113 results about "Strongly coupled" patented technology

The invention discloses a regional energy internet hierarchical control method considering flexible loads, which comprises the following steps of: constructing a global layer, dividing a regional energy internet into a plurality of electrically strongly coupled sub-regions by adopting a dynamic partitioning algorithm, aggregating equivalent adjustable capacities of the flexible loads in the sub-regions, and performing rolling optimization in an hour-level time scale, an improved Benders decomposition algorithm is adopted to accelerate solving and generate a partition-level power regulation instruction; constructing a regional layer, designing a thermodynamic-economic hybrid constraint model according to flexible load physical characteristics and user behavior elasticity, introducing a user comfort recovery time constraint and elastic electricity price response function, and decomposing an adjustment instruction issued by the global layer to each flexible load cluster in a minute-level time scale by adopting a parallel ADMM optimization algorithm; and constructing a load layer, constructing a millisecond-level adjustable load priority switching mechanism based on FPGA hardware acceleration and finite-state machine logic, calculating a load adjustable power boundary in real time, and dynamically adjusting a response queue in combination with system frequency deviation.

The invention belongs to the technical field of dry quenching, and provides a dry quenching boiler inlet temperature prediction method based on a DeepSeek large model. According to the method, the inlet temperature T6 of the coke dry quenching boiler is predicted by using the data, namely the process variable, changed in real time in the coke dry quenching process; the key control variable is a manually controlled variable, the change of the process variable is controlled / adjusted by changing the key control variable, and automatic control is realized by utilizing a prediction-control joint optimization method and by means of reinforcement learning. According to the prediction method provided by the invention, a complex nonlinear relationship and interaction among multiple variables can be captured, and strong coupling and nonlinear influence among multiple process parameters in the dry quenching process can be processed; through fine tuning on a large amount of historical data, the method can automatically learn and adapt to different working conditions, reduces the dependence on artificial feature engineering, and improves the generalization capability and prediction performance of the model.

The invention belongs to the technical field of automatic control, and particularly relates to an improved linear gain digital control method for a power supply system. The problems that a traditional fixed linear gain digital controller is poor in control precision, robustness, response speed and stability, prone to overshoot and oscillation, heavy in calculation burden, large in hardware consumption and the like in a complex dynamic environment and a nonlinear or strong coupling system are solved. According to the method, an operation state and environment parameters are obtained and processed, and system dynamic characteristics are intelligently evaluated; calculating an optimal control gain factor in combination with a gain scheduling rule or an adaptive learning model, and applying the optimal control gain factor to a linear digital controller for dynamic adjustment; and non-linear compensation processing is carried out on the adjusted control quantity, so that accurate closed-loop control is realized. The control precision can be obviously improved, the system response can be accelerated, the external interference suppression capability can be enhanced, the hardware dependence can be reduced, overshoot and oscillation can be eliminated, and high-precision, low-power-consumption and wide-adaptability digital control can be realized.

The invention provides an LLM command line generation tool based on OpenHarmony and a transplantation method, belongs to the technical field of industrial Internet of Things, and aims to solve the problems that a traditional TypeScript / Node. Js scheme has limitation, is insufficient in local reasoning performance, safety performance and reliability and is high in deployment cost. The tool comprises an ETS development environment module, a code migration module, a dependency library adaptation module, a security enhancement module certificate and a deployment and operation module, a zero-dependency transplantation scheme is designed for special environments of unpacked management tools such as OpenHarmony, the problem of strong coupling of a traditional Node.js application to the environment during operation is solved, and the safety of the Node. Js application is improved. And the deployment threshold of the heterogeneous system is further reduced.

The invention discloses a geographic dependency coupling parameter optimization method based on four-dimensional set variation analysis, and belongs to the technical field of ocean data assimilation. Comprising the steps that independent points are selected, a parameter initial disturbance field set is generated, A-4DEnVar data assimilation is carried out, and full-grid parameters are obtained based on independent point parameter value interpolation. According to the method, the mode parameters closely related to the interaction of the coupling components in the coupling mode are optimized based on the A-4DEnVar data assimilation method under the strong coupling data assimilation framework, the influence of strong nonlinearity caused by the coupling mode can be well treated, meanwhile, an adjoint mode is not needed, the problem that the calculation amount of a traditional algorithm is huge is solved, and the method is suitable for large-scale popularization and application. And the problem of simulation precision reduction caused by globally unified coupling parameters is avoided. A corresponding independent point selection scheme does not need to be formulated for a specific problem, and the workload of manual point selection is also reduced. In conclusion, the coupling mode deviation can be efficiently corrected by utilizing the method, technical support is provided in the aspect of improving the accuracy of climate prediction and forecasting, and the method has a relatively good application prospect.

