Adaptive rectifying control method for deep cooling nitrogen production under variable load conditions and related equipment

Abstract

The present application relates to the technical field of bulk gas, in particular to a deep cold nitrogen production variable load condition adaptive rectification control method and related equipment. The steps of the control method include: real-time acquisition of multi-source dynamic parameters of the rectification system, including the temperature gradient distribution matrix of each flow channel outlet of the main heat exchanger, the change amount of the oxygen-rich liquid air component concentration of the tower kettle, the turbulent flow intensity coefficient of the outlet gas phase of the expander, and the fluctuation rate of the rock mass shear stress fed back by the geological monitoring module. Through the geological structure stress field reconstruction algorithm and the multiphase flow mass transfer coupling model, the quantitative mapping of the rock mass dynamics behavior to the rectification working condition characteristics is realized, specifically, the stratum dislocation factor is coded into the non-steady state working condition characteristic tensor, so that the rectification system can perceive the energy accumulation on the stress wave propagation path in advance.

Description

Technical Field

[0001] This invention relates to the field of bulk gas technology, specifically to an adaptive distillation control method and related equipment for cryogenic nitrogen production under variable load conditions. Background Technology

[0002] Under varying load conditions, the gas-liquid two-phase equilibrium of the distillation column in a cryogenic air separation nitrogen production system is susceptible to latent disturbances caused by tectonic movements. Existing control strategies primarily address process parameter fluctuations, neglecting the cross-scale coupling effect between the geological stress field and distillation mass transfer. Discontinuous fluctuations in rock mass shear stress are transmitted to the distillation system through the column base, inducing temperature gradient distortion in the main heat exchanger, aerodynamic instability in the expander, and chaotic oscillations in the bottom component concentration. This combined disturbance causes parameter drift in traditional material balance-based control models. In particular, when the microseismic frequency band of the rock strata resonates with the natural frequency of the distillation column, it leads to a sharp decrease in fractionation efficiency and even loss of control over liquid nitrogen purity. Therefore, the technical problem proposed in this invention is: how to establish a real-time decoupling mechanism for geological-process coupled disturbance sources and reconstruct an adaptive control architecture under varying load conditions. Summary of the Invention

[0003] This disclosure proposes an adaptive distillation control method and related equipment for cryogenic nitrogen production under variable load conditions, aiming to overcome at least one defect in the prior art.

[0004] To achieve the above objectives, the technical solution disclosed in this invention is as follows:

[0005] According to one aspect of this disclosure, an adaptive distillation control method for cryogenic nitrogen production under variable load conditions is provided, the control method comprising the following steps:

[0006] Real-time acquisition of multi-source dynamic parameters of the distillation system, including the temperature gradient distribution matrix of each channel outlet of the main heat exchanger, the change in the concentration of oxygen-enriched liquid air components in the bottom of the column, the gas phase turbulence intensity coefficient at the outlet of the expander, and the rock mass shear stress fluctuation rate fed back by the geological monitoring module.

[0007] By using a geological structural stress field reconstruction algorithm and a multiphase mass transfer coupling model, the multi-source dynamic parameters are encoded into unsteady-state characteristic tensors that include formation faulting factors.

[0008] A spatiotemporal separation convolutional neural network is used to decouple the load mutation component of the working condition feature tensor. A tunable bandpass filter is used to extract the microseismic spectrum features of the rock strata. At the same time, wavelet packet analysis is used to analyze the inherent wave components of the stripping process.

[0009] Based on the bifurcation control theory to drive distillation rebalancing, the safety margin of the reboiler thermosiphon pressure difference is dynamically set according to the decoupled load mutation component, and the chaotic suppression signal of the condenser refrigerant feedforward injection is generated by combining the nitrogen Raman spectrum offset at the top of the column.

[0010] The gas-liquid two-phase flow cooperative controller is constructed, and the chaotic suppression signal is converted into a dynamic coupling instruction of the lower column backflow ratio regulation gradient and the upper tower oxygen-rich liquid air throttle valve opening degree, which is used to maintain the liquid nitrogen purity calibration value of the rectifying column under the condition of rock mass stress fluctuation.

[0011] Further, the geological structure stress field reconstruction algorithm and the multiphase flow mass transfer coupling model perform the following steps:

[0012] Based on the principle of crustal strain energy release, a vibration-rectification mass transfer correlation function is established:

[0013] Wherein, G(t) is a vibration-rectification mass transfer correlation function, used to quantify the coupling strength between rock mass vibration energy release and rectifying column pressure field, Ω is the three-dimensional space range of the rock mass, is the rock mass displacement vector field, and β is the empirical damping coefficient, is the rectifying column pressure gradient distribution field;

[0014] The abnormal diffusion flux of the gas phase component is solved by the non-equilibrium statistical mechanics equation:

[0015] Wherein, D ki eff is the stress-sensitive tensor diffusion coefficient, is the concentration gradient of component k in the i direction, is the second-order mixed derivative of volume strain with respect to time t and spatial coordinate x j , γ k is the strain energy response coefficient of component k.

[0016] Further, the bifurcation control theory executes a variable order control law:

[0017] Wherein, Δu(t) represents the control amount increment of the controller output, K p is the proportional coefficient, τ d is the differential time constant, α is the correction factor, and ν is the fractional order related to the Poisson ratio of the rock mass, which is used to control the parameters K p , τ d According to the online optimization of tower body vibration modal energy distribution, e(t) is the load deviation signal.

[0018] Further, the synergic controller performs the following operations: a high-resolution piezoelectric ceramic array sensor is integrated at the liquid oxygen reflux pump outlet flange for real-time monitoring of fluid micro-vortex energy spectrum density; when the vortex shedding frequency is detected to exceed the rock mass base frequency, a dual-mode anti-disturbance mechanism is triggered, the reflux pump driving power is raised to a set threshold and maintained for an integer number of oscillation periods, and a two-stage pre-cooled liquid methane anti-phase change buffer is injected into the condenser evaporation cavity at the same time; the injection flow rate of the buffer is negatively correlated with the formation longitudinal wave velocity, and sub-micron aerosol is formed by ultrasonic atomizing nozzles to ensure full coverage of the condenser tube bundle.

[0019] Further, the geological monitoring module includes: a borehole strain gauge array arranged according to the Mohr-Coulomb failure criterion, with a drilling depth exceeding the bottom boundary of the weathering layer; a microseismic wave pickup of a three-component geophone group, each geophone unit being equipped with a constant-temperature compensation type signal amplifier; the sampling frequency is dynamically adjusted according to the background noise cross-correlation function, the adjustment range covering the rock mass failure characteristic frequency band, and the industrial vibration interference component is real-time stripped by Hilbert-Huang transform.

[0020] Further, the load mutation component decoupling includes condenser cold end optimization, the steps including: establishing a fractal control model of the temperature difference field of the finned tube bundle, iteratively calculating the evolution path of the frost phase interface with the Hausdorff dimension change rate as the index, and dynamically adjusting the opening sequence of the refrigerant distributor; when the fractal dimension deviates from the reference value by more than the critical amplitude, a liquid nitrogen backwashing defrosting program is started, the liquid nitrogen at the bottom of the tower is injected into the refrigerant distribution main pipe after being boosted in pressure, and the corresponding branch refrigerant supply is cut off at the same time, the defrosting duration is positively correlated with the integral of the deviation of the fractal dimension.

[0021] Further, the control method further includes implementing oxygen-enriched liquid air throttle valve stress corrosion protection, the steps including: depositing an amorphous Ni-W-P ternary alloy coating on the valve core sealing surface, the coating thickness being set in stages according to the hydrogen sulfide concentration of the fluid before the valve; applying a reverse polarization voltage during the valve action period, the voltage amplitude being dynamically calculated according to the oxygen content sensor feedback value by the Nernst equation; activating the in-situ polishing program after every set period of operation, driving the diamond grinding head to perform spiral trajectory cutting along the valve seat cone surface, and the cutting depth being controlled by a residual stress probe closed loop.

[0022] Further, the control method further includes heat exchanger defrosting optimization, the steps including: calculating the dynamic mapping relationship between the fin frosting rate and the thermal conductivity in real time according to the rock layer water content data fed back by the ground resistivity probe; predicting the optimal working time window of each defrosting interval through a finite element thermal-flow coupling model to control the partitioned start and stop of the electric heating belt; when a fault activation signal is detected, the full-area pulse strong heat defrosting mode is triggered in advance: the power of the frequency conversion power supply is instantaneously raised to the rated value and maintained for a set period of time, and the defrosting cycle counter is reset after defrosting is completed.

[0023] Further, when the three-component detector group detects that the shear strain rate exceeds the threshold, a three-level interlock response is activated, including: starting the rectifying column emergency depressurization module: sequentially opening the overhead safety valve and the bottom liquid discharge valve, and the depressurization rate is controlled by a PID controller; switching to the standby expander group: disconnecting the output shaft of the faulty expander group through a hydraulic clutch, while starting the magnetic suspension bearing standby group and soft starting; injecting a process stabilizer: injecting critical state carbon dioxide powder into the cold end of the main heat exchanger, and the injection amount is in logarithmic relationship with the fault slip distance.

[0024] Further, the dynamic determination of the fractional order includes: scanning the tower foundation rock surface using a non-contact laser Doppler vibrometer to collect Rayleigh wave group velocity dispersion curves; based on the anisotropic medium elastic wave equation, the instantaneous gradient value of the rock shear modulus is inverted; the modulus change rate is substituted into the preset nonlinear transfer function to generate an update instruction of the fractional order; when the initial arrival of the P wave is detected, the order update is frozen and switched to the preset anti-shock mode parameter group.

