velocity and roll angular velocity, and track slip rate as output.

[0037] It should be noted that the track slip ratio s is calculated based on the output speed ω of the travel motor, using the encoder's measured value and the body speed v estimated based on acceleration integral, according to the formula s=(ω×rv) / (ω×r).

[0038] Next, based on the established nonlinear dynamic model of large slope, the central intelligent controller uses the NMPC nonlinear model predictive control algorithm to perform rolling time-domain optimization, as follows: The prediction time domain is set to 3 seconds, the control time domain to 1 second, and the monitoring cycle to 5 seconds.

[0039] Using θ1, θ2, F1, and F2 as the current state variables, a quadratic cost function is constructed. Its optimization objective is to make the pitch angular velocity approach zero, the roll angular velocity approach zero, and the track slip ratio s approach zero within the future monitoring period, while simultaneously using the maximum permissible tension force F1 of the track tensioning cylinder. max and the maximum permissible support force F2 of the rear support cylinder max These are constraints.

[0040] Within each 5-second monitoring cycle, the optimal control sequence under constraints is solved using the interior point method to obtain the cost function, and the target tension F1 corresponding to the current control cycle is extracted. ref and target support force F2 ref .

[0041] The central intelligent controller determines the target tension F1. ref and target support force F2 refA pulse width modulation (PWM) signal is generated and a control command is sent to the electro-hydraulic proportional valve group via the CAN bus to adjust the valve core opening, so that the track tensioning cylinder performs a pre-tensioning action and achieves the target tension force, and the rear support cylinder performs a pre-lifting action and achieves the target support force.

[0042] Then, during the adjustment process, the central intelligent controller continuously obtains the actual tension force F1 of the track tensioning cylinder through the tension pressure transmitter. act And the actual supporting force F2 of the rear support cylinder act When |F1 act -F1 ref | / F1 ref <5% and |F2 act -F2 ref | / F2 ref When <5%; Determine the actual tension force F1 of the track tensioning cylinder act And the actual supporting force F2 of the rear support cylinder act Once the verification is successful, a start command for the travel motor is output to the proportional driver via the CAN bus to control the output torque of the travel motor. If the actual tension force F1 of the track tensioning cylinder is determined... act And the actual supporting force F2 of the rear support cylinder act If the verification fails, return to step S1 to repeat the sensing and control loop until the verification is successful.

[0043] This embodiment establishes a nonlinear dynamic model for steep slopes and combines it with NMPC rolling optimization. It can calculate the optimal tension and support force in real time under different slope change conditions, actively adjust the track grounding conditions and machine stability, avoid landslides and overturning accidents of the tunneling machine, and improve the safety of heavy-duty steep slope operations.

[0044] Example 3

[0045] This embodiment is implemented based on embodiment 2, further adding hydraulic oil tank level safety control function and level data processing and displacement protection strategy, specifically including the following: Based on the content described in Example 2, in the established multi-dimensional sensing module, the coordinates of the six liquid level detection points on the top cover of the hydraulic oil tank are clearly set as follows: point 1 (-300, -200, 0), point 2 (300, -200, 0), point 3 (-300, 200, 0), point 4 (300, 200, 0), point 5 (0, -200, 0), and point 6 (0, 200, 0), all in millimeters.

[0046] Next, each submersible level transmitter outputs the raw level value without filtering.

[0047] After completing Example 2 and starting the walking motor, the control method of this example further includes performing the following steps: First, the central intelligent controller performs real-time filtering processing on the liquid level data collected from six points by the array-type liquid level sensor using the AEKF adaptive extended Kalman filter algorithm.

[0048] The state vector [a, b, c] is based on the dynamic liquid surface equation z = ax + by + c. T A filtering model is constructed, where the state transition process is considered as a random walk. The observation equation is the liquid level value at each detection point: zi = a × xi + b × yi + c, where i = 1, 2, ..., 6, [a, b, c]. T In this context, the superscript T indicates transpose.

[0049] The AEKF adaptive extended Kalman filter algorithm adaptively adjusts the process noise covariance matrix Qk and the measurement noise covariance matrix Rk according to the information sequence in each filtering cycle, eliminating non-Gaussian measurement noise caused by tunneling machine vibration and oil sloshing, and outputting the optimal liquid level estimate at each detection point.

