Tunnel water-rich fault fracture zone construction emergency early warning collaborative disposal system

Abstract

The application relates to the technical field of tunnel engineering disaster prevention and control, and discloses a tunnel water-rich fault fracture zone construction emergency early warning and collaborative treatment system which comprises a sensing network device, a grouting execution device and data transmission and calculation equipment comprising a collaborative treatment controller; the controller carries out smoothing and derivation on pore water pressure and deformation displacement time sequence data, and extracts first-order derivatives; the first-order derivatives are used to calculate gradient asynchrony indexes and evolution rates, and target grouting pressure and flow instructions are generated and sent to double-liquid grouting pumps; fluid transmission time lag time is calculated in combination with grouting pipeline volume and flow instructions; rock mass fracture acceleration is extracted by twice derivation of the indexes, and double-liquid mixture volume ratio target instructions are generated in combination with the transmission time lag time and output to a double-path proportional regulating valve. The application realizes dynamic matching compensation of material gel time and rock mass fracture evolution, and improves the reliability of water inrush disaster emergency plugging.

Description

Technical Field

[0001] This invention relates to the field of tunnel engineering disaster prevention and control technology, specifically to an emergency early warning and collaborative response system for construction in water-rich fault fracture zones of tunnels. Background Technology

[0002] During the construction of deep-buried tunnels traversing water-rich fault fracture zones, geological hazards such as sudden water and mud inrushes can easily trigger serious engineering accidents. Currently, pore water pressure gauges and displacement gauges are typically installed at the warning sections for safety monitoring at engineering sites, and dual-liquid grouting technology is used for reinforcement and emergency sealing when signs of water inrush appear.

[0003] Traditional disaster early warning methods often rely on static thresholds for single parameters to trigger alarms, typically using a sustained rise in pore water pressure as the primary signal for predicting danger. However, during the microfracture expansion stage before macroscopic rock fractures occur, the volume expansion within the surrounding rock causes localized pore water to be drawn into newly formed fractures, exhibiting a physical phenomenon of short-term drop or stagnation in pore water pressure coupled with accelerated strain. Existing single-dimensional monitoring logic struggles to identify such asynchronous divergent characteristics of multiple physical field parameters, easily leading to missed alarms before mudslides occur. In subsequent disaster response, existing automated grouting equipment often employs open-loop control, with grouting pressure, flow rate, and two-liquid mixing ratio typically set to fixed preset values. This static execution mechanism fails to correlate with the dynamically evolving fracture state within the surrounding rock, preventing the grouting equipment from adaptively adjusting its operation based on the initiation or penetration of rock fractures. Furthermore, in actual engineering projects, the grouting pumps are often located far from the working face, requiring the mixed two-liquid grout to flow through long, high-pressure grouting pipelines to reach the fractured surrounding rock. Existing grouting control strategies do not incorporate the fluid transport time in the pipeline into the control loop for compensation. When geological conditions deteriorate drastically, even if the ratio of the quick-setting agent is adjusted, due to the physical transport time lag in the pipeline, the grout with the adjusted ratio often misses the effective consolidation phase transformation stage when it reaches the rock mass. It is easily diluted and washed away by high-pressure water, leading to failure of emergency blocking and sealing. Summary of the Invention

[0004] The technical problem solved by this invention is that the existing methods for early warning and grouting treatment of water-rich fault fracture zones in tunnels have the following problems: independent monitoring of multiple physical field data leads to easy omission of early warning signs; physical time delays in mechanical response and pipeline transmission cause the grout gel phase transformation to miss the effective consolidation stage, thus resulting in the failure to block water inrush disasters.

[0005] To address the above problems, the present invention provides the following technical solution:

[0006] This invention provides an emergency early warning and collaborative response system for construction in water-rich fault fracture zones of tunnels, comprising:

[0007] The sensor network device includes a pore water pressure gauge that collects raw time-series data of pore water pressure and a multi-point displacement gauge that collects raw time-series data of deformation displacement.

[0008] Grouting execution equipment, including a dual-liquid grouting pump and a dual-path proportional regulating valve;

[0009] Data transmission and computing device, including a collaborative processing controller; the collaborative processing controller is used to perform:

[0010] Polynomial smoothing and differentiation operations are performed on the original time series data of pore water pressure and the original time series data of deformation displacement to extract the first derivative of pore water pressure and the first derivative of strain.

[0011] The gradient asynchronicity index and its evolution rate are calculated based on the first derivative of the pore water pressure and the first derivative of the strain.

[0012] Based on the gradient asynchronicity index and its evolution rate, a target total grouting pressure command and a target total flow rate command are generated and sent to the dual-liquid grouting pump;

[0013] Calculate the transmission delay time by combining the known volume parameters with the target total flow command;

[0014] The gradient asynchronicity index is subjected to a second derivative operation to extract the rock mass fracture acceleration characteristic quantity. The rock mass fracture acceleration characteristic quantity is combined with the transmission time delay to generate a target command for the two-liquid mixing volume ratio and output to the dual-path proportional control valve.

[0015] Preferably, the step of performing polynomial smoothing and differentiation operations on the original time-series data of pore water pressure and the original time-series data of deformation displacement to extract the first derivative of pore water pressure and the first derivative of strain includes:

[0016] A causal sliding time window is allocated and maintained in memory, which stores only the current moment and historical sampling data before the current moment.

[0017] The data points within the causal sliding time window are subjected to dimensionality reduction mapping of relative time coordinates, and an observation matrix in the local coordinate system is constructed based on the mapped relative time coordinates.

[0018] The coefficients of the polynomial equations are solved directly based on the principle of normal equations using the least squares method, and the first-order analytic derivative of the polynomial equations after determining the coefficients is then performed.

[0019] Substitute the relative time coordinate value corresponding to the latest current moment into the first derivative equation to extract the first derivative of pore water pressure and strain without time delay at the latest current moment.

[0020] Preferably, the step of calculating the gradient asynchronicity index based on the first derivative of the pore water pressure and the first derivative of the strain includes:

[0021] In the underlying logic gate operation, it is determined in parallel whether the first derivative of the current strain is greater than zero and whether the first derivative of the current pore water pressure is less than the set pore water pressure attenuation constant.

[0022] Only when both determination results are true, the product of the current strain first derivative and the difference between the pore water pressure attenuation constant and the current pore water pressure first derivative is extracted, and the product is used as the effective discrete divergence feature value of the current control cycle.

[0023] If any determination result is false, the effective discrete divergence feature value will be forcibly set to zero;

[0024] The effective discrete divergence feature values ​​are pushed into the tail of the first-in-first-out queue, and the oldest historical data removed from the head of the first-in-first-out queue is released simultaneously. All effective discrete divergence feature values ​​in the current first-in-first-out queue are traversed and multiplied by the single scan cycle time, and the results are summed to calculate the gradient asynchronicity index of the current control cycle.

[0025] Preferably, the collaborative processing controller further includes anti-saturation limiting protection logic, which is used for:

[0026] A preset anti-lock upper limit threshold is set, which is calculated and calibrated by combining the maximum safe working pressure limit of the dual-liquid grouting pump with the proportional gain parameter of the subsequent closed-loop controller.

[0027] After each discrete accumulation operation is completed, the calculated gradient asynchronicity index is compared with the anti-lock upper limit threshold.

[0028] If the calculated gradient asynchronicity index exceeds the anti-lock upper limit threshold, the calculation is directly truncated and the gradient asynchronicity index is locked to the anti-lock upper limit threshold.

[0029] Preferably, generating the target total grouting pressure command and the target total flow rate command based on the gradient asynchronicity index and its evolution rate includes:

[0030] The risk deviation of the current grouting condition is obtained by subtracting the received gradient asynchronicity index from the internally stored zero reference value. The zero reference value is set to an absolute zero value or the long-term average of the background characteristic data when the construction site is undisturbed and there is no risk of water seepage.

[0031] The pressure regulation loop is operated, and the risk deviation and the evolution rate are converted into pressure increments to compensate for rock mass fracture and expansion using proportional-integral-differential operation rules. The target total grouting pressure command is generated by superimposing the basic pressure-maintaining base.