[0025] According to another aspect of the present disclosure, a deep cold nitrogen production variable load condition adaptive rectification control system is provided for implementing the deep cold nitrogen production variable load condition adaptive rectification control method as described above, and the control system comprises:

[0026] The geological disturbance sensing unit includes a borehole strain gauge array embedded in the stable rock layer of the tower foundation and a four-component microseismic monitor installed on the tower body pile cap, which is used to collect real-time multi-source dynamic parameters of the rectification system;

[0027] The multi-field coupling processing engine is equipped with an FPGA chip set and a hardware acceleration core embedded with a phase change coupling algorithm, which is used to generate a working condition characteristic tensor in real time, and to encode the multi-source dynamic parameters into a non-steady state working condition characteristic tensor including a stratum dislocation factor;

[0028] The load decoupling execution unit includes a tunable band-pass filter circuit group and a 16-channel wavelet packet decomposition module, which outputs digital signals of load mutation components and process fluctuation components, and is used to decouple the load mutation components from the working condition characteristic tensor;

[0029] The bifurcation control center integrates an analog circuit implementation module of a variable order controller, a chaotic signal generator, and a safety interlock logic unit, which is used to drive rectification rebalancing;

[0030] The two-phase flow coordination terminal includes a piezoelectric ceramic array signal demodulator, a proportional electromagnetic valve group of the liquid methane injection system, and a multivariable PID controller card, which is used to build a gas-liquid two-phase flow cooperative controller.

[0031] Further, the geological disturbance sensing unit comprises a corrosion-resistant packaging structure, the detector probe is sealed in a Hastelloy C-276 forged shell, and the shell is coated with a diamond-like carbon coating outside; signal transmission adopts a double-layer shielding coaxial cable, the inner layer is a silver-coated copper wire braid, and the outer layer is covered with a fluororubber low-temperature resistant sheath; the joint is filled with silicone grease and is sleeved with an armored heat-shrinkable tube to meet the IP68 protection level and the explosion-proof certification standard.

[0032] Further, the load decoupling execution unit is equipped with a redundant communication architecture, the main signal channel adopts an optical fiber reflective memory ring network, and the standby channel is a PROFIBUS-DP bus; a double abnormality detection mechanism is arranged, a CRC-32 check module is added at the output end of the FPGA, and a time series prediction algorithm is used to identify the waveform distortion of the load mutation component; when three consecutive frames of data fail to pass the check or the waveform distortion index exceeds a threshold, the communication channel is automatically switched and the wavelet packet decomposition process is restarted, and during the restart, historical data interpolation is used to maintain control continuity.

[0033] Further, the bifurcation control center comprises a thermal management system, a micro-channel copper cold plate is embedded in the FPGA chip heat dissipation substrate, and the cold plate is connected to a secondary cooling circuit; the cooling medium is a perfluoropolyether fluid, which is driven by a magnetic pump to flow through a finned heat exchanger and exchange heat with the exhaust gas of the expander; the control system dynamically adjusts the flow of the cooling liquid according to the chip junction temperature, and automatically reduces the frequency and starts the auxiliary semiconductor cooling module when the junction temperature is detected to exceed the limit.

[0034] Further, the two-phase flow coordination terminal is configured with a safety interlocking mechanism, double cut-off valves are arranged at the outlet of the liquid methane storage tank, and the burst pressures of the two-stage safety valves are respectively set to 1.1 times and 1.3 times of the design pressure; the frequency converter of the reflux pump is configured with a current harmonic filter and a regenerative braking unit; the position feedback of the actuator uses an absolute photoelectric encoder, and the encoder signal is transmitted through double CAN buses; the shell of the coordination terminal is a nitrogen-filled positive pressure explosion-proof box, and the box is provided with an explosion vent and a flame containment grid.

[0035] According to another aspect of the present disclosure, a deep cold nitrogen production variable load condition adaptive rectification control device is provided, which is integrated with the deep cold nitrogen production variable load condition adaptive rectification control system as described above, and further comprises:

[0036] A displacement-resistant tower foundation structure with a tower body support system connected by universal joints, comprising an axial corrugated pipe compensator and a shear-limiting base;

[0037] A shock-resistant electrical cabinet assembly, a partitioned explosion-proof control cabinet, which is provided with an SIL 3-level safety instrument system, and a cabinet positive pressure ventilation system provided with a dew point monitor and an automatic drain valve;

[0038] Online mass spectrometry monitoring module, four-pole mass spectrometer is installed on the top of nitrogen pipeline, ion source chamber is maintained ultra-high vacuum by magnetic suspension molecular pump, and detector output signal is connected to core processor through EMI filter.

[0039] According to still another aspect of the present disclosure, a computer-readable storage medium is provided, which stores executable program codes, when the program codes are read by a computer, a cryogenic nitrogen production variable load condition adaptive rectification control method as described above is implemented, and the steps further implemented include: when it is detected that the output spectrum of the rectification tower foundation accelerometer falls into the rock mass characteristic frequency band, activating the three-level control protocol, reconstructing the fractional order controller parameters; calling the dual-mode anti-disturbance mechanism; starting the tower base hydraulic damper against mode, and controlling the reciprocating movement amplitude and frequency of the damper piston according to the phase difference fed back by the laser range finder, to form a dynamic balance force opposite to the foundation vibration.

[0040] The present application has the following beneficial effects:

[0041] The present application realizes quantitative mapping of rock mass dynamic behavior to rectification condition characteristics by means of a geological structure stress field reconstruction algorithm and a multiphase flow mass transfer coupling model, specifically, the stratum dislocation factor is coded into the non-steady state condition characteristic tensor, so that the rectification system can perceive the energy accumulation on the stress wave propagation path in advance.

[0042] Further, a space-time separation convolutional neural network is used to decouple the geological disturbance and the process disturbance in the frequency domain, the rock stratum microseismic characteristic spectrum is stripped through wavelet packet analysis, and the interference of shear stress fluctuation on component concentration control is eliminated.

[0043] Further, a chaotic suppression signal is generated based on the bifurcation control theory, the dynamic coordination of the reboiler thermosyphon pressure difference and the transient response of the condenser refrigerant injection is realized; a gas-liquid two-phase flow cooperative controller is constructed, the geological disturbance is converted into the dynamic coupling instructions of the lower reflux ratio and the upper tower oxygen-rich liquid air throttle valve, and the control precision of the liquid nitrogen purity calibration value of ±0.1ppm is maintained under the condition of rock mass stress fluctuation. The present application has broken through the multi-physical field strong coupling control problem of geological structure movement and cryogenic rectification process.

[0044] The above description is only a summary of the technical scheme of the present application, in order to more clearly understand the technical means of the present application, and the content of the specification can be implemented, the following is a preferred embodiment of the present application and the detailed description with the help of the drawings. BRIEF DESCRIPTION OF DRAWINGS

[0045] Figure 1 The flow chart of the cryogenic nitrogen production variable load condition adaptive rectification control method in an embodiment of the present application;

[0046] Figure 2 The phase space diagram of the fractional order control system in an embodiment of the present application;

[0047] Figure 3 A response schematic diagram of a fractional order control system in an embodiment of the present application;

[0048] Figure 4 A dynamic adaptive model schematic diagram of a fractional order in an embodiment of the present application;

[0049] Figure 5 A stress corrosion protection surface treatment model schematic diagram in an embodiment of the present application;

[0050] Figure 6 A geological structure stress field reconstruction vector and component abnormal diffusion flux distribution schematic diagram in an embodiment of the present application;

[0051] Figure 7 A multi-source dynamic parameter monitoring distribution schematic diagram in an embodiment of the present application;

[0052] Figure 8 A geological structure stress field reconstruction schematic diagram in an embodiment of the present application;

[0053] Figure 9 A condenser fractal control model schematic diagram in an embodiment of the present application;

[0054] Figure 10 A bimodal anti-disturbance mechanism schematic diagram in an embodiment of the present application;

[0055] Figure 11 A geological monitoring system parameter configuration schematic diagram in an embodiment of the present application. DETAILED DESCRIPTION

[0056] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application.

[0057] The present application provides the following preferred embodiments:

[0058] Embodiment one: in order to solve the problem of rectification stability caused by coupling of variable load working condition and rock mass stress fluctuation in the process of cryogenic nitrogen production, the present embodiment provides a self-adaptive rectification control method for cryogenic nitrogen variable load working condition aiming at multi-source parameter fusion control mechanism, as shown in Figure 1 The flow of the control method is as follows:

[0059] S100: real-time acquisition of multi-source dynamic parameters of the rectification system, including temperature gradient distribution matrix of each flow passage outlet of the main heat exchanger, oxygen-rich liquid air component concentration change of the tower kettle, turbulent intensity coefficient of the expander outlet gas phase, and rock mass shear stress fluctuation rate fed back by the geological monitoring module.

[0060] S200: Through the geological structural stress field reconstruction algorithm and the multiphase mass transfer coupling model, the multi-source dynamic parameters are encoded into unsteady-state characteristic tensors including the stratum faulting factor.

[0061] S300: It uses a spatiotemporal separation convolutional neural network to decouple the load mutation components of the working condition feature tensor, uses a tunable bandpass filter to extract the microseismic spectrum features of the rock strata, and uses wavelet packet analysis to analyze the inherent wave components of the stripping process.

[0062] S400: Based on bifurcation control theory, the distillation rebalancing is driven. The safety margin of the reboiler thermosiphon pressure difference is dynamically set according to the decoupled load mutation component. The chaotic suppression signal of the condenser refrigerant feedforward injection is generated by combining the nitrogen Raman spectrum offset at the top of the column.

[0063] S500: Construct a gas-liquid two-phase flow collaborative controller to convert the chaos suppression signal into a dynamic coupling command between the lower column reflux ratio adjustment gradient and the upper column oxygen-enriched liquid air throttle valve opening, which is used to maintain the liquid nitrogen purity calibration value of the distillation column under rock mass stress fluctuation conditions.