[0050] It should be noted that the innovation sequence refers to the sequence of differences between the actual measured value and the predicted measured value calculated from the current state prediction value in each filtering cycle; In the AEKF adaptive extended Kalman filter algorithm, the algorithm monitors the covariance of the innovation sequence in real time. If the actual fluctuation of the innovation sequence is found to be inconsistent with the theoretical prediction, it indicates that the original noise model: process noise covariance matrix Qk and measurement noise covariance matrix Rk are inaccurate. At this point, the algorithm adaptively adjusts Qk and Rk based on the information sequence, thereby making the filtering result more stable and closer to the real situation.

[0051] Then, using the least squares method with the spatial coordinates (xi, yi) of six points and the corresponding optimal liquid level estimates as input, the fitting coefficients a, b, and c of the dynamic liquid level equation z=ax+by+c are obtained.

[0052] The spatial coordinates of the hydraulic pump's suction port in the oil tank are determined to be (xs, ys, zs), which are known design values. The central intelligent controller calculates in real time the Euclidean normal distance from the suction port coordinates to the dynamic liquid level equation. The vertical safety distance d = |a×xs+b×ys+c-zs| / √(a 2 +b 2 +1).

[0053] Next, the safety thresholds preset by the operator are obtained, including the first preset threshold D1=150mm and the second preset threshold D2=80mm, where D1 is greater than D2.

[0054] The central intelligent controller compares the calculated vertical safety distance d with the safety threshold: When d≥D1, the liquid level is determined to be safe, and the hydraulic pump maintains the current displacement. When D2≤d<D1, the central intelligent controller outputs a displacement limit command to the hydraulic displacement adjustment unit via the CAN bus, generating a 200Hz PWM signal to limit the proportional electromagnet control current of the variable pump to within 70% of the rated displacement of the corresponding pump, thereby limiting the current displacement of the hydraulic pump to within the preset upper limit value. When d < D2, the central intelligent controller outputs a forced shutdown command to the hydraulic displacement adjustment unit, directly setting the displacement control current to zero. At the same time, it outputs an alarm signal to the remote operation and maintenance terminal via Ethernet. The alarm frame includes the vertical safety distance, the liquid level equation coefficient, and a timestamp.

[0055] This embodiment uses AEKF adaptive filtering and least squares liquid level fitting to determine the real-time position of the dynamic liquid level in the oil tank under strong vibration environment. It also implements graded displacement limitation and shutdown protection based on the vertical safety distance between the oil inlet and the liquid level to avoid pump cavitation damage caused by hydraulic pump cavitation and ensure the reliable operation of the hydraulic power system under extreme slope conditions.

[0056] Example 4

[0057] This embodiment, based on the system and method of the aforementioned embodiments, adds equipment health management functions, covering in-depth utilization of vibration and temperature sensors and remote operation and maintenance communication, as detailed below: During each monitoring cycle of normal operation of steep slope anti-slip, anti-tipping control, and oil tank level safety control, the central intelligent controller synchronously executes equipment health management steps, as follows: The central intelligent controller receives vibration and temperature signals from the three axes of the hydraulic pump in real time from the vibration and temperature sensors.

[0058] For a triaxial vibration signal, calculate the root mean square velocity v in the frequency band from 10Hz to 1000Hz. rms The unit is mm / s, and it is used as an index of vibration intensity; Simultaneously, the pump casing temperature H is obtained, in °C.

[0059] The central intelligent controller has a pre-stored standard vibration health threshold v. th =7.1 mm / s and the maximum allowable temperature threshold H for normal pump operation th =90℃.

[0060] v obtained in real time rms With v th Comparison of H and H th Compare.

[0061] When v rms >v th Or H > H th Upon receiving the warning, the central intelligent controller immediately generates predictive maintenance early warning information. This information is a JSON format string containing equipment identifiers, such as the main hydraulic pump, anomaly types, such as vibration exceeding limits and temperature exceeding limits, and the current status (v). rms H value and the timestamp of the trigger time.

[0062] The central intelligent controller uses an Ethernet interface and the MQTT protocol to publish predictive maintenance early warning information to remote operation and maintenance terminals for operators to view in real time.

[0063] In this embodiment, the remote operation and maintenance terminal can receive real-time equipment operation status data, slope attitude data, liquid level safety distance and various alarm signals from the central intelligent controller, so as to realize remote centralized monitoring of multiple large-slope tunneling machines.

[0064] This embodiment, by integrating online monitoring of both vibration and temperature parameters, can proactively detect anomalies and issue predictive maintenance warnings in the early stages of hydraulic pump performance degradation, transforming passive emergency repairs into planned maintenance, reducing unplanned downtime of critical power components in steep slope tunneling operations, and improving the overall utilization rate of equipment.