Methods and systems for mitigating the effects of defects in a quantum processor are provided. A mitigation system includes a quantum processor comprising a plurality of qubits. The system includes a light emitting source that can be tuned to produce light pulses of different wavelengths. The system includes an array of bandpass filters. Each bandpass filter is aligned with a qubit on the quantum processor and has a unique pass band. The system may include a controller configured to receive a selection of a qubit and to tune the light emitting source to emit a light pulse having a wavelength that falls within a range of a bandpass filter that is aligned with the selected qubit. The light pulse is used to scramble an ensemble of strongly coupled two-level system (TLS) in the processor.

The invention discloses a fluctuation signal coupling noise suppression method based on a neural network, which is a coupling noise suppression method in an auto-encoder fluctuation signal based on adjacent position constraint and first arrival pre-sample guidance, and is applied to the field of fluctuation signal data processing. In order to solve the problem that effective signals are prone to being greatly damaged when the coupled noise is suppressed in the prior art, the invention designs a method for representing the characteristics of waveform similarity of adjacent spatial positions of the coupled noise and fluctuation signals based on the weight of a noise sample before first arrival, and under a robust principal component analysis framework, on the basis of a model and data combined driving strategy. Effective signals and strong coupling noise are separated in an unsupervised mode, and damage to the effective signals is reduced.

The invention relates to the technical field of medical detection, in particular to a dynamic modular tumor intelligent staging system and method based on an MCP protocol and LLM. Different staging rules are packaged into independent dynamic staging calculation modules respectively and stored in a dynamic staging calculation module library; meanwhile, tool functions corresponding to different staging rules are registered in the MCP coordination service module through an MCP protocol to form a tool function list, so that the staging rules exist in a modular tool form, and strong coupling of rules and core codes in a traditional system is avoided; after semantic analysis is carried out by the LLM intelligent agent module, a corresponding tool function in a tool function list is automatically matched and called, so that flexible adaptation from multi-modal data to structured TNM data is realized; the analysis logic does not need to be reconstructed when the data source type is newly added; on the other hand, when the staging rules are updated or increased, core codes of the system do not need to be modified, the system does not need to be redeployed, and the flexible response capacity of the system to staging rule iteration and multi-modal data adaptation is greatly improved.

There is described herein a method for determining a parameter of a coupling element in a quantum circuit, the quantum circuit comprising a qubit, the coupling element coupled to the qubit, and a readout element. The method comprises selecting the qubit or the readout element as a probing element, tuning a frequency of the qubit, performing a measurement of a parameter of the probing element on the readout element while an effective coupling rate of the probing element and the coupling element is in the strong coupling regime, and determining the parameter of the coupling element from the parameter of the probing element.

The application relates to the technical field of yaw slip ring state monitoring, and specifically discloses a multi-dimensional state operation monitoring method and system based on a yaw slip ring, which comprises the following steps: obtaining an original operation data set of the yaw slip ring, wherein the original operation data set contains operation parameters in the electrical, mechanical, attitude and environmental dimensions which are collected via sensors and subjected to redundancy checking; constructing a noise suppression matrix based on the physical mechanism of the yaw slip ring, filtering isolated fluctuations and cross-modal abnormal values in the original operation data set by using the noise suppression matrix, and inputting the noise suppression matrix into a convolutional neural network model to obtain a fusion feature vector. By introducing the noise suppression matrix constructed based on the physical mechanism and the channel interaction mechanism guided by the physical topology, the technical problem that a lightweight model is difficult to extract cross-modal strong coupling features at an edge end is effectively solved, and the sensitivity and accuracy for early composite faults are significantly improved without increasing the computing power burden.

The invention discloses a reactor nuclear material cross-scale simulation method coupling microstructure simulation and mechanical-thermal performance calculation. The method comprises the following steps: a, planning a most unit sampling point through a most unit method; b, simulating through a mesoscale phase field model to obtain force thermal parameters; c, the force and heat parameters are input into a finite element model to be simulated to obtain force and heat performance; d, the mechanical and thermal performance is input into the phase field model for simulation to obtain the structure morphology of the target nuclear material; e, obtaining a macroscopic scale-mesoscopic scale coupling model by circulating the step c and the step d; f, obtaining a strong coupling model through physical time axis control; and g, constructing a database of the relationship between the mechanical and thermal parameters and the mechanical and thermal performance through a data-driven neural network algorithm. According to the method, strong coupling is achieved through parameter transmission of the mesoscale phase field model and the macroscale finite element, the macroscale and the mesoscale are connected through the most element method, the parameter transmission precision can be adjusted according to convergence, and novel reactor material mechanical and thermal performance prediction can be achieved.