[0064] Specifically, in the real-time acquisition of multi-source dynamic parameters, such as Figure 7 The monitoring distribution architecture shown employs a density of 0.5 units / cm² at the outlet of each flow channel of the main heat exchanger. 2 An array-type temperature sensor was used to construct a three-dimensional temperature gradient distribution matrix containing 12 axial monitoring sections and 5 radial temperature measurement points. The sampling period was set to 200ms to match the dynamic response characteristics of the flow channel. The change in the concentration of oxygen-enriched liquid air components in the bottom of the tower was monitored in real time by an online mass spectrometer at a frequency of 1Hz, and the rate of change of the mole fraction of oxygen and nitrogen components was obtained in seconds through differential calculation. The gas phase turbulence intensity coefficient at the expander outlet was based on the axial velocity pulsation signal collected by a laser Doppler velocimeter. The geological monitoring module... Figure 11 The stress sensor network shown comprises eight sets of triaxial stress gauges arranged in the four principal stress directions of the foundation rock mass. The regional rock mass shear stress fluctuation rate was obtained by fitting the data using Kriging interpolation. The sampling frequency was synchronized with the process parameters. It is important to understand that these parameters encompass the process heat and mass transfer characteristics and the dynamics of the geological environment, providing multi-dimensional input for coupled analysis.

[0065] Furthermore, the parameter encoding process is combined with Figure 6 The geological tectonic stress field reconstruction vector and component anomaly diffusion flux distribution are shown. The geological tectonic stress field reconstruction algorithm is based on the elastic foundation beam theory. Shear stress fluctuation rate and historical geostress data are input into the finite element model. The dimensionless factor characterizing the risk of stratum slippage is obtained through inversion calculation. The multiphase flow mass coupling model uses the temperature gradient matrix of the main heat exchanger as the boundary condition to establish the mass conservation equation and energy conservation equation for the gas-liquid two-phase flow in the distillation column.

[0066] , ; the time-space distribution function of the mass transfer unit height under non-steady state conditions is solved. After standardization, the two types of characteristics are spliced into a 100x40 non-steady state condition characteristic tensor in the time dimension and the space dimension according to 32 process monitoring points + 8 geological monitoring points.

[0067] In the working condition characteristic processing stage, the space-time separation convolutional neural network adopts a three-layer three-dimensional convolution structure, the first layer of convolution kernel size is 5x5x3, the load mutation related feature weight is strengthened through the channel attention mechanism, and the decoupling extraction of the load mutation component AL(t) in the characteristic tensor is realized; as shown in the tunable band-pass filter module, according to the preset rock stratum microseismic characteristic frequency range of the geological monitoring, the frequency domain screening of the decoupled vibration signal is carried out, and the frequency spectrum component related to the rock mass microseism is extracted; the wavelet packet analysis adopts db4 wavelet basis for 3-layer decomposition, and the process inherent fluctuation component with a period of 10-30s is stripped through the reconstruction algorithm, so as to ensure the purity of the subsequent control signal. Figure 4

[0068] Further, in the rectification and rebalancing process based on the bifurcation control theory, as shown in the fractional order control phase space reconstruction, Figure 2 the bifurcation control law adopts the formula:

[0069] , wherein K p is a proportional coefficient, τ d is a differential time constant, α is a correction factor, and ν is a fractional order degree related to the Poisson ratio of the rock mass, which is used to control the parameters K p , τ d According to the online optimization of the tower vibration modal energy distribution, e(t) is the load deviation signal. In the variable order control law, Δu(t) represents the control amount increment of the controller output, that is, the dynamic adjustment amount for adjusting the system control action calculated in real time according to the current load deviation signal. According to the amplitude of the decoupled Δu(t), the differential pressure safety margin ΔP th is dynamically adjusted through the fluid mechanics model of the reboiler thermosyphon cycle, the margin and the reboiler liquid level height h and the temperature difference ΔT satisfy the nonlinear relationship ΔP th =f(h,ΔT); at the same time, the frequency shift Δν monitored by the tower top nitrogen Raman spectrometer is used to generate the condenser coolant feedforward injection signal u c (t), which is used to suppress the chaotic oscillation trend caused by the load mutation.

[0070] Further, the gas-liquid two-phase flow cooperative controller is constructed as shown in Figure 10 ​The dual-mode disturbance rejection mechanism shown converts the chaotic suppression signal into the lower column reflux ratio adjustment gradient ∇R(t) and the upper column throttle valve opening command Δθ(t). It employs a two-variable decoupled control strategy, with the reflux ratio adjustment based on the lower column gas-liquid balance equation:

[0071] The critical equilibrium adjustment amount is calculated by combining the pressure P and temperature T distribution within the tower; the throttling valve opening control models the oxygen-enriched liquid air throttling process as a first-order process with hysteresis.

[0072] The predictive control algorithm achieves dynamic matching between opening adjustment and reflux ratio change, ensuring that the purity of liquid nitrogen remains at the calibrated value when rock mass stress fluctuates.

[0073] The advantage of this embodiment is that, through the systematic integration of a multi-dimensional parameter monitoring system, geological-process coupled feature encoding, fractional-order bifurcation control strategy, and dual-modal collaborative adjustment mechanism, an adaptive control closed loop covering parameter acquisition, feature processing, and control command generation is formed, which effectively copes with the combined effects of variable load conditions and geological disturbances, and provides a structured control scheme for the stable operation of the cryogenic nitrogen distillation system.

[0074] Example 2: In order to solve the problem of instability caused by the coupling effect of geological stress fluctuation and distillation mass transfer process in cryogenic nitrogen production, this example further refines the vibration-distillation correlation mechanism and the solution logic of abnormal diffusion flux for the geological structure stress field reconstruction algorithm and multiphase mass transfer coupling model.

[0075] Furthermore, when constructing the vibration-distillation mass transfer correlation function based on the principle of crustal strain energy release, such as Figure 6 The diagram showing the reconstructed vector of the geological structural stress field and the distribution of component anomaly diffusion flux is obtained through a three-dimensional stress sensor network deployed on the foundation rock mass of the device, such as... Figure 11 The geological monitoring system parameter configuration architecture shown indicates that it collects rock mass displacement vector fields in real time. The field quantity was spatially interpolated and fitted using the Kriging interpolation method to eight sets of triaxial stress gauge data, forming a regional displacement distribution model containing 12 monitoring nodes. (Distillation column pressure gradient distribution field) The spatial resolution is calculated using the finite volume method based on real-time data from 16 pressure sensors within the tower (8 sets in the upper tower, 6 sets in the lower tower, and 2 sets in the reboiler), achieving a resolution of 0.5 meters per node. The correlation function is:

[0076] Where G(t) is the vibration-distillation mass transfer correlation function, used to quantify the coupling strength between the vibration energy release of the rock mass and the pressure field of the distillation column, Ω is the integration region, representing the rock mass region in three-dimensional space, i.e., the three-dimensional spatial range of the rock mass, and the second-order time derivative term. Characterizing the distribution of rock mass vibration acceleration, exponential decay factor This method is used to quantify the damping effect of the distillation pressure field on the transmission of vibrational energy, where β is an empirical damping coefficient determined by fitting historical data. The integral operation is performed in a Cartesian coordinate system, dividing the rock mass into 20×20×15 three-dimensional grid cells. The Gaussian integral method is used to solve for the coupling strength between regional strain energy release and the distillation pressure field.

[0077] Furthermore, the solution process for the anomalous diffusion flux of gas-phase components incorporates non-equilibrium statistical mechanics theory, such as... Figure 8 The diagram shown illustrates the reconstruction of the geological stress field, with the stress-sensitive tensor diffusion coefficient D. ki eff Calibrated by rock mechanics experiments, its value varies with the rock mass shear stress τ. s It exhibits nonlinear variation:

[0078] When τ s <τ cr At (critical shear stress), When τ s ≥τ cr At that time, a crack propagation correction term is introduced:

[0079] , where D ki 0 Let k be the initial diffusion coefficient. d m is the linear stress sensitivity coefficient. f This is a factor influencing fracture development. The diffusion flux expression is:

[0080] The first term is the concentration gradient-driven Fick diffusion term, and the second term is the strain energy change rate coupling term, γ k ϵ is the strain energy response coefficient of component k, reflecting the dynamics of geological strain energy. v Additional driving force for gas phase diffusion. Let i be the concentration gradient of component k along the i-th direction. For volumetric strain versus time t and spatial coordinate x j The second-order mixed derivative, spatial derivative term and The solution is obtained by using the central difference method and the second-order spatiotemporal precision scheme respectively, ensuring that transient change features are captured within the sampling period.

[0081] It needs to be understood that the coupling logic of the above two expressions lies in that the correlation function establishes the space-time mapping of the rock mass vibration energy release and the rectification pressure field, providing boundary excitation conditions for the multiphase mass transfer model; and the solution of the abnormal diffusion flux embeds the stress-sensitive diffusion characteristics into the component transport equation, correcting the limitations of the traditional Fick's law under non-steady-state geological disturbance. In specific implementation, first, the real-time data such as rock mass displacement and rectification pressure gradient are obtained through the multi-source dynamic parameter monitoring distribution architecture as shown in Figure 7 , and after preprocessing, they are input into the geological structure stress field reconstruction module to calculate the strain energy release rate and the stress-sensitive diffusion coefficient; then the coefficient is substituted into the multiphase mass transfer coupling model, combined with the temperature gradient distribution of the main heat exchanger and the component concentration change amount of the tower kettle, and the three-dimensional gas-liquid two-phase flow control equation is solved by the finite element method, and finally the non-steady-state working condition characteristic tensor containing the formation dislocation factor is output.

[0082] The benefit of this embodiment is that by constructing cross-field correlation functions based on the crustal strain energy release principle, the geological vibration energy attenuation mechanism is dynamically coupled with the rectification pressure field, and a stress-sensitive diffusion coefficient is introduced to correct the component transport model, forming a bidirectional mapping mechanism between geological disturbance and process parameters. This structured modeling method effectively captures the abnormal influence of rock mass stress fluctuation on the rectification mass transfer process under variable load conditions, providing a theoretical support and implementation path for the accurate encoding of subsequent working condition characteristic tensors and the precise generation of control instructions.

[0083] Embodiment Three: In order to solve the dynamic response lag problem caused by the mismatch between control parameters and geological disturbance modes in the variable load process of cryogenic nitrogen production, this embodiment further refines the parameter coupling mechanism and online optimization logic of the variable order control law for the engineering implementation of bifurcation control theory.