[0065] All data in the formulas described above have been calculated with dimensions removed. Furthermore, any content not described in detail in this specification is existing technology known to those skilled in the art.

[0066] The above description is merely an example and illustration of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the invention or exceed the scope defined in the claims, all of which should fall within the protection scope of the present invention.

[0067] It should be stated that all user data collected in this application was collected with the user's consent and authorization. Furthermore, the uses of user data are legal and compliant, and the use and processing of user data comply with the relevant laws, regulations, and standards of the relevant regions.

Claims

1. An adaptive safety control system for a steep gradient tunneling machine, characterized in that, The system includes: The central intelligent controller is deployed in the main electrical control cabinet of the tunneling machine; The multi-dimensional sensing module, which communicates with the central intelligent controller, includes: Dual-axis tilt sensors detect the attitude data of the tunneling machine body, including pitch and roll angles; An array-type liquid level sensor consists of liquid level detection points distributed at different preset positions in the hydraulic oil tank, which collects liquid level data at multiple points in the oil tank; The tension pressure transmitter detects the pressure signals of the track tensioning cylinder and the rear support cylinder. Vibration and temperature sensors detect vibration and temperature signals from the hydraulic pump. The execution control module, which communicates with the central intelligent controller, includes: Electro-hydraulic proportional valve assembly controls the tension adjustment of the track tensioning cylinder; An electric proportional drive controls the speed and torque output of the travel motor; Hydraulic displacement adjustment unit, used to adjust the displacement of the hydraulic pump; The central intelligent controller interacts with the multi-dimensional sensing module and the execution control module via CAN bus and RS485 bus, respectively. The central intelligent controller has a built-in control algorithm kernel that generates control commands based on real-time data collected by the multi-dimensional sensing module and inputs the control commands to the execution control module.

2. The system according to claim 1, characterized in that, The dual-axis tilt sensor transmits attitude data to the central intelligent controller via a CAN bus; The array-type liquid level sensor transmits multi-point liquid level data to the central intelligent controller via RS485 bus; The tension pressure transmitter and vibration temperature sensor transmit pressure, vibration, and temperature signals to the central intelligent controller via RS485 bus, respectively. The electro-hydraulic proportional valve assembly and electro-proportional actuator receive control commands from the central intelligent controller via the CAN bus. The central intelligent controller also communicates with the remote operation and maintenance terminal via Ethernet.

3. An adaptive safety control method for a tunnel boring machine with a steep gradient, employing the system described in claims 1 and 2, characterized in that, The method includes: S1 collects real-time attitude data of the tunneling machine, pressure signals of the track tensioning cylinder and rear support cylinder, multi-point liquid level data in the oil tank, vibration signals and temperature signals of the hydraulic pump through the multi-dimensional sensing module. S2, based on attitude data and pressure signals, determines the pitch angle θ1, tilt angle θ2, current tension force F1, and current support force F2 of the tunneling machine, and establishes a large-slope nonlinear dynamic model with θ1, θ2, F1, and F2 as input variables and attitude data change rate as output. S3, based on a large-gradient nonlinear dynamic model, takes the attitude data change rate approaching zero and the track slip rate approaching zero as optimization objectives within the future monitoring cycle, and takes the maximum allowable tension force of the track tensioning cylinder and the maximum allowable support force of the rear support cylinder as constraints. The NMPC nonlinear model predictive control algorithm is used to optimize the rolling time domain and solve for the target tension force and target support force in the current monitoring cycle. S4, the central intelligent controller sends control commands to the electro-hydraulic proportional valve group, which respectively control the track tensioning cylinder and the rear support cylinder to perform pre-tensioning and pre-lifting actions.

4. The method according to claim 3, characterized in that, The method further includes: S5, the central intelligent controller obtains the actual tension force of the track tensioning cylinder and the actual support force of the rear support cylinder through the tensioning pressure transmitter; Verify the actual tension force and actual support force against the target tension force and target support force; If the verification is successful, a start command for the travel motor is output to the electro-proportional driver to control the output torque of the travel motor; If the verification fails, return to step S1 and readjust the control commands until the verification is successful. S6, the central intelligent controller processes the multi-point liquid level data collected by the array liquid level sensor, uses the AEKF adaptive extended Kalman filter algorithm for filtering and noise reduction, uses the least squares method to fit the noise-reduced multi-point liquid level data, constructs the dynamic liquid surface equation of the oil, and calculates the Euclidean normal distance from the oil suction port coordinates to the dynamic liquid surface equation in real time, which is denoted as the vertical safety distance. S7, the central intelligent controller determines whether the vertical safety distance is less than the preset safety threshold; If the displacement is less than the safety threshold, the hydraulic pump displacement is limited by the hydraulic displacement adjustment unit. If the current hydraulic pump displacement is greater than or equal to the safety threshold, maintain the current displacement.