[0032] The operation flow regulation loop, when the gradient asynchronicity index and its evolution rate cross the set critical warning threshold, rapidly reduces the target total flow command value through an active negative feedback mechanism. The critical warning threshold is determined by combining the tensile strength and porosity parameters of the anti-outburst rock mass obtained in the early geological exploration stage with fluid-structure interaction numerical simulation to obtain the theoretical upper limit of the deviation index at macroscopic fracture.

[0033] Preferably, the calculation of the transmission delay time based on the known volume parameters of the high-pressure grouting pipeline includes:

[0034] A minimum flow safety threshold and a corresponding maximum allowable physical time delay are preset; the minimum flow safety threshold is determined by back-calculation based on the mechanical displacement corresponding to the minimum stable frequency conversion operating frequency of the main drive motor of the dual-liquid grouting pump under the premise of no stall or overheating.

[0035] The maximum permissible physical time delay is calculated and calibrated in reverse based on the initial setting time of the current two-component formula under no external disturbance and with a safety factor reserved to prevent pipeline blockage.

[0036] The read target total traffic command is numerically compared with the minimum traffic safety threshold.

[0037] If the target total traffic instruction is greater than or equal to the minimum traffic safety threshold, the known volume parameter is divided by the target total traffic instruction to obtain the quotient value, and the quotient value is used as the transmission delay time.

[0038] If the target total flow instruction is lower than the minimum flow safety threshold, the division operation is directly stopped, and the preset maximum allowable physical delay limit is forcibly output.

[0039] Preferably, the step of generating the target command for the two-liquid mixing volume ratio by combining the rock mass fracture acceleration characteristic quantity with the transmission time delay includes:

[0040] The rock mass fracture acceleration characteristic value calculated by itself and the transmission time delay are read synchronously as the basic input variables, and the deep coupling control law equation is called; the deep coupling control law equation includes the basic mixed volume proportional constant, the feedforward control gain coefficient, the time delay penalty compensation coefficient and the step function;

[0041] The advanced compensation logic is executed through the deep coupling control law equation to calculate and generate the theoretical target command for the two-liquid mixing volume ratio.

[0042] The calculated theoretical target volume ratio of the two-liquid mixture is compared with the maximum allowable mixing ratio threshold. If the calculated value exceeds the maximum allowable mixing ratio threshold, the output command is directly truncated and locked to the maximum allowable mixing ratio threshold. The maximum allowable mixing ratio threshold is calibrated based on the upper limit of the ratio corresponding to the extremely short safe pumpable time of the two-liquid material under the pumping pressure.

[0043] Preferably, the step of performing a second derivative operation on the gradient asynchronicity exponent to extract the rock mass fracture acceleration feature includes:

[0044] The read evolution rate is smoothed using a first-order inertial low-pass filter.

[0045] Perform time difference operations on the smoothed evolution rate to obtain the feature values;

[0046] The calculated characteristic value is subjected to directional validity constraint judgment. If the characteristic value is greater than zero, it is allowed to proceed and the rock mass fracture acceleration characteristic value is confirmed. If the characteristic value is less than or equal to zero, the characteristic value is forcibly set to zero.

[0047] Preferably, the collaborative processing controller further includes hardware self-locking protection logic for:

[0048] The actual pumping pressure is collected in parallel from the physical pressure sensor installed at the slurry manifold of the dual-liquid grouting pump.

[0049] Whenever the actual pumping pressure collected touches or exceeds the hardware fuse pressure threshold, the underlying interrupt protection program is activated first to directly cut off the control of the dual-path proportional regulating valve.

[0050] The hardware fusing pressure threshold is set by back-calculating 80% of the rated burst pressure of the high-pressure grouting pipeline.

[0051] A preset reference safety ratio electrical signal is forcibly sent to the servo driver of the dual-path proportional regulating valve to execute the closing action of the accelerator slurry inlet channel, so that the actuator only retains the single liquid base slurry or clear water pumping state and enters the mechanical self-locking state.

[0052] Preferably, the collaborative processing controller further includes electromechanical jamming detection logic, used for:

[0053] The actual mechanical opening data fed back by the displacement feedback sensor installed at the valve shaft of the dual-channel proportional control valve is read in real time and converted into the actual physical mixing ratio;

[0054] The actual physical mixing ratio is compared with the theoretical target command for the two-liquid mixing volume ratio.

[0055] When the actual physical mixing ratio deviates from the target command of the two-liquid mixing volume ratio by more than the preset allowable error dead zone, and the duration of the deviation exceeds the set delay confirmation time window, it is determined that the underlying actuator has mechanically jammed, the rheological feedforward adjustment logic is stopped and the equipment abnormality alarm is triggered.

[0056] The allowable error dead zone is comprehensively calibrated by combining the mechanical gear backlash of the dual-path proportional control valve and the measurement accuracy of the displacement sensor;

[0057] The length of the delayed confirmation time window is set by inversely calculating the inertial response time of the servo drive motor under maximum load.

[0058] This invention provides an emergency early warning and collaborative response system for construction in water-rich fault fracture zones of tunnels. It has the following beneficial effects:

[0059] 1. This invention performs time-delay-free polynomial smoothing and differentiation on the time-series data of pore water pressure and deformation displacement, and combines the step function to perform continuous time-domain integration on the gradient asynchrony divergence characteristics of the two, thereby calculating the gradient asynchrony index. This method effectively filters out high-frequency mechanical noise interference in the construction environment, accurately extracts the characteristics of pore pressure drop and strain increase caused by micro-fracture expansion before macroscopic rock mass fracturing, avoids false alarms caused by traditional single static threshold monitoring, and improves the accuracy and timeliness of early warning of water inrush disasters.

[0060] 2. This invention constructs a deep-coupled control law equation that integrates the characteristic quantity of rock mass fracture acceleration and the fluid transport time delay. It is used to dynamically output the target command for the mixing volume ratio of the two liquids. By calculating the physical delay of the grouting pipeline in real time and increasing the mixing ratio of the quick-setting agent in the accelerated stage of fracture evolution, the gelation time of the grout is actively shortened, which compensates for the time misalignment caused by long-distance pipeline transmission. This allows the grout to undergo phase change consolidation as needed when it reaches the fracture zone of the working face, thus meeting the collaborative sealing requirements under the critical condition of water inrush.

[0061] 3. The present invention sets up multi-dimensional boundary protection logic in the computation and execution layers of the collaborative processing controller, including integral anti-saturation limit, division-to-zero protection, maximum mixing ratio limit, and hardware fuse self-locking mechanism. The above boundary constraints not only avoid the integral lock-up and division overflow anomalies that may occur in the lower controller, but also quickly cut off the feedforward control and force the execution of single liquid safety pumping when abnormal overpressure occurs in the pipeline or mechanical jamming of the proportional valve, thus ensuring the reliability of the system operation and the safety of the equipment in complex geological and construction environments. Attached Figure Description

[0062] Figure 1 This is a diagram illustrating the architecture of the emergency early warning and collaborative response system for tunnel construction in water-rich fault fracture zones, as described in this invention.

[0063] Figure 2 This is a flowchart of the emergency early warning and collaborative response system for tunnel construction in water-rich fault fracture zones, as described in this invention.

[0064] Figure 3 This is a comparison diagram of the evolution of early warning features and the effect of collaborative processing in this invention. Detailed Implementation

[0065] The technical solutions in 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.

[0066] See attached document Figure 1 This invention provides an emergency early warning and collaborative response system for construction in water-rich fault fracture zones of tunnels. The system may include sensor network equipment, data transmission and computing equipment, and grouting execution equipment.

[0067] The sensor network equipment is deployed inside the surrounding rock at the tunnel face and the early warning section. The sensor network equipment includes several pore water pressure gauges and multi-point displacement gauges. The pore water pressure gauges are used to collect raw time-series data of pore water pressure inside the surrounding rock at the monitoring points. The multi-point displacement gauges are used to collect raw time-series data of deformation and displacement of the surrounding rock at the early warning section.

[0068] The data transmission and computing equipment includes an edge computing gateway and a collaborative processing controller. The signal outputs of the pore water pressure gauge and the multi-point displacement gauge are connected to the signal input ports of the edge computing gateway. The edge computing gateway is connected to the collaborative processing controller via an industrial Ethernet bus. The collaborative processing controller is constructed using a programmable logic controller (PLC).