[0084] Specifically, the expression of the variable order control law is constructed in the framework of fractional calculus theory, as shown in Figure 3 , which converts the elastic characteristics of geological structures into dynamic adjustment parameters of the control algorithm by introducing a fractional order ν related to the rock mass Poisson's ratio. The proportional coefficient K p and the differential time constant τ d are not fixed parameters, but are optimized online based on the fractional order dynamic adaptation model as shown in Figure 4 , according to the energy distribution of the tower vibration mode. Specifically, the vibration energy spectrum is obtained through the triaxial acceleration sensor installed at the tower base, and when the energy proportion in the 10-50Hz frequency band exceeds 40%, the parameter adaptation module is triggered, and the control order is adjusted according to the mapping relationship between the rock mass Poisson's ratio μ and ν, and the values of the proportional coefficient K p and the differential time constant τ d are simultaneously corrected.

[0085] Further, the correction factor a in the control law is related to the material damping characteristics of the rectifying tower. A nonlinear relationship between the tower body natural frequency and a is established through finite element simulation. When the deviation between the measured vibration frequency and the simulated natural frequency exceeds 5%, the a compensation value under the corresponding working condition is automatically retrieved. The fractional derivative term is discretized within a 1ms control period, with a calculation step of 0.1ms to ensure that high-frequency component changes are captured during the variable load transient process. Figure 2 As shown in the phase space diagram of the fractional order control system, the component concentration oscillation term c k , the diffusion flux J k and the geological strain energy e v constitute a three-dimensional phase space. The variable order control law changes the convergence characteristics of the system trajectory by adjusting v. When the rock mass Poisson's ratio mu increases (i.e., the rock mass brittleness increases), v is increased to 0.9, corresponding to the "fracture zone" working condition curve in Figure 3 , which enhances the inhibition of high-frequency deviations by the differential element; conversely, when mu decreases (rock mass plasticity increases), v is reduced to 0.3, corresponding to the "hard rock mass" working condition curve in Figure 3 , which weakens the differential action to avoid excessive fluctuations in the control quantity.

[0086] It should be understood that the online optimization process of the control parameters K p and tau d is deeply coupled with the multi-source parameter monitoring data in Figure 7 . First, the temperature gradient distribution at the outlet of the main heat exchanger is used to judge the current cold distribution state. If the standard deviation of the temperature gradient exceeds 0.5K / m, adaptive correction of K p is triggered; then, tau d is adjusted in combination with the gas phase turbulent intensity coefficient at the outlet of the expander. When the turbulent intensity is greater than 0.6, tau d is dynamically reduced by 20% to improve the response speed. This dual driving mechanism based on geological parameters and process parameters enables the control law to adapt in real time to the coupled changes in rock mass stress state and rectification working conditions, avoiding the phase lag problem of traditional integer order control under complex disturbances.

[0087] The benefits of this embodiment are that by embedding the rock mass Poisson's ratio into the dynamic adjustment logic of the fractional order, a direct mapping between geomechanical parameters and control algorithm parameters is established, enabling the control law to autonomously optimize the proportional and differential parameters based on the tower vibration modal energy distribution. This variable order control method, which integrates bifurcation control theory, not only retains the ability of fractional calculus to describe non-integer order dynamic systems, but also realizes the working condition adaptation of control parameters through deep integration of multi-source monitoring data, providing a systematic control solution for deep cold nitrogen production devices to suppress the influence of geological disturbances and maintain the stability of the rectification process during the variable load process.

[0088] Embodiment four: In order to solve the problem of rectification stability caused by fluid dynamic disturbance of the deep cooling nitrogen production system under variable load conditions, the anti-disturbance execution mechanism of the cooperative controller is further optimized. Specifically, a high-resolution piezoelectric ceramic array sensor is integrated on the flow channel wall of the liquid oxygen reflux pump outlet flange. The sensor array is arranged in an equidistant ring shape, which can capture the spatiotemporal distribution characteristics of fluid micro-vortex energy spectrum density in real time. It should be understood that an abnormal increase in micro-vortex energy spectrum density usually indicates an intensification of vortex shedding. When the sensor detects that the vortex shedding frequency and the rock mass base frequency form a resonant coupling, the system immediately triggers a dual-mode anti-disturbance mechanism.

[0089] Further, the first mode of the dual-mode mechanism is the dynamic adjustment of the reflux pump driving power. The power is increased to a preset threshold value through a servo motor controller, which is calculated in advance based on a fluid dynamics model to ensure that the increment of driving power is sufficient to suppress the resonance effect and the maintenance time is an integer number of oscillation periods to avoid control failure caused by phase interference. At the same time, the second mode synchronously activates the buffer injection system of the condenser evaporation cavity. The system converts liquid methane after two-stage pre-cooling into sub-micron aerosol through an ultrasonic atomizing nozzle. The spray particle size distribution is calibrated in real time by a laser particle size analyzer to ensure that a uniform anti-phase change protection layer is formed on the surface of the condenser tube bundle. It should be noted that the injection flow rate of the buffer is real-time feedback controlled by the formation longitudinal wave velocity, and the two are negatively correlated. The higher the wave velocity, the greater the rock stiffness, and the required buffer dose is correspondingly reduced. This dynamic matching mechanism is realized through an embedded fuzzy logic controller, which can effectively offset the influence of formation vibration on condensation efficiency.

[0090] In terms of hardware configuration, the ultrasonic atomizing nozzle adopts an anti-clogging structure design. The internal flow channel is polished to reduce the surface energy, and a pre-positioned impurity filtering device is used to ensure the reliability of long-term operation. The signal acquisition frequency of the sensor array is dynamically matched with the fluid Reynolds number. The main frequency component of the energy spectrum density is analyzed in real time through Fourier transform to avoid false triggering caused by noise interference. It can be understood that this anti-disturbance mechanism forms a two-way suppression of vortex shedding vibration through the synergistic effect of power regulation and phase change buffering, which not only improves the stability of the mechanical structure, but also maintains the high-efficiency heat transfer performance of the condenser, thereby establishing a dynamic balance system of multi-physical field coupling under variable load conditions.

[0091] The benefit of this embodiment is that through the integrated design of sensor array and dual-mode control, accurate identification and layered control of fluid dynamic disturbance are realized, avoiding the limitations of single adjustment method, and providing a systematic solution for stable operation of the deep cooling nitrogen production system under complex geological conditions.

[0092] Embodiment five: To solve the problem of the influence of geological environment changes on the monitoring accuracy of the deep cold nitrogen production equipment, the structure design and signal processing mechanism of the geological monitoring module are further refined. Specifically, the borehole strain gauge array is arranged according to the Mohr-Coulomb failure criterion, the drilling depth penetrates the bottom boundary of the weathered layer and enters the intact bedrock stratum, ensuring that the sensor can capture the true stress state of the rock mass. Each borehole strain gauge adopts a triaxial structure to synchronously measure the linear strain in three orthogonal directions, and the layout spacing is optimized according to the elastic modulus gradient of the stratum to form a three-dimensional monitoring network of the tectonic stress field.

[0093] Further, the microseismic wave pickup of the three-component geophone group is equipped with a constant temperature compensation type signal amplifier, which uses a semiconductor refrigerating plate to maintain the working temperature stable and eliminate the influence of environmental temperature fluctuations on the sensitivity of the sensor. The sampling frequency of the signal acquisition system is dynamically adjusted through the background noise cross-correlation function, covering the rock mass failure characteristic frequency band, ensuring effective capture of microseismic signals. In the signal processing link, the Hilbert-Huang transform is used for time-frequency analysis of the original signal, and the industrial vibration interference components are stripped through empirical mode decomposition, including the following steps: first, the signal is decomposed into intrinsic mode functions (IMF) through EMD, then the time-frequency distribution is constructed through Hilbert transform, and finally the interference modes are identified and removed based on the energy threshold.

[0094] It should be understood that the application of the Mohr-Coulomb criterion ensures the scientificity of the strain gauge array layout, which can sensitively reflect the stress state of the potential failure surface of the rock mass; the constant temperature compensation technology solves the measurement error problem caused by the temperature drift of traditional sensors, improving the reliability of long-term monitoring data; the combination of dynamic sampling frequency and Hilbert-Huang transform processing effectively separates the effective signal in complex background, providing a high-precision data basis for subsequent stress field reconstruction. In addition, the monitoring module also integrates an automatic calibration function, which regularly calibrates the sensor sensitivity through a standard signal source to ensure the long-term stability of the entire monitoring system.

[0095] The benefit of this embodiment is that through targeted sensor arrangement, hardware optimization and signal processing technology, a high-reliability geological monitoring system is constructed, which can real-time and accurately obtain the stress state of the rock mass and microseismic activity information, providing key environmental parameter support for the adaptive control of the deep cold nitrogen production system, effectively improving the prediction ability of the system to geological disturbances.

[0096] Embodiment six: To solve the problem of heat transfer efficiency reduction caused by condenser fin frosting, this embodiment further optimizes the cold end control strategy in load mutation component decoupling, and establishes a dynamic control model based on fractal theory. Specifically, first, a fractal control model of the finned tube bundle temperature difference field is constructed. The model takes the Hausdorff dimension change rate as the key indicator of the evolution of the frosting phase interface, and calculates the spatial distribution and growth path of the frost thickness through numerical iteration, as shown in Figure 9 The input parameters of the model include the inlet and outlet temperatures of the cold end, the fluid flow rate and humidity, and the output is the frost rate distribution of each fin unit, thereby realizing the fine description of the frosting process.

[0097] When the fractal dimension monitoring value deviates from the reference value by more than the critical amplitude, the system automatically starts the liquid nitrogen backwashing defrosting program. The specific process is as follows: the liquid nitrogen at the bottom of the tower is pressurized to a set pressure by a plunger pump, and is injected into the refrigerant distribution main pipe in the reverse direction through the pipeline distribution system, while the electromagnetic valve is turned off to cut off the supply of refrigerant to the corresponding branch, forming a local cold source-free environment to accelerate the melting of the frost layer. The duration of defrosting is dynamically determined by the integral of the deviation of the fractal dimension, that is, the required energy input is calculated by accumulating the deviation value, to ensure that the defrosting process is sufficient and avoids waste of cold energy caused by excessive flushing. It should be understood that the backwashing pipeline is designed with insulation to reduce the gasification loss of liquid nitrogen during transportation, and a pressure sensor is installed at the end of the pipeline to monitor whether the backwashing flow meets the preset parameters in real time.