This invention relates to the field of wireless communication network optimization technology, specifically a three-dimensional communication coverage method that jointly optimizes the downtilt angle of ground base stations and the deployment location of airborne base stations. Based on an integrated air-ground network, this method designs a two-layer optimization architecture for network optimization. By handling the location selection of airborne base stations and the adjustment of antenna tilt angles of ground base stations separately, it avoids strong coupling between discrete and continuous variables, thereby preventing superlinear expansion of the solution space. Simultaneously, by introducing SPSA simulated gradient technology and the ZO-Adam optimization algorithm, it overcomes the non-smoothness and non-convexity problems of the system's spectral efficiency function, ensuring the stability and convergence of the optimization process. Furthermore, the two-layer optimization architecture effectively alleviates the path dependency problem of greedy algorithms, achieving a balance between global search and timeliness.

The present invention discloses a method, system and medium for adjusting the power setting value of a reactor control system; relates to the technical field of reactor power control; performs online fault detection on an actuator of the reactor control system; when a fault in the actuator is detected, determines a fault symptom variable and an operating power setting value; pre-processes the fault symptom variable to obtain an input variable, inputs the input variable into a constructed fuzzy tuning model to calculate an operating power compensation value; adjusts the power setting value of the reactor control system by combining the operating power setting value and the operating power compensation value; for a multi-variable, strongly coupled nonlinear small pressurized water reactor system, when a feedwater valve fault and a control rod drive mechanism jam occur, a fuzzy tuning model with strong uncertainty processing capability is used to establish a relationship between the operating power compensation value and the fault symptom variable, automatically adjusts the reactor power setting value, and realizes fault-tolerant control of the reactor control system.

A molecular sensing device and molecular sensing method that are based on quantum plexcitonic states on strongly coupled nanophotonic systems are disclosed. The molecular sensing device includes a metasurface, a layer of a semiconducting and / or dielectric material disposed onto a layer of electrically conducting material, a metal nanohole array disposed onto the metasurface, a biological recognition element disposed onto a surface of the metal nanohole array, and where the surface of the metal nanohole array is configured to receive a bioanalyte and a plurality of quantum emitters. The molecular sensing device can be a mixture of quantum emitters, comprising gold, having more than one aspect ratio. The molecular sensing device may include a spectroscopic instrument, including a portable detector. The one or more biological recognition elements are configured to recognize a communicable disease in response to the presence of a specific biological recognition element.

This invention belongs to the technical field of digital design of Roots blower equipment, specifically involving a digital twin-based digital design method for the dynamic performance of blowers. This method includes constructing a parametric structural model of multiple components of the blower, establishing a multi-physics coupled model encompassing aerodynamics, structure, heat, electromagnetics, and acoustics, applying dynamic boundary conditions such as flow disturbances, inertial forces, and base disturbances, constructing a digital twin model synchronized with the physical blower state, performing a master-slave strongly coupled joint solution, extracting multi-field response indices and performing normalization analysis, and finally outputting optimized design parameters for the blower through a multi-objective optimization feedback mechanism. This invention enables multi-domain balanced design among the blower's structural performance, electromagnetic response, thermal stability, and acoustic characteristics, improving modeling accuracy and parameter optimization capabilities, and is suitable for the digital development of high-reliability Roots blower equipment under complex operating conditions.

This invention relates to the field of optical technology, specifically to a multi-mode FP-anapole-exciton system that achieves multiple strongly coupled Rabi splits. The system comprises basic units arranged in a two-dimensional periodic pattern, with a square period. Each basic unit includes, from bottom to top, a first dielectric layer, a first metal layer, a composite layer, a second metal layer, and a second dielectric layer. The composite layer includes a transparent dielectric and a heterogeneous disk embedded within it. The heterogeneous disk includes, from top to bottom, a first dielectric disk, a semiconductor disk, and a second dielectric disk. This invention utilizes incident light waves to induce excitation of multiple FP-cavity modes, anapoles, and excitons. Through precise structural design and parameter adjustment, it achieves strong coupling between these multiple FP-cavity modes, anapoles, and excitons, expanding the number of Rabi splits, enhancing the strength of the three-mode coupling, and obtaining multiple large Rabi splitting energies. It has broad application prospects in ultrafast optical switches, Bose-Einstein condensates, and low-threshold lasers.