[0069] The grouting equipment includes a dual-liquid grouting pump and a dual-path proportional control valve. The dual-path proportional control valve is installed at the intersection of the two grout inlet pipes of the dual-liquid grouting pump. The analog output port of the coordinating controller is connected to the main drive frequency converter of the dual-liquid grouting pump and the driver of the dual-path proportional control valve, respectively. The discharge port of the dual-liquid grouting pump is connected to the grouting hole reserved at the tunnel face through a high-pressure grouting pipeline.

[0070] The collaborative treatment controller contains multiple system function modules. These include a data robustness analysis module, an early warning feature calculation module, a grouting energy distribution module, a fluid time delay measurement module, and a rheological characteristic feedforward tuning module.

[0071] See attached document Figure 2 The overall collaborative workflow of the system may include:

[0072] The S100 and edge computing gateway synchronously receive time-series data collected by the pore water pressure gauge and multi-point displacement gauge according to the set sampling frequency, and send the time-series data to the data robust analysis module inside the collaborative processing controller. The data robust analysis module performs polynomial smoothing and differentiation operations on the time-series data to extract the first derivative of pore water pressure and the first derivative of strain at the corresponding time node.

[0073] The S200 early warning feature calculation module receives the first derivative of pore water pressure and the first derivative of strain output by the robust data analysis module, and calculates the gradient asynchronicity index and the evolution rate of the gradient asynchronicity index according to the set integration operation rules.

[0074] The S300 grouting energy distribution module reads the gradient asynchronicity index and evolution rate output by the early warning feature calculation module, and generates the target total grouting pressure command and the target total flow command. The grouting energy distribution module sends the target total grouting pressure command and the target total flow command to the main drive frequency converter of the dual liquid grouting pump to adjust the output power of the grouting execution equipment.

[0075] The S400 fluid time delay calculation module reads the target total flow command generated by the grouting energy distribution module in real time, and calculates the transmission time delay of the current grout flowing through the high-pressure grouting pipeline by combining the known volume parameters of the high-pressure grouting pipeline.

[0076] The S500 rheological characteristic feedforward tuning module performs a second derivative operation on the gradient asynchronicity index output by the early warning feature calculation module to extract the rock mass fracture acceleration feature quantity; the rheological characteristic feedforward tuning module combines the rock mass fracture acceleration feature quantity with the transmission delay time calculated by the fluid time delay measurement module to calculate and generate the dual-liquid mixing volume ratio target command, and converts the dual-liquid mixing volume ratio target command into an electrical control signal output to the dual-path proportional regulating valve.

[0077] In one embodiment of the present invention, the internal implementation details of the robust data parsing module are provided. The edge computing gateway is responsible for data interaction with the front-end sensor network devices. In the actual construction environment of tunnel engineering, conventional drilling and blasting operations and the operation of large construction machinery generate continuous mechanical vibrations. These vibration signals are usually coupled to the pore water pressure gauges and multi-point displacement gauges inside the surrounding rock, resulting in the superposition of high-frequency noise signals in specific frequency bands in the acquired raw time-series data. If the traditional finite difference method is used directly to perform derivative calculations on the discrete time-series data with high-frequency noise, the high-frequency noise will be amplified during the difference calculation process, which may cause deviations in the rate of change features extracted by the system. To improve this high-frequency noise amplification phenomenon and eliminate the physical time delay caused by conventional smoothing algorithms, in this embodiment, the robust data parsing module performs polynomial fitting and time-delay-free derivative extraction operations based on asymmetric causal window technology, ensuring strict time synchronization between the feature output and the front-end sampling. The internal implementation details of the robust data parsing module may include:

[0078] S110: The edge computing gateway synchronously acquires analog signals from the pore water pressure gauge and multi-point displacement gauge according to the set sampling frequency, and converts the analog signals into digital form of raw pore water pressure and strain time-series data. The hardware circuit design for the analog-to-digital conversion and low-pass filtering within the edge computing gateway is well-known in the field and will not be described further here. The edge computing gateway then transmits the time-series data, after basic electrical filtering, to the robust data parsing module within the collaborative processing controller via industrial Ethernet.

[0079] S120, the robust data parsing module allocates and maintains a causal sliding time window in memory. This causal sliding time window only stores the current moment and historical sampling data up to the current moment, and does not contain any future data points. The length of this causal sliding time window is set to... , The value is a pre-set positive integer. During the initial data accumulation phase at the beginning of system startup, when the number of collected data points has not reached the length of the sliding time window... During this process, the robust data parsing module stores the data in a cache and temporarily refrains from outputting derivative features until the cached data fills the sliding time window, at which point it enters the normal sliding step calculation process. When the edge computing gateway pushes a new discrete data point, the robust data parsing module pushes this new data point onto the right endpoint of the causal sliding time window and simultaneously removes the oldest data point on the left side of the window. To prevent the absolute timestamp value from becoming too large as the system runs continuously, thus causing floating-point overflow in the underlying matrix operations, the robust data parsing module needs to perform dimensionality reduction mapping on the data points within the causal sliding time window using relative time coordinates. As a preferred approach, the robust data parsing module maps the time coordinate of the oldest data point in the window to 0, and maps the time coordinate of the most recently acquired data point to... Within the causal sliding time window, the remaining data points are assigned local coordinate system values ​​at equal intervals according to the sampling period.

[0080] S130, the robust data analysis module constructs the observation matrix in the local coordinate system. The order of the fitted polynomial is set to... The robust data parsing module constructs a structure with dimensions based on the relative time coordinates after mapping. The Vandermonde observation matrix. Any element in the matrix represents the corresponding power of a specific relative time coordinate. The raw physical samples within the causal sliding time window are arranged into a matrix with dimension [missing information]. The column vectors are then used. Subsequently, the robust data parsing module, based on the principle of the normal equation using the least squares method, utilizes the observation matrix and its transpose, combined with the column vectors containing the original sampled data, to directly solve the coefficients of the polynomial equation through underlying matrix algebra operations. For ill-conditioned matrix non-invertibility issues that may arise during the underlying matrix inversion operation due to data collinearity, those skilled in the art can use singular value decomposition (SVD) to solve for the generalized inverse matrix to ensure the algorithm's continued operation; this will not be elaborated upon here.

[0081] S140. The robust data analysis module calculates the analytical derivative of the polynomial equation at the right endpoint of the causal sliding time window based on the obtained polynomial coefficients. In the relative time coordinate system, the right endpoint of the causal sliding time window corresponds to the latest sampling time. In specific implementation, taking the extraction of the time-delay-free first derivative of pore water pressure as an example, the robust data analysis module directly performs first-order analytical differentiation on the polynomial equation after determining the coefficients, then substitutes the relative time coordinate value corresponding to the latest time into the first-order derivative equation, and obtains the specific transient slope value through algebraic addition and multiplication operations, which is used as the first derivative of pore water pressure at the latest time. The calculation process for the first derivative of strain is similar.

[0082] The S150 robust data parsing module uses the calculated, latency-free first derivative at the latest moment as a valid dynamic feature value and directly outputs it to the early warning feature calculation module. By anchoring the calculation evaluation point to the rightmost edge of the fitted curve, the system achieves time synchronization with the front-end data acquisition while extracting the trend of change. After completing a single parsing output, the robust data parsing module resets its internal register parameters and waits for the edge computing gateway to push the sampling data for the next moment to start a new round of coordinate mapping and calculation.

[0083] In one embodiment of the present invention, the operating mechanism of the integral accumulator in the early warning feature calculation module is described in detail. In the underlying computing architecture of an industrial control system, since microprocessors cannot directly process theoretical continuous-time differential equations, the early warning feature calculation module incorporates discretization transformation logic. This logic transforms the theoretical gradient asynchronous exponential continuous integral model into a time-domain discrete accumulation task based on the system scan cycle, thereby helping to enable the early warning features to be accurately executed by the programmable logic controller. The internal operating mechanism of the integral accumulator in the early warning feature calculation module may include:

[0084] S210. The early warning feature calculation module synchronously reads the first derivative of pore water pressure and the first derivative of strain currently output by the robust data analysis module according to the set control scan cycle, and calculates the effective discrete deviation feature value of the current control cycle. In this embodiment, the early warning feature calculation module performs Boolean logic substitution on the step function in the mathematical model, and performs two numerical comparison judgments in parallel in the underlying logic gate operation: determining whether the current first derivative of strain is greater than zero, and determining whether the current first derivative of pore water pressure is less than the set pore water pressure attenuation constant. Only when both judgment results are true, the early warning feature calculation module extracts the product of the first derivative of strain and the difference between the pore water pressure attenuation constant and the first derivative of pore pressure, and uses it as the effective discrete deviation feature value of the current control cycle. If either judgment result is false, the early warning feature calculation module forcibly sets the effective discrete deviation feature value of the cycle to zero. For the floating-point comparison and memory address offset operations at the underlying level of the early warning feature calculation module, those skilled in the art can use conventional instruction sets for programming, and the specific execution process of its instruction calls is a well-known technology in the field, which will not be described in detail here.