[0098] In terms of hardware design, the refrigerant distributor uses an electric regulating valve group, and the opening sequence is controlled by the fractal model output signal, which can realize precise distribution of cold energy to each fin unit. The calculation of the evolution path of the frosting phase interface is processed by a field programmable logic controller (FPGA) in parallel, ensuring real-time requirements. It can be understood that the introduction of fractal theory provides a multi-scale description method for the frosting process, which can more accurately capture the nonlinear characteristics of frost growth compared to traditional models, and the defrosting trigger mechanism based on fractal dimension avoids the blindness of timed defrosting, achieving a balance between energy saving and efficiency.

[0099] The benefit of this embodiment is that by combining the fractal control model with the dynamic defrosting strategy, an intelligent regulation system for the cold end of the condenser is established, effectively suppressing the impact of frosting on heat transfer performance and improving the adaptive ability of the system under variable load conditions, providing key thermal management technical support for the stable operation of the deep cold nitrogen production process.

[0100] Embodiment seven: To solve the stress corrosion problem of the oxygen-enriched liquid air throttle valve in a wet oxygen environment, this embodiment further optimizes the valve surface protection and maintenance process. Specifically, an amorphous Ni-W-P ternary alloy coating is deposited on the valve core sealing surface using magnetron sputtering technology, as shown in Figure 5The coating thickness is set according to the hydrogen sulfide concentration in the fluid before the valve: the higher the concentration, the thicker the coating, forming a differential protection against corrosive media. During the coating process, the element distribution is monitored in real time by a glow discharge spectrometer to ensure that the coating composition uniformity meets the design requirements.

[0101] During the valve action cycle, the control system dynamically calculates and applies a reverse polarization voltage based on the Nernst equation according to the oxygen content sensor feedback value to suppress the electrochemical corrosion process. It should be understood that the application of the Nernst equation takes into account the relationship between ion concentration in the solution and electrode potential, and by adjusting the voltage amplitude in real time, the valve core surface is always in a thermodynamically stable potential interval, thereby effectively reducing the corrosion current density. The voltage application system is equipped with an isolation transformer to avoid the influence of strong electric interference on the control signal, and an overvoltage protection device is also set to ensure the safety of the equipment.

[0102] After each operating cycle, the system activates an in-situ polishing program, which drives a diamond grinding head along the valve seat cone surface in a spiral trajectory through a servo motor, and the cutting depth is controlled by a residual stress probe in a closed loop. The probe uses a strain gauge principle to monitor the stress changes in the valve seat material during polishing in real time, avoiding excessive cutting that can lead to a decrease in sealing performance. The polishing trajectory is pre-planned by computer-aided design to ensure that it covers the entire sealing surface area, and the grinding head pressure is precisely controlled by a pneumatic servo system to achieve uniform removal of the worn layer.

[0103] It should be noted that the amorphous structure of the amorphous coating itself has excellent corrosion resistance, and the gradient setting is optimized for the uneven distribution of corrosive media; the dynamic adjustment of the reverse polarization voltage realizes the intelligentization of electrochemical protection; the in-situ polishing program removes the surface fatigue layer through mechanical processing, forming a multiple protection system for the valve sealing surface. In addition, the entire protection system is linked with the valve state monitoring module, which can adaptively adjust the maintenance cycle according to factors such as the number of on-off times and fluid medium parameters, improving the reliability and service life of the equipment.

[0104] The benefits of this embodiment are that through the synergistic effect of surface coating, electrochemical protection and mechanical maintenance, a comprehensive stress corrosion protection system for oxygen-rich wet environments is established, effectively solving the sealing failure problem of throttling valves in severe working conditions, and ensuring the long-term stable operation of key components of deep cold nitrogen production systems.

[0105] Embodiment Eight: To solve the problem of balancing the defrosting efficiency and energy consumption of the heat exchanger under complex geological conditions, this embodiment further optimizes the defrosting control strategy and realizes intelligent regulation and control of the defrosting process combined with geological monitoring data. Specifically, the system collects rock layer water content data in real time through the ground resistivity probe, establishes a dynamic mapping relationship between the fin frosting rate and the thermal conductivity, and this relationship model is trained based on historical operation data through a machine learning algorithm, which can accurately reflect the change law of frost layer thermal resistance under different water content conditions.

[0106] Further, the optimal working time window of each defrosting interval is predicted using a finite element heat-flow coupling model, which takes into account the fin structure, heating band power distribution, and environmental temperature. The start and stop timing of each heating partition is determined through numerical simulation to avoid thermal interference between adjacent regions. The electric heating band uses nanometer carbon fiber material, which has the characteristics of fast response speed and uniform heating. Its control circuit realizes high-frequency switching through solid-state relays to ensure precise control of the time window.

[0107] When the three-component detector group detects fault activation signals, the system triggers the full-area pulse strong heat defrosting mode in advance: through the variable frequency power supply, the heating power is instantaneously increased to 120% of the rated value, and maintained for a set period of time to quickly remove the potential thick frost layer. After defrosting is completed, the system automatically resets the defrosting cycle counter to avoid the impact of frequent triggering on equipment life. It should be understood that the ground resistivity data reflects the change of the water content of the stratum, and the water content is closely related to the air humidity, thereby indirectly affecting the frosting rate. This cross-physical-field parameter correlation provides a more comprehensive basis for defrosting control; the design of the pulse strong heat mode is aimed at sudden geological disturbances to ensure the safe operation of the heat exchanger under extreme working conditions.

[0108] In terms of system integration, the defrosting control module and the geological monitoring module communicate in real time, sharing data such as fault activity and rock layer water content, forming a closed-loop control system. The heating band partitions are divided according to the flow channel layout of the heat exchanger, and each partition is equipped with an independent temperature sensor that provides real-time feedback on heating effectiveness and corrects model prediction errors.

[0109] The benefits of this embodiment are that through the coupling analysis of geological parameters and thermal parameters, an intelligent defrosting control model is constructed, dynamic optimization and emergency response of the defrosting process are realized, and the reliability of the heat exchanger under complex environments is effectively improved, providing a solid thermal management guarantee for the stable operation of the deep cold nitrogen production system.

[0110] Embodiment Nine: To solve the problem of shear strain rate exceeding the standard caused by geological structure activity impacting the deep cold nitrogen production system, this embodiment further perfects the three-level interlock response mechanism to ensure system safety under extreme working conditions. Specifically, when the three-component detector group detects that the shear strain rate exceeds the preset threshold, the control system activates the three-level response measures in priority order:

[0111] The first level response is the start of the emergency pressure relief module of the rectifying tower, which quickly reduces the pressure in the tower to a safe range by sequentially opening the overhead safety valve and the bottom liquid drain valve. The pressure relief rate is accurately constrained by a PID controller to avoid liquid flooding caused by sudden pressure drop. The controller parameters are pre-set according to the tower structure size and material properties to ensure the stability of the pressure relief process. The safety valve uses a pilot structure and is equipped with a pressure sensor to provide real-time feedback of the valve pressure. The liquid drain valve is interlocked with the liquid accumulation tank to prevent secondary hazards caused by liquid splashing.

[0112] The second level response is the switching of the standby expander unit, which quickly disconnects the output shaft of the faulty unit through a hydraulic clutch and starts the magnetic suspension bearing standby unit. The magnetic suspension bearing system has the characteristics of non-contact and low power consumption. During the start-up process, a soft starter is used to control the gradual increase of the speed to the rated value, avoiding the impact of current surge on the power grid. During the switching process of the unit, the flow distribution system automatically adjusts the guide vane angle to ensure the continuity of the cold supply and reduce the disturbance of load fluctuation on the rectification process.

[0113] The third level response is the injection of a working condition stabilizer, which sprays critical state carbon dioxide powder into the cold end of the main heat exchanger. The injection amount is logarithmically related to the fault slip distance, which is determined by a geomechanical model to ensure that the stabilizer dosage matches the disturbance intensity. Critical state carbon dioxide has high diffusivity and rapid heat absorption characteristics, which can quickly suppress abnormal phase change in the heat exchanger and stabilize the heat transfer conditions.

[0114] It should be understood that the three-level interlock response mechanism forms a multi-level protection against geological impact through the coordinated action of pressure control, power switching, and working condition stabilization. Each response link is equipped with independent sensors and actuators, and is coordinated by a redundant controller to avoid protection failure caused by single component failure.

[0115] The benefits of this embodiment are that through the systematic three-level interlock design, a rapid response system is established to deal with sudden geological disturbances, which can quickly cut off the risk source, switch to standby equipment, and stabilize the process parameters when the shear strain rate exceeds the standard, minimizing the damage of extreme conditions to the deep cold nitrogen production system and ensuring the safety of the equipment and the continuity of production.

[0116] Example Ten: To solve the problem of parameter self-adaptation of fractional order control system under complex geological conditions, this embodiment further clarifies the dynamic determination method of fractional order, improving the environmental adaptability of the control strategy. Specifically, first, a non-contact laser Doppler vibration meter is used to scan the tower foundation rock surface to collect Rayleigh wave group velocity dispersion curves. The measurement accuracy of this instrument reaches microns, which can capture the subtle vibration characteristics of the rock surface. The dispersion curve contains wave velocity information of different frequency components, reflecting the distribution of elastic properties of the shallow rock mass.

[0117] Further, based on the anisotropic medium elastic wave equation, the collected dispersion data is inverted to calculate the instantaneous gradient value of the shear modulus of the rock mass. The inversion process adopts a regularization algorithm, and combines prior geological data to construct a constraint condition to ensure the uniqueness and accuracy of the inversion result. The change of the shear modulus gradient value reflects the spatial inhomogeneity of the stiffness of the rock mass, and is the key basis for determining the fractional order controller parameters.