The invention relates to a phase modifier network-related test simulation platform based on multi-physics field coupling. The platform comprises a power grid fault generation device, an electromagnetic field theoretical calculation device, a thermal field theoretical calculation device, a structure field calculation device and a multi-source input coupling device. According to the multi-physical field coupling simulation platform constructed in the scheme, the limitation of a traditional isolated calculation method is thoroughly overcome through a real-time interaction mechanism of dynamically integrating an electromagnetic field, a thermal field and a structural field. Under a power grid fault working condition, the platform accurately simulates a strong coupling effect among electromagnetic loss, temperature distribution and mechanical deformation (such as local overheating caused by eddy-current loss, air gap offset caused by thermal expansion and closed-loop influence of electromagnetic imbalance aggravated by eccentricity); and the output actual parameter set (electromagnetic force density, temperature rise curve, air gap eccentricity and the like) is highly close to the real operation state of the phase modifier set.

The invention discloses a metamaterial non-reciprocal absorber based on a'lying 'structure, a conductive patch is attached to the side surface, a supporting surface is made of polyimide, a layer of substrate is arranged above and below the structure, the absorber comprises a substrate and a'lying' metal unit array periodically arranged on the surface of the substrate, the inclination angles of the metal units are continuously increased in the preset direction, through asymmetric distribution of space angles, forward incident electromagnetic waves and the metal units generate strong coupling absorption, reverse incident electromagnetic waves are high in transmission due to weakening of the coupling effect and air waveguide, and efficient non-reciprocal response can be achieved without an external magnetic field. According to the invention, the structure is in a lying type and has the characteristics of high integration level, nonreciprocity, broadband response and the like, the working bandwidth can reach ten times of frequency, and the device can be applied to scenes of flexible electronics, on-chip optical communication, ultrathin radar stealth coatings and the like.

This invention discloses a bandpass frequency-selective surface structure based on strongly coupled pairs, composed of multiple basic structural units. Each basic structural unit has a vertically symmetrical structure along its thickness direction. Both the upper and lower symmetrical portions include a first metal layer and a first dielectric layer in symmetrical positions. The first metal layer contains a first resonator pair, a second resonator pair, and a third resonator pair, each using metal patches as resonators. A second metal layer and a second dielectric layer, separated along the thickness direction of the basic structural unit and in asymmetrical positions, are disposed between the upper and lower symmetrical portions. The upper and lower symmetrical portions are connected by metal conductive posts. This invention operates stably at a center frequency of 5 GHz, introduces four transmission zeros, achieves a fourth-order frequency response within 8.4% of the relative bandwidth, has an insertion loss of less than 0.49 dB, and 40° angular stability, exhibiting superior performance.

The invention belongs to the technical field of photocatalyst composite materials, and particularly relates to a covalent bond bonded MoBTx / CdS composite material and a preparation method and application thereof.According to the MoBTx / CdS composite material, rod-shaped CdS and two-dimensional layered MoBTx are compounded through an in-situ growth strategy, covalent bonding between Mo atoms and S atoms is achieved at an interface, and the MoBTx / CdS composite material is obtained. Therefore, a strong-coupling heterogeneous interface is constructed. The covalent bond mediated composite structure can significantly enhance visible light absorption and optimize carrier separation and transmission kinetics, thereby improving the activity and stability of hydrogen production by photocatalytic decomposition of water. The material is excellent in environmental adaptability and can keep good hydrogen evolution performance under complex water quality and low-temperature conditions. The research provides a new interface engineering idea for designing an efficient and stable semiconductor-metal boride photocatalyst.

The application discloses a shortwave miniaturized antenna based on electromagnetic super-metamaterial resonance mechanism and belongs to the technical field of wireless communication and antenna. The antenna comprises a main radiator, a plurality of electromagnetic super-metamaterial resonance units and a feeding structure, wherein the physical length of the main radiator is far less than the working wavelength; the super-metamaterial resonance units are arranged in the near-field region of the main radiator and are strongly coupled with the main radiator, magnetic resonance or electric resonance is generated at a target shortwave frequency point, a frequency splitting effect is generated, the resonance frequency of the antenna is greatly reduced, and deep miniaturization is realized; and the feeding structure is used for coupling radio frequency energy. The application can reduce the size of the shortwave antenna to less than 1 / 5 of the size of a conventional antenna, meanwhile, a relatively high radiation efficiency is maintained, frequency electric tuning can be realized through integrated adjustable elements, and the application is suitable for shortwave communication occasions with limited space, such as mobile platforms and portable devices.