[0085] The S220 early warning feature calculation module allocates and maintains a first-in-first-out (FIFO) queue storage area in its internal registers to represent the sliding integration time window. During the cold start phase of the system's initial power-on, the early warning feature calculation module actively performs an initialization and zeroing operation on this FIFO queue storage area to prevent any residual random and disordered data in memory from interfering with the initial accumulation calculation, thereby avoiding erroneous instruction issuance during the initial system startup. As a preferred method, the early warning feature calculation module divides the set time window length by the single scan cycle time to calculate the total number of discrete nodes required for the FIFO queue. Each time a valid discrete deviation feature value is calculated, the early warning feature calculation module pushes the value to the tail of the FIFO queue and simultaneously releases the oldest historical data removed from the head of the queue to maintain a fixed span and dynamic stepping of the integration time window in the time domain.

[0086] S230, the early warning feature calculation module iterates through all valid discrete divergence feature values ​​in the current first-in-first-out queue and performs an accumulation operation to calculate the gradient asynchronicity index of the current control cycle. From an engineering implementation perspective, this calculation step approximates the complex continuous-time mathematical integration by using a discretized multiply-accumulate loop in the time-series state. In specific implementation, the early warning feature calculation module multiplies each valid discrete divergence feature value in the first-in-first-out queue with the single scan cycle time of the controller, and then sums all the product results in the queue. By introducing the single scan cycle as a time product term, the early warning feature calculation module achieves dimensional alignment with the continuous integral dimension of real physical time in the underlying discrete summation operation, thereby deriving the gradient asynchronicity index used to characterize the risk of water inrush at the current moment.

[0087] The S240 early warning feature calculation module performs differential calculations between the gradient asynchronicity index of the current control cycle and the gradient asynchronicity index of the previous control cycle to extract the evolution rate of the gradient asynchronicity index and implements anti-saturation limiting protection for the integral accumulator. In complex engineering monitoring environments, if the rock mass maintains a slight deviation state for a long time, the value inside the integral accumulator will continue to grow slowly, which may lead to the integral lock-up phenomenon commonly seen in industrial control theory. This phenomenon can easily cause abnormal saturation of the system output, making it unable to respond to subsequent reverse exit safety signals. To improve this type of algorithm dead zone problem, the early warning feature calculation module has a preset anti-lock-up upper limit threshold. Regarding the determination of this anti-lock-up upper limit threshold, field technicians usually perform reverse calculation and calibration based on the maximum safe working pressure limit allowed by the grouting execution equipment, combined with the proportional gain parameter of the subsequent closed-loop controller, to ensure that the integral accumulation output does not exceed the physical equipment's tolerance boundary when converted into control commands. After each discrete accumulation operation, the early warning feature calculation module compares the calculated gradient asynchronicity index with this anti-lock-up upper limit threshold. If the calculated value exceeds the threshold, the early warning feature calculation module directly truncates the calculation and locks the gradient asynchronicity index to the anti-lock upper limit threshold, which helps the early warning feature data to be within a safe physical response range when it is transmitted to the subsequent grouting energy distribution module.

[0088] In one embodiment of the present invention, the dual closed-loop control logic of the grouting energy distribution module is detailed. In the coordinated operation of rock mass early warning and grouting treatment, grouting equipment operating with fixed parameters is difficult to adapt to the dynamically evolving rock mechanical state. Mapping characteristic indices representing fracture risk to drive commands for mechanical equipment enables the establishment of a linkage control mechanism from algorithmic judgment to physical execution. In this embodiment, the grouting energy distribution module dynamically adjusts the pressure and flow output of the dual-liquid grouting pump according to the evolution of the gradient asynchronicity index based on its internally configured control logic. Specific internal implementation details of the grouting energy distribution module performing closed-loop energy distribution may include:

[0089] S310, the grouting energy distribution module acquires the gradient asynchronicity index and its evolution rate output by the early warning feature calculation module in real time according to the set calculation cycle. To construct the system deviation basis required for closed-loop control, the grouting energy distribution module internally stores a zero-point reference value in a safe state. During the initial calibration phase of system operation, engineers can choose to set this zero-point reference value to absolute zero, or extract the long-term average of the background feature data when the construction site is undisturbed and there is no risk of seepage as the zero-point reference value. The grouting energy distribution module subtracts the gradient asynchronicity index received in the current cycle from this zero-point reference value to obtain the risk deviation of the current grouting condition. This risk deviation and its corresponding time change rate are input in parallel to the pressure regulation loop and flow regulation loop of the controller.

[0090] The S320 grouting energy distribution module operates a pressure regulation loop, converting risk deviations into specific pressure intervention signals. In its implementation, the pressure regulation loop employs a proportional-integral-differential (PI-DI) arithmetic rule. The grouting energy distribution module directly responds to the current risk deviation using a proportional gain coefficient, while simultaneously using a differential gain coefficient to amplify the evolution rate in the input parameters, thereby calculating the pressure increment used to compensate for rock mass fracture and expansion. The module then superimposes this pressure increment with a preset baseline pressure to generate the theoretical target total grouting pressure command. For the implementation of the specific algebraic equations for the standard integral accumulation and differential differentiation at the bottom layer of the control algorithm, those skilled in the art can call the standard operational function blocks built into the programmable logic controller (PLC). The basic code construction is well-known in the field and will not be elaborated upon here.

[0091] The S330 grouting energy distribution module operates a flow regulation loop, executing reverse cooperative constraint logic corresponding to the pressure command. Based on fluid mechanics and rock grouting engineering principles, pressure and flow rate need to be precisely matched to specific geological fracture stages. High flow rate is required during the micro-fracture stage, but if a high flow rate is maintained when the fracture is about to open and water inrush occurs, grout loss can easily occur, preventing the establishment of effective sealing back pressure. As a preferred approach, the flow regulation loop uses a segmented mapping rule based on the numerical range of the gradient asynchronicity index. When the gradient asynchronicity index is in a low range, the system determines that the rock mass is in the micro-fracture permeability stage, and the grouting energy distribution module maintains a high target total flow rate command, prompting the grout to quickly enter the surrounding rock. When the gradient asynchronicity index and its evolution rate cross the set critical warning threshold, the system determines that the rock mass has entered the critical breakthrough stage, and the flow regulation loop rapidly reduces the target total flow rate command value through an active negative feedback mechanism, thereby cooperating with the high pressure command to form a pressure-maintaining consolidation and sealing condition. The determination of the aforementioned critical warning threshold is usually based on the tensile strength and porosity parameters of the rock mass in the area obtained during the early geological exploration stage. The theoretical upper limit of the deviation index when macroscopic rupture is about to occur is calculated by combining fluid-structure interaction numerical simulation, and this is used as the basis for threshold calibration.

[0092] The S340 grouting energy distribution module performs over-limit protection and smoothing filtering on the calculated target total grouting pressure and target total flow commands. To prevent damage to mechanical equipment caused by over-limit values ​​output by the control algorithm, the grouting energy distribution module internally presets the maximum safe mechanical pressure limit allowed by the grouting execution equipment and the minimum operating flow limit to prevent grout sedimentation in the pipeline. The module compares and truncates the theoretical command with the above-mentioned limiting boundaries. Subsequently, the grouting energy distribution module converts the target total grouting pressure and target total flow commands, after being limited and protected, into standard industrial analog electrical signals and sends them to the main drive frequency converter of the dual-liquid grouting pump. To mitigate the mechanical shock caused by command transients, the grouting energy distribution module incorporates a first-order inertial low-pass filter on the signal output side, which helps to smoothly transition the frequency adjustment process of the frequency converter drive motor, thereby facilitating a robust closed loop for adaptive energy transfer in the system.