[0118] Further, the shear modulus change rate is substituted into the preset nonlinear transfer function, which is established by system identification method, and can map the change of physical parameters to the update instruction of the fractional order ν. The input of the transfer function is the shear modulus gradient ΔG(t), and the output is the adjustment step Δν of ν. The mathematical form considers the matching of the rock mass dynamics and the control algorithm, and ensures the smoothness and effectiveness of the order adjustment.

[0119] Further, when the laser vibration detector detects the P-wave first arrival, the system immediately freezes the order update process and switches to the preset anti-shock mode parameter group. The anti-shock mode parameter group is determined in advance through seismic working condition simulation, and contains a higher proportional coefficient and an adaptive fractional order, which can enhance the damping characteristics of the system and suppress the structure vibration caused by ground motion. It should be understood that the detection of the P-wave first arrival is realized by a real-time signal processing algorithm, which uses the time difference between P-wave and S-wave to give an early warning and gain a time window for switching control parameters.

[0120] The benefit of the embodiment is that through the integrated design of geological vibration measurement, elastic parameter inversion, and control parameter linkage, a dynamic adaptation system of fractional order is constructed, so that the control system can match the dynamics under complex geological conditions in real time.

[0121] Embodiment eleven: In order to solve the adaptive rectification control problem of the deep cryogenic nitrogen production system under variable load conditions, the hardware architecture of the control system and the coordination mechanism of each functional unit are further refined. The geological disturbance sensing unit of the control system collects the formation stress change signal in real time through the drill hole strain gauge array embedded in the stable rock layer of the tower foundation, and synchronously obtains the dynamic parameters such as vibration acceleration and displacement through the four-component microseismic monitor installed on the tower body pile cap, forming a multi-source heterogeneous monitoring data chain. It should be understood that the drill hole strain gauge array adopts a distributed networking method, and its layout density and depth are determined according to the tower foundation geological exploration report to ensure coverage of the potential geological disturbance area; the four-component microseismic monitor integrates a three-axis acceleration sensor and a one-way displacement sensor, which can capture structure vibration signals of different modes.

[0122] Further, the multi-field coupling processing engine takes the FPGA chip set as the core carrier, and its embedded phase change coupling algorithm hardware acceleration core realizes real-time coding of multi-source dynamic parameters through customized logic circuits. The processing engine first synchronously samples the analog signals output by the borehole strain gauge and the microseismic monitor, inputs the signals into the feature extraction module after analog-digital conversion, maps the non-steady state working condition characteristics into high-dimensional feature tensors by constructing a tensor space containing formation dislocation factors, vibration amplitude phase difference and other parameters. It can be understood that the hardware acceleration core adopts a pipeline architecture design and supports multi-channel data parallel processing, ensuring real-time requirements under variable load conditions.

[0123] The load decoupling execution unit receives the feature tensors from the multi-field coupling processing engine, divides the signal into frequency bands through a set of tunable band-pass filter circuits, and realizes fine analysis in the time-frequency domain through a 16-channel wavelet packet decomposition module. The unit separates the load mutation components from the process fluctuation components in the feature tensors through an adaptive threshold determination algorithm, and outputs the corresponding digital signals to the bifurcation control hub. It should be noted that the center frequency and bandwidth of the tunable band-pass filter can be dynamically adjusted according to the real-time working condition, and the wavelet packet decomposition module uses an orthogonal basis function set to ensure the completeness and reconstruction accuracy of signal decomposition.

[0124] The bifurcation control hub integrates an analog circuit implementation module of the variable order controller, constructs a continuously adjustable controller transfer function through a voltage-controlled resistor network and an operational amplifier array, cooperates with the perturbation signal generated by the chaotic signal generator, and realizes tracking and locking of the dynamic equilibrium point of the rectification system. The safety interlocking logic unit monitors the controller output and the process parameter threshold in real time, and triggers the predefined safety strategy when abnormal fluctuations are detected. It can be understood that the order adjustment range of the variable order controller covers 0.1 to 2.0, and the hardware implementation of the analog circuit avoids the quantization error in digital control and improves the continuity of system response.

[0125] The two-phase flow coordination terminal obtains the gas-liquid interface fluctuation signal through a piezoelectric ceramic array signal demodulator, drives the proportional electromagnetic valve group of the liquid methane injection system to adjust the reflux ratio, and simultaneously constructs a gas-liquid two-phase flow coordination control model based on a multivariable PID controller card. It should be understood that the piezoelectric ceramic array is uniformly arranged in the middle section of the rectification tower to sense the subtle changes in liquid phase distribution; the proportional electromagnetic valve group adopts a pilot type structure, and the response time is controlled within 50 ms, ensuring rapid tracking of load changes. Through the synergistic effect of each unit, the control system realizes full-link closed-loop control from geological disturbance sensing to rectification process rebalancing, providing hardware support for stable operation under variable load conditions.

[0126] Embodiment twelve: To solve the reliability problem of the geological disturbance sensing unit in the deep cold environment and complex geological conditions, the hardware protection structure and signal transmission scheme of the geological disturbance sensing unit are further optimized. The detector probe of the geological disturbance sensing unit is packaged in a Hastelloy C-276 forged shell with a wall thickness of not less than 8 mm, which is formed by integral forging process, effectively resisting mechanical impact and corrosion medium erosion caused by formation stress. The shell is coated with a 3-5 μm thick diamond-like carbon coating on the outside, which forms a dense protective layer through physical vapor deposition process, significantly improving the surface hardness and chemical inertness, and adapting to corrosive environments containing hydrogen sulfide, moisture, etc.

[0127] Further, the signal transmission system uses a double-layer shielded coaxial cable, the inner layer is a silver-plated copper wire braid, which provides high-efficiency electromagnetic interference shielding capability, and the outer layer is covered with a fluororubber low-temperature resistant sheath, ensuring flexibility and insulation performance in the temperature range of -196°C to 85°C. The cable joint is filled with high-thermal-conductivity silicone grease to fill the gap, and the outside is sleeved with an armored heat-shrinkable tube to form a sealed structure. The connection assembly treated in this way meets the IP68 protection level requirement and can withstand 10 meters of water immersion and reliable operation in a dusty environment. It should be understood that the fluororubber sheath adds nano-sized anti-aging fillers, which improves the ultraviolet and ozone resistance through formula optimization, prolonging the service life of the cable.

[0128] Further, the synergistic design of the corrosion-resistant packaging structure and the shielding transmission scheme effectively solves the problem of interference and device damage of monitoring signals caused by environmental factors such as low temperature, high humidity, and strong corrosion in the deep cold nitrogen production scene. During signal transmission, the double-layer shielding structure suppresses common-mode interference to more than 60 dB, ensuring that the raw signals of the borehole strain gauge and the microseismic monitor are transmitted to the processing unit with distortion less than 0.1%, providing a reliable data basis for subsequent multi-field coupling analysis. Through multiple means of material selection, structure design, and process optimization, this embodiment constructs a high-reliability geological disturbance sensing front end, ensuring the stable operation of the control system in harsh environments.

[0129] Embodiment thirteen: To solve the communication reliability problem of the load decoupling execution unit in the high real-time control scene, this embodiment further designs a redundant communication architecture and an exception handling mechanism. The load decoupling execution unit is configured with a main and backup dual communication channel. The main channel uses a fiber optic reflective memory ring network to realize nanosecond-level data synchronization using the high-speed transmission characteristics of optical signals. The ring network nodes are connected through a star coupler, supporting plug-and-play and fault isolation of any node. The backup channel is a PROFIBUS-DP bus, which uses a shielded twisted pair as the transmission medium to provide RS-485 differential signal transmission, ensuring basic communication functions when the main channel fails.

[0130] Further, the system sets up a double abnormality detection mechanism, integrates a CRC-32 check module at the output end of the FPGA, performs cyclic redundancy check on each data frame, constructs a waveform model of the load mutation component through a time series prediction algorithm, and calculates a distortion index of the current signal in real time. When three consecutive data checks fail or the distortion index exceeds a preset threshold, the communication management module automatically triggers the channel switching logic, disconnects the faulty channel and accesses the standby channel, and sends a restart instruction to the wavelet packet decomposition module.

[0131] Further, the combination of the redundant communication architecture and the abnormality processing mechanism effectively improves the fault tolerance of the load decoupling process. It can be understood that the communication delay of the fiber reflection memory ring network is controlled within 200 ns, meeting the real-time transmission requirements of high-frequency signals; the PROFIBUS-DP bus adopts a transmission rate of 12 Mbps, which can carry the feature data after wavelet packet decomposition. The abnormality detection algorithm dynamically adjusts the threshold to adapt to changes in signal characteristics under different working conditions, ensuring that the false positive rate is less than 0.01%. Through the collaborative design of hardware redundancy and software fault tolerance, this embodiment constructs a high-reliability signal transmission and processing link, providing communication support for accurate decoupling of the load mutation component.

[0132] Embodiment fourteen: To solve the problem of heat dissipation of the bifurcation control center in the operation of high-density integrated circuits, the thermal management system architecture of the bifurcation control center is further optimized. The FPGA chip heat dissipation substrate of the bifurcation control center is embedded with a micro-channel copper cold plate. The cold plate uses a precision etching process to form a parallel flow channel network with a diameter of 0.5 mm, and the flow channel density is 20 channels / cm, which is connected to the chip substrate through a brazing process with high thermal conductivity. The cold plate is connected to a secondary cooling circuit, and perfluoropolyether is used as the cooling medium. This medium has low viscosity, high boiling point and electrical insulation properties, which is suitable for the heat dissipation needs of high-density electronic devices.

[0133] Further, the cooling circuit is driven by a magnetic pump, and the medium exchanges heat with the expander exhaust when flowing through the finned heat exchanger, realizing pre-cooling using the low-temperature waste heat resources of the cryogenic system. The control system monitors the junction temperature in real time through the thermocouple integrated in the chip substrate, dynamically adjusts the cooling liquid flow based on the fuzzy control algorithm, and automatically triggers the frequency reduction mechanism to reduce the chip power consumption when the junction temperature approaches the critical value, while starting the auxiliary semiconductor cooling module to further reduce the substrate temperature through the Peltier effect.