5. The method according to claim 4, characterized in that, In S2 and S3, the method for establishing the nonlinear dynamic model for large slopes is as follows: Taking the tunneling machine as the research object, the overall mass, center of gravity position and track ground contact size are obtained as basic parameters. The overall mass, center of gravity position and track ground contact size are all determined by the tunneling machine's own control system and are considered as known values. Using pitch angle θ1 and roll angle θ2 as model inputs, gravity is decomposed into a sliding component along the slope direction and a normal force component perpendicular to the slope direction. Using the current tension force F1 and the current support force F2 as the input to the actuator, the force balance equation along the slope direction and the moment balance equation around the track ground contact boundary of the tunneling machine are established. The acceleration of the tunneling machine along the slope and the pitch attitude angular velocity are used as the attitude data change rate to form the output of the nonlinear dynamic model for large slopes. The central intelligent controller combines a large-slope nonlinear dynamic model and uses the NMPC nonlinear model predictive control algorithm in real time to perform rolling time-domain optimization, and solves the target tension force and target support force in the current monitoring cycle. Using the pitch angle θ1, roll angle θ2, current tension force F1, and current support force F2 as the current state variables, the preset prediction time domain and control time domain are extracted. With the optimization objectives of the pitch attitude angular velocity approaching zero and the track slip rate approaching zero within the future monitoring cycle, and with the maximum allowable tension force of the track tensioning cylinder and the maximum allowable support force of the rear support cylinder as constraints, a cost function is constructed, wherein the monitoring cycle is 5 seconds. The optimal solution of the cost function under constraints is solved in real time within each monitoring cycle, resulting in the optimal control sequence that includes the target tension force and target support force from the current moment to each future monitoring cycle. Extract the target tension force and target support force outputs for the current control cycle from the optimal control sequence, and perform rolling optimization again in the next monitoring cycle.

6. The method according to claim 5, characterized in that, In S4 and S5, the control command is a pulse width modulation (PWM) signal. The central intelligent controller controls the valve core opening of the electro-hydraulic proportional valve group by adjusting the duty cycle of the PWM signal, thereby adjusting the actual tension and support force of the track tensioning cylinder and the rear support cylinder. Before outputting torque to the walking motor, the central intelligent controller continuously receives feedback from the tension pressure transmitter on the actual tension force and actual support force during the monitoring cycle. When the actual tension force reaches the target tension force and the actual support force reaches the target support force, the central intelligent controller outputs a walking motor start command to the electro-proportional drive.

7. The method according to claim 6, characterized in that, In S6, the AEKF adaptive extended Kalman filter algorithm performs real-time filtering on multi-point liquid level data to eliminate measurement noise caused by tunneling machine vibration and oil sloshing, and outputs the optimal liquid level estimate for each liquid level detection point. The least squares method fits the dynamic liquid level equation z=ax+by+c based on the spatial coordinates of each liquid level detection point and the corresponding optimal liquid level estimate, where a, b, and c are all fitting coefficients.

8. The method according to claim 7, characterized in that, In S7, the preset security threshold includes a first preset threshold and a second preset threshold, wherein the first preset threshold is greater than the second preset threshold. When the vertical safety distance is less than the first preset threshold and greater than or equal to the second preset threshold, the central intelligent controller outputs a displacement limit command to the hydraulic displacement adjustment unit via the CAN bus to limit the current displacement of the hydraulic pump to within the preset displacement upper limit. When the vertical safety distance is less than the second preset threshold, the central intelligent controller outputs a forced shutdown command to the hydraulic displacement adjustment unit and simultaneously outputs an alarm signal.

9. The method according to claim 8, characterized in that, The method also includes equipment health management steps: The central intelligent controller receives vibration and temperature signals collected by the vibration and temperature sensors in real time, and compares the vibration and temperature signals with preset vibration health thresholds and temperature health thresholds, respectively. When the vibration signal exceeds the vibration health threshold or the temperature signal exceeds the temperature health threshold, the central intelligent controller generates predictive maintenance warning information and sends it to the remote operation and maintenance terminal via Ethernet.

Images