[0093] In one embodiment of the present invention, the dynamic analytical logic of the fluid time delay calculation module is described in detail. In dual-liquid grouting operations, dual-path proportional control valves are typically located on the grout inlet side or at the front confluence of the dual-liquid grouting pump. The mixed grout needs to flow through a pre-set length of high-pressure grouting pipeline to reach the grouting hole at the tunnel face corresponding to the warning section. This time-consuming fluid transmission in physical space results in a time lag between the control command and the actual phase change consolidation of the grout. To match the dynamically evolving rock mass fracture state, in this embodiment, the fluid time delay calculation module evaluates this lag time in real time based on fluid dynamics principles, providing a time compensation benchmark for subsequent material proportioning intervention. The internal implementation details of the fluid time delay calculation module may include:

[0094] S410, the fluid time delay calculation module reads the fixed internal volume parameters of the high-pressure grouting pipeline entered by the system during the initial operation phase. During the actual engineering deployment phase, on-site technicians calculate these fixed internal volume parameters by measuring the physical inner diameter and laying length of the pipeline between the grout mixing port of the dual-liquid grouting pump and the grouting hole at the tunnel face. As a preferred method, if multiple sections of flexible hoses or steel pipes of different diameters are spliced ​​together during on-site construction, these fixed internal volume parameters represent the sum of the geometric volumes of the inner cavities of each independent pipe section. Considering the physical characteristics of high-pressure hoses in actual engineering, the pipeline will undergo a certain degree of radial expansion under high-pressure grouting conditions. When inputting the volume parameters, technicians can introduce a compensation margin based on the rated expansion coefficient of the hose to improve the accuracy of the volume calculation. After calculation and calibration, this volume value is stored as a static basic characteristic parameter in the register unit of the fluid time delay calculation module.

[0095] S420, the fluid time delay calculation module reads the target total flow command generated by the grouting energy distribution module in real time according to the system's set calculation cycle. Based on the law of conservation of fluid mass, the transmission time of fluid within a closed pipe is directly proportional to the pipe volume and inversely proportional to the delivery flow rate. Based on this fluid dynamics principle, the fluid time delay calculation module divides the fixed internal volume parameter of the high-pressure grouting pipeline by the target total flow command acquired at the current moment, calculates the quotient through algebraic division, and directly uses this quotient as the transmission time delay of the grout flowing through the high-pressure grouting pipeline at the current moment. Through this calculation process, the system can transform the fluid flow characteristics in the energy execution dimension into physical delay characteristics in the time dimension.

[0096] S430. In handling abnormal operating conditions, the fluid time delay calculation module implements division-to-zero protection and amplitude limiting intervention logic. To ensure the robust operation of the industrial control program, the fluid time delay calculation module has an internal mechanism for handling computational anomalies. When the grouting energy distribution module issues a shutdown command or is in an extremely low flow infiltration condition, the target total flow command value approaches zero. At this time, the divisor region of the aforementioned dynamic time delay analysis calculation logic is at risk of triggering a division-to-zero overflow anomaly, which can easily cause the underlying calculation chip to report an error or output invalid infinite large data. To effectively avoid this underlying computational overflow anomaly, the fluid time delay calculation module pre-sets a minimum flow safety threshold and a corresponding maximum allowable physical time delay upper limit. The range of this minimum flow safety threshold is usually determined by back-calculating the mechanical displacement corresponding to the minimum stable variable frequency operating frequency of the dual-liquid grouting pump main drive motor under the premise of no stall or overheating. The basis for determining the maximum permissible physical time delay is usually calculated and calibrated by engineers based on the initial setting time of the current two-component formula under no external disturbance, and by reserving a certain safety factor to prevent pipeline blockage.

[0097] Before performing the division operation, the fluid time delay calculation module (S440) compares the read target total flow command with the minimum flow safety threshold. If the target total flow command is greater than or equal to the minimum flow safety threshold, the fluid time delay calculation module normally calls the dynamic time delay parsing division operation logic to perform time delay analysis; if the target total flow command is lower than the minimum flow safety threshold, the fluid time delay calculation module directly stops the parsing operation and forcibly outputs the preset maximum allowable physical time delay upper limit. This processing mechanism not only maintains the continuity of the underlying control code execution but also helps to limit the time delay compensation signal to within a reasonable physical boundary. After completing the parsing and evaluation, the fluid time delay calculation module sends the finally determined transmission time delay to the subsequent rheological characteristic feedforward tuning module, providing forward-looking compensation parameters for the deep coupling calculation of the two-liquid mixing ratio.

[0098] In one embodiment of the present invention, the internal implementation details of the rheological property feedforward tuning module are provided. In the coordinated treatment of rock mass water inrush prevention, the distribution of grouting energy mainly changes the physical pressure injection state of the fluid. When dealing with the critical condition of mudslide penetration, it is often necessary to intervene in the chemical phase transformation process of the injected material. Since the gel time of the two-liquid grouting material is controlled by the mixing volume ratio of the two liquids, predicting the macroscopic instability node of rock mass fracture penetration in advance helps the grouting fluid to complete the consolidation phase transformation when it reaches the fracture surface. In this embodiment, the rheological property feedforward tuning module provides a decision basis for the advance intervention of material ratio by extracting high-order dynamic parameters from the early warning feature data. The internal implementation details of the rheological property feedforward tuning module may include:

[0099] S510, the rheological characteristic feedforward tuning module synchronously reads the gradient asynchronicity index and its first-order evolution rate output by the early warning feature calculation module according to the set control cycle. Based on the fluid-structure interaction damage theory in rock mechanics, the gradient asynchronicity index characterizes the cumulative total amount of microcrack expansion damage inside the rock mass, and its first-order evolution rate reflects the current speed of damage propagation. When the rock mass approaching the critical threshold of macroscopic mudslide penetration, the damage propagation rate will show a sharp upward evolution trend. To capture this transient critical characteristic, the rheological characteristic feedforward tuning module is equipped with secondary analytical calculation logic to perform high-order time-dimensional calculations on the above dynamic parameters.

[0100] The S520 rheological characteristic feedforward tuning module performs time-difference operations on the read first-order evolution rate to extract the rock mass fracture acceleration characteristic quantity. In engineering signal processing practice, high-order numerical differentiation operations can easily amplify the small fluctuations remaining in the underlying signal. To avoid high-frequency characteristic pulses that cause distortion during the secondary analysis process, which in turn cause abnormal fluctuations in subsequent ratio commands, the rheological characteristic feedforward tuning module smooths the first-order evolution rate using an internally set low-pass filter before performing the difference operation. As a preferred method, the system uses a first-order inertial low-pass filter to perform this smoothing operation. Field engineers usually calibrate and determine the filtering time constant of this filter based on the inherent physical response delay time of the front-end pore water pressure gauge and multi-point displacement gauge to achieve a balance between noise reduction and feature preservation. For the algorithm construction of the basic discrete difference equation, those skilled in the art can perform conventional matching design based on the hardware scan step size of the controller. Its specific logic implementation is a well-known technology in this field and will not be elaborated here. After filtering and time difference analysis, the rheological characteristic feedforward tuning module obtains the characteristic quantity of rock mass fracture acceleration. This characteristic quantity directly represents the urgency of the countdown before the anti-outburst rock mass structure is about to be penetrated by high-pressure water at the physical level.

[0101] The S530 rheological property feedforward tuning module performs directional validity constraint judgment on the calculated rock mass fracture acceleration characteristic quantity. In the actual evolution process of rock mass grouting reinforcement, there is not only a deterioration stage of accelerated damage expansion, but also a repair stage where damage expansion gradually slows down due to the previous grout filling. In the repair stage, the rock mass fracture acceleration characteristic quantity is mathematically negative. To avoid the rheological material intervention logic being falsely triggered under low-risk conditions, the rheological property feedforward tuning module performs non-negative amplitude limit judgment on the rock mass fracture acceleration characteristic quantity. Specifically, if the calculated characteristic quantity value is greater than zero, the rheological property feedforward tuning module confirms it as a valid positive acceleration signal and allows it to proceed; if the characteristic quantity value is less than or equal to zero, the module forcibly sets the variable to zero, indicating that the current rock mass is not in the accelerated fracture danger period. Through this validity constraint judgment mechanism, the system can actively filter out the deceleration characteristic variable generated during the calculation process, providing boundary-safe parameter input for the subsequent generation of deep coupling control law equations.