[0134] Further, the design of the thermal management system effectively solves the problem of heat accumulation in the FPGA chip during high-speed operation. The finned heat exchanger uses the -150℃ low-temperature gas discharged by the expander to maintain the inlet temperature of the cooling liquid below 40℃, reducing the load of the auxiliary cooling system. This embodiment combines passive heat dissipation and active cooling to build an efficient thermal management solution suitable for cryogenic environments, ensuring the long-term safe operation of the bifurcation control center.

[0135] Example 15: To solve the problem of safety control of the two-phase flow coordination terminal in a high-pressure cryogenic environment, the safety interlocking mechanism and signal transmission design are further improved. The two-phase flow coordination terminal is provided with double cut-off valves at the outlet of the liquid methane storage tank, which adopts a series structure. The first valve is a pneumatic ball valve, and the secondary valve is an electric gate valve, which has a manual emergency operation function. The burst pressures of the two-stage safety valves are set to 1.1 times and 1.3 times of the design pressure, respectively, forming a staged discharge protection mechanism.

[0136] Further, the reflux pump frequency converter is configured with a current harmonic filter, which uses an LC filter circuit to suppress high-order harmonics, and integrates a regenerative braking unit to feed the braking energy back to the DC bus, avoiding capacitor overvoltage failure. The position feedback of the actuator uses an absolute photoelectric encoder with a resolution of 16 bits. The signal is transmitted through double CAN bus branches, each bus is independently powered and configured with a signal isolator to ensure the integrity of the feedback data.

[0137] Further, the design of the safety interlocking mechanism comprehensively improves the intrinsic safety of the two-phase flow system. It should be understood that the redundant configuration of the double cut-off valves meets the SIL 2 safety level requirement, and regular valve group sealing tests ensure that the failure probability is less than 10 -6 / h; the double CAN bus architecture adopts a hot standby redundancy mode, and the master and slave buses are real-time synchronized. When a single bus fails, it automatically switches over, with a switching time of less than 50μs. The nitrogen-filled positive pressure explosion-proof box is equipped with a dew point monitor. When the dew point of the gas in the box is higher than -40℃, the drying process is automatically started to avoid damage to electrical components by condensed water.

[0138] Example 16: To solve the problem of integrated control of the deep cold nitrogen making device under complex working conditions, the overall architecture of the control device is further optimized, integrating an anti-displacement tower foundation structure, a shock-resistant electrical cabinet component, and an online mass spectrometry monitoring module. The anti-displacement tower foundation structure adopts a tower body support system with a universal hinge. The axial bellows compensator can absorb vertical displacement of ±50mm. The shear limiting base is buffered by a damping rubber pad to resist horizontal vibration. The natural frequency of the support system is designed to be 2.5Hz, avoiding the main energy frequency band (5-15Hz) of geological disturbance and reducing the impact of vibration transmission on the rectifying tower.

[0139] Further, the anti-vibration electrical cabinet assembly is a partitioned explosion-proof control cabinet, which is internally configured with a SIL 3 level safety instrument system, including an independent emergency shutdown loop and a process interlock module. The positive pressure ventilation system in the cabinet adopts a centrifugal fan to maintain a positive pressure of 50-100 Pa, and a dew point monitor and an automatic drain valve are arranged at the air inlet to start a heating and dehumidification process when the detected air dew point is higher than -20℃, so as to avoid electrical failure caused by condensation. The on-line mass spectrometry monitoring module is installed with a quadrupole mass spectrometer at the top of the nitrogen gas pipeline, the ion source chamber is maintained at a super-high vacuum of 10 Pa by a magnetic suspension molecular pump, and the detector output signal is connected to a core processor after being filtered by an EMI filter to filter high-frequency noise, so as to realize real-time monitoring of the nitrogen purity. -6 Pa super-high vacuum, and the detector output signal is connected to a core processor after being filtered by an EMI filter to filter high-frequency noise, so as to realize real-time monitoring of the nitrogen purity.

[0140] This embodiment cooperatively designs mechanical structure, electrical safety and on-line monitoring to build an integrated control device with anti-vibration, explosion-proof and high-precision detection capabilities, and provides a solution for stable operation of the cryogenic nitrogen production under variable load conditions.

[0141] Embodiment Seventeen: To solve the adaptability problem of the cryogenic nitrogen production control method in the geological disturbance scene, this embodiment further refines the control logic of the executable program in the computer readable storage medium. Based on the implementation of the adaptive rectification control method, the program code adds a geological vibration response module: when it is detected that the output frequency spectrum of the rectification tower foundation accelerometer falls into the rock mass characteristic frequency band, the three-level control protocol is activated, first the fractional order controller parameters are reconstructed by the fuzzy logic algorithm, and the integral order is dynamically adjusted from 0.8 to 1.2, which improves the damping characteristics of the system to low-frequency vibration; then the dual-mode anti-disturbance mechanism is called to switch between the conventional PID control and the sliding mode control, and the corresponding control mode is selected according to the vibration amplitude to suppress the influence of disturbance on the rectification process.

[0142] Further, the program code triggers the tower foundation hydraulic damper to the butt mode, and the relative displacement phase difference between the tower body and the foundation is obtained in real time by a laser range finder, and the reciprocating motion amplitude and frequency of the damper piston are controlled based on the phase difference calculation result to form a dynamic balance force opposite to the foundation vibration. It should be understood that the laser range finder adopts a phase measurement principle, with an accuracy of ±0.1 mm and a sampling frequency of 1000 Hz, which ensures real-time capture of vibration phase changes; the servo valve of the hydraulic damper has a response frequency of 50 Hz, which can track high-frequency vibration signals, and the piston stroke is controlled within a range of ±50 mm to avoid structural stress concentration caused by excessive compensation.

[0143] Further, the design of the control logic enhances the system's active response capability to geological disturbances. It can be understood that the three-level control protocol realizes adaptive adjustment of control parameters by matching vibration characteristics through a pre-defined working condition database; the dual-mode anti-disturbance mechanism improves response speed while ensuring system stability by switching control strategies. The adaptive filtering technology is adopted in the inverse control algorithm of the hydraulic damper to estimate the vibration amplitude and phase in real time, ensuring the precise application of the balancing force. The present embodiment builds a control system with active compensation capability for geological disturbances through the deep synergy of software algorithms and hardware execution mechanisms, improves the operation reliability of the deep cryogenic nitrogen generation system under non-steady state working conditions, and realizes the full-process of the control method from data acquisition to strategy execution.

[0144] Although the present application has been described above with reference to the preferred embodiments, it is to be understood that the application is not limited to the above-described embodiments, and various modifications and changes can be made by those skilled in the art without departing from the spirit of the present application, and such modifications and changes shall fall within the scope of the appended claims and their equivalents.

Claims

1. A method for adaptive rectification control of deep cold nitrogen production variable load condition, characterized in that, The steps of the control method comprise: Real-time acquisition of multi-source dynamic parameters of the rectification system, including temperature gradient distribution matrix of each flow passage outlet of the main heat exchanger, concentration change amount of air-liquid components in the tower cauldron, turbulent intensity coefficient of gas phase at the outlet of the expander, and fluctuation rate of rock mass shear stress fed back by the geological monitoring module; Through a geological structure stress field reconstruction algorithm and a multiphase flow mass transfer coupling model, the multi-source dynamic parameters are coded into a non-steady state working condition characteristic tensor including a stratum dislocation factor; A space-time separation convolutional neural network is used to decouple the load mutation component of the working condition characteristic tensor, a tunable band-pass filter is used to extract the microseismic frequency spectrum characteristics of the rock stratum, and a wavelet packet analysis is used to strip the inherent fluctuation component of the process; Based on the bifurcation control theory, the rectification is rebalanced, the reboiler thermal siphon pressure difference safety margin is dynamically set according to the decoupled load mutation component, and the chaotic suppression signal of the condenser cold agent feed injection is generated combined with the nitrogen Raman spectrum shift at the top of the tower; A gas-liquid two-phase flow cooperative controller is constructed, the chaotic suppression signal is converted into the dynamic coupling instruction of the lower backflow ratio regulation gradient and the upper tower oxygen-rich liquid air throttling valve opening degree, and is used to maintain the liquid nitrogen purity calibration value of the rectification tower under the condition of rock mass stress fluctuation.

2. The deep-cooling nitrogen production adaptive rectification control method of claim 1, wherein, The geological structure stress field reconstruction algorithm and the multiphase flow mass transfer coupling model perform the following steps: Based on the principle of crustal strain energy release, a vibration-rectification mass transfer correlation function is established: wherein G(t) is a vibration-distillation mass transfer correlation function, used to quantify the coupling strength between the vibration energy release of the rock mass and the pressure field of the distillation column, Ω is the three-dimensional spatial range of the rock mass, is the displacement vector field of the rock mass, and β is an empirical damping coefficient, is the pressure gradient distribution field of the distillation column. The abnormal diffusion flux of gas phase components is solved by a non-equilibrium statistical mechanics equation: , where D ki eff The stress-sensitive tensor diffusion coefficient, Let i be the concentration gradient of component k along the i-th direction. For volumetric strain versus time t and spatial coordinate x j The second mixed derivative, γ k Let be the strain energy response coefficient of component k.

3. The deep-cooling nitrogen production adaptive rectification control method for varying load conditions according to claim 1, characterized in that, The bifurcation control theory executes a variable order control law: where Δu(t) represents the control increment of the controller output, K p is the proportional coefficient, τ d is the differential time constant, α is the correction factor, and v is the fractional order related to the Poisson ratio of the rock mass, which is used to control the parameters K p , τ d According to the online optimization of the tower body vibration modal energy distribution, e(t) is the load deviation signal.