[0102] In traditional automated grouting projects, controllers typically adjust equipment power based solely on the current state, often failing to account for the physical spatial lag in material transport through pipelines. This time misalignment can cause the injected grout to miss the effective phase transformation and consolidation stage. To address these control deficiencies, in this embodiment, the rheological feedforward tuning module, based on cross-coupling theory, combines the fracture evolution state of rock mechanics with the spatial transport time lag of fluid dynamics to generate dynamic proportioning commands for real-time intervention in the gelation time of the two-liquid mixture. Specific implementation details of the deeply coupled control law generated by the rheological feedforward tuning module may include:

[0103] S540. During each control cycle, the rheological characteristic feedforward tuning module synchronously reads the rock mass fracture acceleration characteristic quantity calculated by itself. And the transmission delay time transmitted by the fluid time delay measurement module. These two characteristic parameters, belonging to different physical domains, serve as the basic input variables for the deeply coupled control law equations, providing a comprehensive judgment basis for the system that takes into account both the rock mass deterioration trend and the fluid transport state.

[0104] The S550 rheological property feedforward tuning module calls the deeply coupled control law equation according to the preset control algorithm. From the perspective of the intersection of materials rheology and control theory, this equation not only utilizes the characteristic quantities of rock mass fracture acceleration... A basic feedforward intervention loop was constructed, and a product term including a time compensation factor was introduced to correct the phase lag caused by spatial delay. The formula for calculating the target volume ratio command for the two-liquid mixture is as follows:

[0105] ;

[0106] In the formula, The target command for the two-liquid mixing volume ratio calculated at the current moment; This is the basic mixing volume ratio constant for a two-component grouting system; The feedforward control gain coefficient; This is a characteristic quantity of rock mass fracture acceleration; This is the time delay penalty compensation coefficient; This refers to the transmission delay time. It is a step function.

[0107] S560. To ensure the smooth execution of the aforementioned deeply coupled control law equations in industrial applications and reduce the risk of algorithm dead zones caused by abstract parameters, on-site technicians need to perform engineering definitions for various static calibration constants. As a preferred method, the basic mixing volume ratio constant... The feedforward control gain coefficient is usually determined based on the basic formula test of the grouting material on site, representing the benchmark mix ratio under normal working conditions without the risk of water inrush. The sensitivity of material mix design to rock mass fracture acceleration can be adjusted by referring to the rheological time sensitivity curves of specific accelerators obtained in the laboratory; time lag penalty compensation coefficient. This is used to weigh the immediate impact of transport lag on gelation time. The actual calibration process requires a reverse test that comprehensively considers the risk limit of premature coagulation and pipe blockage caused by the rapid-setting slurry in the pipeline.

[0108] S570. In actual calculations, the rheological characteristic feedforward tuning module executes advance compensation logic through the above equations. When the characteristic quantity of rock mass fracture acceleration... When an upward trend occurs and the system is currently operating at low traffic, the transmission latency time is... The value is relatively large. At this point, the lead compensation term within the formula will generate an amplified positive gain. The rheological property feedforward tuning module calculates the target command for the two-liquid mixing volume ratio. The command value for increasing the proportion of accelerator is accordingly increased. This computational mechanism actively reduces the intrinsic gel time of the mixed slurry, helping to offset the time consumed during long-distance pipeline transmission. However, from a chemirheological perspective, excessive addition of accelerator can easily lead to instantaneous solidification of the slurry at the mixer outlet, potentially causing a complete blockage of the high-pressure pipeline. To avoid this dead zone in the theoretical formula output, the rheological characteristic feedforward tuning module incorporates anti-blockage limiting logic before the output command.

[0109] S580. Regarding the output of control commands, the rheological characteristic feedforward tuning module has a preset maximum allowable mixing ratio threshold. This threshold is typically defined by technicians based on the upper limit of the mixing ratio corresponding to the extremely short safe pumpable time of the specific two-component material at the pumping pressure. The rheological characteristic feedforward tuning module will then calculate the target mixing volume ratio command for the two-component mixture. The calculated value is compared with the maximum permissible mixing ratio threshold. If the calculated value exceeds the threshold, the output command is truncated and locked to the maximum permissible mixing ratio threshold. After limiting, the rheological characteristic feedforward tuning module sets the final target two-liquid mixing volume ratio command. The signal is converted into corresponding electronic control signal parameters. For the underlying analog-to-digital conversion process, those skilled in the art can call the built-in analog-to-digital conversion function block of the industrial controller to execute it. The hardware conversion and drive protocol are well-known technologies in the field and will not be elaborated upon here. The rheological characteristic feedforward tuning module continuously outputs the converted control signal to drive actuators such as the front-end dual-path proportional control valve to change the physical flow channel state, thereby cooperating with the pressure control loop to construct a collaborative sealing system against sudden water inrush disasters.

[0110] After calculating the control law instructions, the digital signals are converted into the physical displacement of the underlying mechanical actuator. During this process, a hardware-level protection mechanism is implemented, which is of practical significance for maintaining the safety of grouting projects under complex geological conditions. In this embodiment, the collaborative handling controller not only drives the dual-path proportional control valve to dynamically change the material ratio, but also establishes a forced self-locking protection mechanism based on the underlying hardware status to cope with potential overpressure or mechanical blockage in the physical flow channel. The specific implementation details of the mechanical response and self-locking protection of the dual-path proportional valve actuator may include:

[0111] The S590 and collaborative handling controller send the target two-liquid mixing volume ratio command, ultimately determined by the rheological characteristic feedforward tuning module, to the servo driver of the dual-channel proportional control valve via an analog output channel or industrial fieldbus. The dual-channel proportional control valve internally houses an electric actuator and a mechanical transmission link. Upon receiving the electrical control signal, the servo driver drives the electric actuator to generate displacement, which in turn drives the mechanical transmission link to synchronously change the valve core opening within the two independent feed branches for the base slurry and the accelerator. By altering the effective cross-sectional area of ​​the physical flow channel, the dual-channel proportional control valve achieves a dynamic proportional distribution of the two fluid intake volumes at the physical level, thereby responding to the rheological adjustment requirements at the algorithm level.

[0112] S591. In the high-pressure, pulsating fluid environment of industrial sites, the mechanical structure of regulating valves is subject to the objective risk of jamming due to slurry particles. To address the execution deviation between control commands and actual mechanical responses, a displacement feedback sensor is installed at the valve shaft of the dual-channel proportional regulating valve. The collaborative handling controller reads the actual mechanical opening data fed back by this displacement feedback sensor in real time and converts it into an equivalent actual physical mixing ratio. As a preferred method, the system compares the actual physical mixing ratio with the theoretical target command for the two-liquid mixing volume ratio. When the numerical deviation exceeds the preset allowable error dead zone, and the duration of this deviation exceeds the set delayed confirmation time window, the collaborative handling controller determines that the underlying actuator has mechanically jammed, thereby suspending the rheological feedforward regulation logic and triggering an equipment abnormality alarm to the external system. Regarding the value of the allowable dead zone, on-site technicians usually combine the mechanical gear clearance of the dual-path proportional control valve itself and the measurement accuracy of the displacement sensor for comprehensive calibration; while the length of the delayed confirmation time window is set by reverse calculation based on the inertial response time of the servo drive motor under maximum load, so as to avoid false alarms caused by transient lag during normal adjustment.

[0113] S592. Based on the potential for localized premature setting and pressure buildup caused by high-concentration, rapidly setting grout at specific flow rates, the system incorporates a higher-priority hardware self-locking protection mechanism on top of the mechanical response architecture. Although the pre-processing algorithm includes a logical soft limit for the maximum permissible mixing ratio threshold, abnormal increases in actual grouting pressure within the discharge pipeline can still occur when geological conditions change abruptly. In this situation, the collaborative handling controller collects the actual pumping pressure feedback from the physical pressure sensor installed at the grouting manifold of the dual-liquid grouting pump in parallel. The system has a pre-calibrated hardware fusing pressure threshold characterizing the pipeline's ultimate bearing capacity. Regarding the range of this hardware fusing pressure threshold, on-site technicians typically use 80% of the rated burst pressure of the high-pressure grouting pipeline as a back-calculation setting to reserve a safety margin for the project.

[0114] S593. Whenever the actual pumping pressure reaches or exceeds the hardware fuse pressure threshold, the interrupt protection program at the bottom layer of the collaborative handling controller is activated first. This hardware interrupt protection program has the highest execution priority during the system operation cycle. Once triggered, the collaborative handling controller directly cuts off the control of the rheological characteristic feedforward tuning module over the dual-channel proportional control valve and forcibly sends a preset reference safety ratio electrical signal to the servo driver. In specific implementation, this reference safety ratio electrical signal drives the mechanical linkage to perform the closing action of the quick-setting agent inlet slurry channel, so that the actuator only retains the single-liquid-based slurry or clean water pumping state. Based on the analysis of the grouting engineering maintenance principle, the residual dual-liquid mixture inside the high-pressure grouting pipeline is quickly flushed out using single liquid or clean water to avoid overall solidification and scrapping of the pipeline due to direct emergency stop. At this time, the dual-channel proportional control valve enters the mechanical self-locking state, maintaining the opening of a single channel. Once this self-locking state is triggered, the system blocks all subsequent dynamic proportioning intervention commands until on-site technicians complete the safe depressurization of the pipeline and manually execute the self-locking reset command through the controller panel. Only then will the system resume closed-loop feedforward control. Through this self-locking protection mechanism, the system can effectively reduce the engineering risk of high-pressure pipeline rupture and help compensate for the logical blind spots of pure software algorithms when dealing with sudden physical blockages.

[0115] Specific application examples and experimental results comparison:

[0116] Project background and characteristic parameter settings:

[0117] This embodiment uses the construction process of a deep-buried tunnel traversing a water-rich fault fracture zone (segment K12+450) as its background. This section has fractured rock mass and an extremely high risk of water and mud inrush. The system deployed pore water pressure gauges and multi-point displacement gauges on-site, and included a high-pressure grouting pipeline with a dual-liquid grouting pump (with known internal fixed volume parameters). Connect to the working face.

[0118] The basic static calibration parameters for system operation are set as follows: pore water pressure attenuation constant. Calibrated as a slight background fluctuation value; basic mixing volume ratio constant. Set to 1.0 (i.e., the volume ratio of base slurry to accelerator is 1:1); feedforward control gain coefficient With time delay penalty compensation coefficient Determined based on on-site material rheological tests.

[0119] See attached document Figure 3 , Figure 3 This is a comparison chart of the evolution of early warning features and the effect of collaborative handling according to a specific application embodiment of the present invention. Figure 3 a is the evolution state diagram of the characteristic derivatives of the multiphysics field. Figure 3 b is a graph showing the evolution of early warning characteristic indices and rupture acceleration. Figure 3 c is a comparison diagram of feedforward intervention control laws for rheological properties. Figure 3 Figure d shows a comparison of the actual consolidation phase transformation state of the grout at the tunnel face. Using the time axis (minutes) as the horizontal axis, this figure fully illustrates the differences in data evolution and control effects between the proposed method and the traditional fixed-ratio method, from the initiation of micro-fractures in the rock mass to the macroscopic interconnection stage.

[0120] Response and dynamic early warning of water inrush (combined with integral model):

[0121] exist During the normal tunneling phase of minutes, the surrounding rock structure is stable, and the first derivative of strain is... First derivative of pore water pressure All fluctuate slightly around zero. At this point, the step function in the integral model... and If the result is deemed false, the early warning feature calculation module will force the integral to zero.

[0122] when At a certain time, the rock mass designed to prevent rock bursts enters the microfracture expansion phase. For example... Figure 3 As shown in Figure a, the solid line represents the first derivative of strain. It shows an accelerating upward trend (indicating intensified rock mass deformation), while the dotted line with circles represents the first derivative of pore pressure. A significant negative drop occurred and it fell below... (This characterizes the short-term decrease in local pore pressure caused by expansion and water absorption).

[0123] At this point, the underlying logic gates of the system simultaneously meet the triggering conditions, and the early warning feature calculation module continuously calls the formula:

[0124] ;

[0125] like Figure 3 The solid line representing b indicates the gradient asynchronicity exponent. The energy is rapidly accumulated over time. Based on this, the grouting energy distribution module switches the equipment operating mode to a low-pressure, high-flow-rate infiltration mode.

[0126] Deeply Coupled Control Law Intervention and Effect Verification (Combined with Feedforward Equations):

[0127] when At the minute mark, the rock mass approached the critical period of breakthrough. For example... Figure 3 The dashed line in b represents the rock mass fracture acceleration characteristic quantity extracted by the system's secondary analysis. A sharp increase occurred. Simultaneously, the grouting energy distribution module significantly reduced the target total flow command in order to establish the sealing back pressure. This leads to the transmission delay time calculated by the fluid delay measurement module based on the transmission delay analytical equation being... It has become noticeably longer.

[0128] Under this extreme condition, the rheological characteristic feedforward tuning module activates the deeply coupled control law equation:

[0129] ;

[0130] like Figure 3 As shown in c:

[0131] Traditional schemes (dotted lines) always maintain a fixed basic mixing volume ratio. It failed to respond to rock mass evolution and pipeline delays.

[0132] The present invention (solid line) is based on the above formula, at the fracture acceleration Increase and time delay Under the extended dual positive gain, the command curve shows a significant feedforward surge. The rheological property feedforward tuning module quickly drives the dual-path proportional control valve to increase the proportion of accelerator addition to near the set maximum allowable mixing ratio threshold, actively reducing the intrinsic gel time of the mixed slurry.

[0133] Comparison and verification of treatment effects:

[0134] like Figure 3 As shown in figure d, this figure represents the rheological state when the grout reaches the fracture surface of the rock mass (the higher the value, the greater the consolidation strength).

[0135] When using the traditional method (dashed line), because the transmission time delay of the physical pipeline is not considered, when the rock mass is... When a macroscopic rupture occurs within minutes, the grout reaching the working face is still in a liquid flow stage, making it extremely susceptible to being washed away by sudden high-pressure water inrush. The consolidation state is severely delayed, leading to a water inrush disaster.

[0136] When employing the scheme of this invention (solid line), thanks to the time-look-ahead compensation brought about by the deep coupling control law, the chemical gel phase transition point of the grout, prepared with a high concentration of accelerator, is precisely shifted and aligned to the moment of macroscopic rock fracture after a long pipeline journey. The grout rapidly completes the solid-state phase transition the instant it reaches the fracture surface, successfully seizing the extremely narrow effective phase transition consolidation stage, and achieving reliable and coordinated blocking of water inrush disasters in water-rich fractured zones.

Claims

1. An emergency early warning and collaborative response system for construction in water-rich fault fracture zones of tunnels, characterized in that, include: The sensor network device includes a pore water pressure gauge that collects raw time-series data of pore water pressure and a multi-point displacement gauge that collects raw time-series data of deformation displacement. Grouting execution equipment, including a dual-liquid grouting pump and a dual-path proportional regulating valve; Data transmission and computing device, including a collaborative processing controller; the collaborative processing controller is used to perform: Polynomial smoothing and differentiation operations are performed on the original time series data of pore water pressure and the original time series data of deformation displacement to extract the first derivative of pore water pressure and the first derivative of strain. The gradient asynchronicity index and its evolution rate are calculated based on the first derivative of the pore water pressure and the first derivative of the strain. Based on the gradient asynchronicity index and its evolution rate, a target total grouting pressure command and a target total flow rate command are generated and sent to the dual-liquid grouting pump; The transmission delay time is calculated by combining the internal fixed volume parameters of the high-pressure grouting pipeline with the target total flow command; The gradient asynchronicity index is subjected to a second derivative operation to extract the rock mass fracture acceleration characteristic quantity. The rock mass fracture acceleration characteristic quantity is combined with the transmission time delay to generate a target command for the two-liquid mixing volume ratio and output to the dual-path proportional control valve.