4. The deep-cooling nitrogen production adaptive rectification control method of claim 1, wherein, The cooperative controller performs the following operations: a high-resolution piezoelectric ceramic array sensor is integrated at the outlet flange of the liquid oxygen reflux pump for real-time monitoring of fluid micro-vortex energy spectrum density; when the vortex shedding frequency exceeds the rock mass fundamental frequency, a dual-mode anti-disturbance mechanism is triggered, the reflux pump driving power is increased to a set threshold and maintained for an integer number of oscillation periods, and a two-stage pre-cooled liquid methane anti-phase change buffer is injected into the condenser evaporation cavity at the same time; the injection flow rate of the buffer is negatively related to the longitudinal wave velocity of the stratum, and a sub-micron aerosol is formed through an ultrasonic atomizing nozzle to ensure full coverage of the condenser tube bundle.

5. The deep-cooling nitrogen production adaptive rectification control method of claim 1, wherein, The geological monitoring module comprises: a borehole strain gauge array arranged according to the Mohr-Coulomb failure criterion, with a drilling depth exceeding the bottom boundary of the weathered layer; a microseismic wave pickup of a three-component geophone group, each geophone unit being equipped with a constant temperature compensation type signal amplifier; the sampling frequency is dynamically adjusted according to the background noise cross-correlation function, the adjustment range covers the rock mass rupture characteristic frequency band, and the industrial vibration interference component is stripped in real time through Hilbert-Huang transform.

6. The deep-cooling nitrogen production variable load condition adaptive rectification control method of claim 1, wherein, The load mutation component decoupling includes condenser cold end optimization, the steps comprising: establishing a fractal control model of the temperature difference field of the finned tube bundle, taking the Hausdorff dimension change rate as the index to iteratively calculate the evolution path of the frost phase interface, and dynamically adjusting the opening sequence of the refrigerant distributor; when the fractal dimension deviates from the reference value by more than the critical amplitude, a liquid nitrogen backwash defrosting program is started, the liquid nitrogen at the bottom of the tower is injected into the refrigerant distribution mother pipe in reverse after being boosted, and the refrigerant supply of the corresponding branch is cut off at the same time, the defrosting duration is positively related to the integral quantity of the deviation of the fractal dimension.

7. The deep-cooling nitrogen production adaptive rectification control method of claim 1, wherein, The control method further comprises implementing oxygen-rich liquid air throttle valve stress corrosion protection, the steps comprising: depositing an amorphous Ni-W-P ternary alloy coating on the valve core sealing surface, the coating thickness being set in stages according to the hydrogen sulfide concentration of the fluid before the valve; applying a reverse polarization voltage during the valve action cycle, the voltage amplitude being dynamically calculated according to the oxygen content sensor feedback value by the Nernst equation; activating an in-situ polishing program after each set period of operation, driving a diamond grinding head to perform spiral trajectory cutting along the valve seat cone surface, the cutting depth being controlled by a residual stress probe closed loop.

8. The deep-cooling nitrogen production adaptive rectification control method of claim 6, wherein, The control method further comprises heat exchanger defrosting optimization, the steps comprising: calculating the dynamic mapping relationship between the fin frosting rate and the thermal conductivity according to the rock moisture content data fed back by the ground resistivity probe; predicting the optimal working time window for each defrosting interval by a finite element thermal-flow coupling model, and controlling the partitioned start and stop of the electric heating belt; when a fault activation signal is detected, triggering the full-area pulse strong heat defrosting mode in advance: instantaneously increasing the power of the frequency conversion power supply to the rated value and maintaining for a set period of time, and resetting the defrosting cycle counter after defrosting is completed.

9. The deep-cooling nitrogen production variable load condition adaptive rectification control method according to claim 5, characterized in that, When the three-component detector group detects that the shear strain rate exceeds the threshold value, a three-level interlock response is activated, the steps comprising: starting the rectifying tower emergency depressurization module: sequentially opening the tower top safety valve and the bottom liquid discharge valve, the depressurization rate being constrained by a PID controller; switching to the standby expander set: disconnecting the output shaft of the faulty set through a hydraulic clutch, while starting the magnetic suspension bearing standby set and performing soft start; injecting a process stabilizer: spraying critical state carbon dioxide powder into the cold end of the main heat exchanger, the spraying amount being in logarithmic relationship with the fault slip distance.

10. The deep-cooling nitrogen production adaptive rectification control method for varying load conditions according to claim 3, wherein, The dynamic determination of the fractional order step comprises: scanning the tower foundation rock surface using a non-contact laser Doppler vibration meter to collect Rayleigh wave group velocity dispersion curves; inversing the instantaneous gradient value of the rock shear modulus based on the elastic wave equation of anisotropic medium; substituting the modulus change rate into a preset nonlinear transfer function to generate an update instruction of the fractional order step; when the initial arrival of the seismic P wave is detected, freezing the step update and switching to a preset anti-shock mode parameter group.

11. A deep-cooling nitrogen production variable load condition adaptive rectification control system for implementing the deep-cooling nitrogen production variable load condition adaptive rectification control method according to any one of claims 1-10, characterized in that, The control system comprises: a geological disturbance sensing unit, including a borehole strain gauge array embedded in the stable rock layer of the tower foundation and a four-component microseismic monitor installed on the tower body pile cap, for real-time acquisition of multi-source dynamic parameters of the rectifying system; a multi-field coupling processing engine equipped with an FPGA chip set and a hardware acceleration core embedded with a phase change coupling algorithm, for real-time generation of a working condition feature tensor, and for encoding the multi-source dynamic parameters into a non-steady working condition feature tensor including a stratum dislocation factor; a load decoupling execution unit including a tunable band-pass filter circuit group and a 16-channel wavelet packet decomposition module, outputting digital signals of load mutation components and process fluctuation components, for load mutation component decoupling of the working condition feature tensor; a bifurcation control center integrated with an analog circuit implementation module of a variable order controller, a chaotic signal generator and a safety interlock logic unit, for driving rectifying rebalancing. The two-phase flow coordination terminal comprises a piezoelectric ceramic array signal demodulator, a proportional electromagnetic valve group of a liquid methane injection system and a multivariable PID controller card, and is used for constructing a gas-liquid two-phase flow coordination controller.

12. The cryogenic nitrogen producing adaptive rectification control system of claim 11, wherein, The geological disturbance sensing unit comprises a corrosion-resistant packaging structure, a detector probe sealed in a Hastelloy C-276 forged shell, and a diamond-like carbon coating plated on the outside of the shell; signal transmission adopts a double-layer shielding coaxial cable, the inner layer is a silver-plated copper wire braid, and the outer layer is covered with a fluororubber low-temperature resistant sheath; the joint is filled with silicone grease and is sleeved with an armored heat-shrinkable tube to meet the IP68 protection level and the explosion-proof certification standard.

13. The cryogenic nitrogen producing adaptive rectification control system of claim 11, wherein, The load decoupling execution unit is equipped with a redundant communication architecture, the main signal channel adopts an optical fiber reflective memory ring network, and the standby channel is a PROFIBUS-DP bus; a double abnormality detection mechanism is arranged, a CRC-32 check module is added at the output end of the FPGA, and a time series prediction algorithm is used to identify the waveform distortion of the load mutation component; when three consecutive frames of data fail to pass the check or the waveform distortion index exceeds a threshold, the communication channel is automatically switched and the wavelet packet decomposition process is restarted, and during the restart period, historical data interpolation is used to maintain control continuity.

14. The cryogenic nitrogen producing adaptive rectification control system of claim 11, wherein, The bifurcation control center comprises a thermal management system, a micro-channel copper cold plate is embedded in the FPGA chip heat dissipation substrate, and the cold plate is connected with a secondary cooling circuit; the cooling medium is perfluoropolyether fluid, which is driven by a magnetic pump to flow through a finned heat exchanger and exchange heat with the exhaust gas of the expander; the control system dynamically adjusts the flow of the cooling liquid according to the chip junction temperature, and automatically reduces the frequency and starts the auxiliary semiconductor cooling module when the junction temperature is detected to exceed the limit.

15. The cryogenic nitrogen producing adaptive rectification control system of claim 11, wherein, The two-phase flow coordination terminal is provided with a safety interlocking mechanism, double cut-off valves are arranged at the outlet of the liquid methane storage tank, the burst pressures of the two-stage safety valves are respectively set to 1.1 times and 1.3 times of the design pressure; the frequency converter of the reflux pump is provided with a current harmonic filter and a regenerative braking unit; the position feedback of the actuator adopts an absolute photoelectric encoder, and the encoder signal is transmitted through double CAN buses; the shell of the coordination terminal is a nitrogen-filled positive pressure explosion-proof box, and the box is provided with a vent and a flame containment grid.

16. A deep-cooling nitrogen production variable load condition adaptive rectification control device integrated with the deep-cooling nitrogen production variable load condition adaptive rectification control system according to any one of claims 11-15, characterized in that, The control device further comprises: The anti-displacement tower foundation structure has a tower body support system with a universal hinge, comprising an axial bellows compensator and a shear limiting base; The shock-resistant electrical cabinet assembly comprises a partitioned explosion-proof control cabinet, which is provided with an SIL 3-level safety instrument system, and a cabinet positive pressure ventilation system provided with a dew point monitoring and automatic drain valve; An online mass spectrometry monitoring module is arranged on the nitrogen gas pipeline at the top of the tower, a quadrupole mass spectrometer is arranged in an ion source chamber, an ultra-high vacuum is maintained in the ion source chamber through a magnetic levitation molecular pump, and the detector output signal is connected to a core processor through an EMI filter.

17. A computer-readable storage medium storing executable program code, wherein the program code, when executed, causes a processor to perform operations comprising: When the program code is read by the computer, the adaptive rectification control method for variable load conditions of cryogenic nitrogen production is realized, and the steps further include: when it is detected that the output frequency spectrum of the rectification tower foundation accelerometer falls into the rock mass characteristic frequency band, activating a three-level control protocol, reconstructing the fractional order controller parameters; calling a dual-mode anti-disturbance mechanism; starting the tower foundation hydraulic damper against mode, controlling the reciprocating movement amplitude and frequency of the damper piston according to the phase difference feedback by the laser range finder, and forming a dynamic balance force opposite to the foundation vibration.

Images