2. The emergency early warning and collaborative response system for tunnel construction in water-rich fault fracture zones according to claim 1, characterized in that, The process of performing polynomial smoothing and differentiation operations on the original time-series data of pore water pressure and deformation displacement to extract the first derivative of pore water pressure and the first derivative of strain includes: A causal sliding time window is allocated and maintained in memory, which stores only the current moment and historical sampling data before the current moment. The data points within the causal sliding time window are subjected to dimensionality reduction mapping of relative time coordinates, and an observation matrix in the local coordinate system is constructed based on the mapped relative time coordinates. The coefficients of the polynomial equations are solved directly based on the principle of normal equations using the least squares method, and the first-order analytic derivative of the polynomial equations after determining the coefficients is then performed. Substitute the relative time coordinate value corresponding to the latest current moment into the first derivative equation to extract the first derivative of pore water pressure and strain without time delay at the latest current moment.

3. The emergency early warning and collaborative response system for construction in water-rich fault fracture zones of tunnels according to claim 1, characterized in that, The calculation of the gradient asynchronicity index based on the first derivative of the pore water pressure and the first derivative of the strain includes: In the underlying logic gate operation, it is determined in parallel whether the first derivative of the current strain is greater than zero and whether the first derivative of the current pore water pressure is less than the set pore water pressure attenuation constant. Only when both determination results are true, the product of the current strain first derivative and the difference between the pore water pressure attenuation constant and the current pore water pressure first derivative is extracted, and the product is used as the effective discrete divergence feature value of the current control cycle. If any determination result is false, the effective discrete divergence feature value will be forcibly set to zero; The effective discrete divergence feature values ​​are pushed into the tail of the first-in-first-out queue, and the oldest historical data removed from the head of the first-in-first-out queue is released simultaneously. All effective discrete divergence feature values ​​in the current first-in-first-out queue are traversed and multiplied by the single scan cycle time, and the results are summed to calculate the gradient asynchronicity index of the current control cycle.

4. The emergency early warning and collaborative handling system for tunnel construction in water-rich fault fracture zones according to claim 3, characterized in that, The collaborative processing controller also includes anti-saturation limiting protection logic, which is used for: A preset anti-lock upper limit threshold is set, which is calculated and calibrated by combining the maximum safe working pressure limit of the dual-liquid grouting pump with the proportional gain parameter of the subsequent closed-loop controller. After each discrete accumulation operation is completed, the calculated gradient asynchronicity index is compared with the anti-lock upper limit threshold. If the calculated gradient asynchronicity index exceeds the anti-lock upper limit threshold, the calculation is directly truncated and the gradient asynchronicity index is locked to the anti-lock upper limit threshold.

5. The emergency early warning and collaborative response system for construction in water-rich fault fracture zones of tunnels according to claim 1, characterized in that, The step of generating the target total grouting pressure command and the target total flow command based on the gradient asynchronicity index and its evolution rate includes: The risk deviation of the current grouting condition is obtained by subtracting the received gradient asynchronicity index from the internally stored zero reference value. The zero reference value is set to an absolute zero value or the long-term average of the background characteristic data when the construction site is undisturbed and there is no risk of water seepage. The pressure regulation loop is operated, and the risk deviation and the evolution rate are converted into pressure increments to compensate for rock mass fracture and expansion using proportional-integral-differential operation rules. The target total grouting pressure command is generated by superimposing the basic pressure-maintaining base. The operation flow regulation loop, when the gradient asynchronicity index and its evolution rate cross the set critical warning threshold, rapidly reduces the target total flow command value through an active negative feedback mechanism. The critical warning threshold is determined by combining the tensile strength and porosity parameters of the anti-outburst rock mass obtained in the early geological exploration stage with fluid-structure interaction numerical simulation to obtain the theoretical upper limit of the deviation index at macroscopic fracture.

6. The emergency early warning and collaborative response system for construction in water-rich fault fracture zones of tunnels according to claim 1, characterized in that, The calculation of transmission delay time, based on the internal fixed volume parameters of the high-pressure grouting pipeline, includes: Pre-set the minimum traffic security threshold and the corresponding maximum allowable physical latency limit; The minimum flow safety threshold is determined by back-calculation based on the mechanical displacement corresponding to the minimum stable frequency conversion operating frequency of the main drive motor of the dual-liquid grouting pump under the premise of no stall or overheating. The maximum permissible physical time delay is calculated and calibrated in reverse based on the initial setting time of the current two-component formula under no external disturbance and with a safety factor reserved to prevent pipeline blockage. The target total traffic command read is compared with the minimum traffic safety threshold. If the target total traffic instruction is greater than or equal to the minimum traffic safety threshold, the internal fixed volume parameter is divided by the target total traffic instruction to obtain the quotient value, and the quotient value is used as the transmission delay time. If the target total flow instruction is lower than the minimum flow safety threshold, the division operation is directly stopped, and the preset maximum allowable physical delay limit is forcibly output.

7. The emergency early warning and collaborative response system for construction in water-rich fault fracture zones of tunnels according to claim 1, characterized in that, The method of generating a target command for the two-liquid mixing volume ratio by combining the rock mass fracture acceleration characteristic quantity with the transmission time delay includes: The rock mass fracture acceleration characteristic value calculated by itself and the transmission time delay are read as basic input variables, and the deep coupling control law equation is called. The deep coupling control law equation includes a basic hybrid volume proportionality constant, a feedforward control gain coefficient, a time delay penalty compensation coefficient, and a step function. The advanced compensation logic is executed through the deep coupling control law equation to calculate and generate the theoretical target command for the two-liquid mixing volume ratio. The calculated theoretical target volume ratio of the two liquids is compared with the maximum allowable mixing ratio threshold. If the calculated value exceeds the maximum allowable mixing ratio threshold, the output command is directly truncated and locked to the maximum allowable mixing ratio threshold. The maximum permissible mixing ratio threshold is calibrated based on the upper limit of the mixing ratio corresponding to the extremely short safe pumpable time of the specific two-component material under the pumping pressure.

8. The emergency early warning and collaborative response system for construction in water-rich fault fracture zones of tunnels according to claim 1, characterized in that, The step of performing a second derivative operation on the gradient asynchronicity exponent to extract the rock mass fracture acceleration feature includes: The read evolution rate is smoothed using a first-order inertial low-pass filter; Perform time difference operation on the smoothed evolution rate to obtain the feature value; The calculated characteristic value is subjected to directional validity constraint judgment. If the characteristic value is greater than zero, it is allowed to proceed and the rock mass fracture acceleration characteristic value is confirmed. If the characteristic value is less than or equal to zero, the characteristic value is forcibly set to zero.

9. The emergency early warning and collaborative response system for tunnel construction in water-rich fault fracture zones according to claim 1, characterized in that, The collaborative processing controller also includes hardware self-locking protection logic for: The actual pumping pressure is collected in parallel from the physical pressure sensor installed at the slurry manifold of the dual-liquid grouting pump. Whenever the actual pumping pressure collected touches or exceeds the hardware fuse pressure threshold, the underlying interrupt protection program is activated first to directly cut off the control of the dual-path proportional regulating valve. The hardware fusing pressure threshold is set by back-calculating 80% of the rated burst pressure of the high-pressure grouting pipeline. A preset reference safety ratio electrical signal is forcibly sent to the servo driver of the dual-path proportional regulating valve to execute the closing action of the accelerator slurry inlet channel, so that the actuator only retains the single liquid base slurry or clear water pumping state and enters the mechanical self-locking state.

10. The emergency early warning and collaborative response system for construction in water-rich fault fracture zones of tunnels according to claim 1, characterized in that, The collaborative processing controller also includes electromechanical jamming detection logic, used for: The actual mechanical opening data fed back by the displacement feedback sensor installed at the valve shaft of the dual-channel proportional control valve is read in real time and converted into the actual physical mixing ratio; The actual physical mixing ratio is compared with the theoretical target command for the two-liquid mixing volume ratio. When the actual physical mixing ratio deviates from the target command of the two-liquid mixing volume ratio by more than the preset allowable error dead zone, and the duration of the deviation exceeds the set delay confirmation time window, it is determined that the underlying actuator has mechanically jammed, the rheological feedforward adjustment logic is stopped and the equipment abnormality alarm is triggered. The allowable error dead zone is comprehensively calibrated by combining the mechanical gear backlash of the dual-path proportional control valve and the measurement accuracy of the displacement sensor; The length of the delayed confirmation time window is set by inversely calculating the inertial response time of the servo drive motor under maximum load.

Images