The invention discloses a material multi-constitutive parameter inversion method and related equipment, and the method comprises the steps: S1, inputting a candidate constitutive parameter combination into a pre-trained proxy model, and obtaining a mechanical response dimension reduction representation corresponding to the candidate constitutive parameter combination; s2, based on the mechanical response dimension reduction representation corresponding to the candidate constitutive parameter combination and a predetermined primary function set, reconstructing to obtain a mechanical response prediction curve corresponding to the candidate constitutive parameter combination; s3, performing matching degree calculation on a mechanical response prediction curve corresponding to the candidate constitutive parameter combination and a target experiment mechanical response curve; s4, judging whether a convergence condition is met or not according to the matching degree: if yes, outputting the current candidate constitutive parameter combination as an inversion result; and if not, updating the candidate constitutive parameter combination and returning to the step S1. The objective of the invention is to overcome the problem of low efficiency caused by strong coupling among parameters and high calculation cost of direct problem solving in a traditional inversion method.

This invention relates to the field of microfluidic control technology and discloses a method and system for detecting and controlling microfluidic droplet formation. The method first constructs an ideal concentric ring geometric model by acquiring the target parameters of the microcapsule and establishes a force-shape coupling response matrix. Based on this, a preset velocity triplet is solved to generate the initial droplet. Then, droplet image sequences are acquired and fitted with real-time concentric ring groups. These are geometrically superimposed with the ideal concentric ring geometric model to generate a deviation ring domain, which is then decomposed hierarchically to obtain geometric deformation characteristic components. Finally, a coordinated velocity adjustment vector is calculated using the force-shape coupling response matrix, a composite fluid correction field is constructed, and the deviation ring ablation cycle is iteratively executed until geometric convergence is achieved. This invention effectively solves the oscillation problem caused by strong parameter coupling in multiphase flow fields by constructing a quantitative mapping and inverse decoupling mechanism between fluid and geometric shape, achieving coordinated and precise control of the microcapsule's external dimensions and internal structure.

This invention discloses a pulsed laser control system integrating dynamic knowledge graphs and deep learning. The invention relates to the field of pulsed laser processing control technology and comprises, in sequence, a dynamic knowledge graph construction module, a deep learning hybrid model module, and a dual-loop collaborative optimization module. This pulsed laser control system integrating dynamic knowledge graphs and deep learning achieves accurate analysis and real-time optimization of the coupling relationships of multi-domain parameters in pulsed lasers through deep collaboration between the dynamic knowledge graph and the deep learning model. By employing cross-domain coupling quantization technology of the dynamic knowledge graph, the interactive influence of time-domain, spatial-domain, and frequency-domain parameters is transformed into an online-updable coupling coefficient matrix, solving the problem of traditional methods struggling to dynamically analyze strong parameter coupling. Through the condition-triggered computational architecture of the hybrid model, based on the dynamic activation of sub-modules using an LSTM-CNN-GNN heterogeneous network, the computational load is significantly reduced while ensuring prediction accuracy.

The present application relates to the technical field of motor design, and particularly relates to a permanent magnet motor iron core size optimization method based on a particle swarm algorithm. The method comprises the following steps: selecting geometric parameters such as stator tooth width to construct particles and initializing a particle swarm; in each round of iteration, an agent model constructed in advance is used to predict the electromagnetic performance index of the particles and to calculate the fitness value; a gradient vector is constructed by perturbing the global optimal solution, and the maximum value of the cosine similarity of the gradient vectors between geometric parameters is calculated as the correlation attribute value; the speed update logic is dynamically corrected in response to the correlation attribute value: an inertia weight inversely proportional to the correlation attribute value is applied to the strongly coupled parameters to suppress oscillation, and an inertia weight proportional to the gradient module length is applied to the weakly coupled parameters to accelerate convergence; until the convergence condition is met, the optimal iron core size is output. The present application effectively solves the oscillation problem caused by strongly coupled parameters, and significantly improves the optimization accuracy and efficiency.

This application discloses a three-dimensional ultra-wideband power divider, belonging to the field of radio frequency communication technology. It includes an input port, a strongly coupled line unit, a weakly coupled line unit, a first output port, and a second output port. The input port, strongly coupled line unit, and weakly coupled line unit are connected sequentially. The first terminal of the second end of the weakly coupled line unit is connected to the first output port, and the second terminal of the second end of the weakly coupled line unit is connected to the second output port. The weakly coupled line unit includes any number of coupling lines and multiple first resistors, with the multiple first resistors bridging between two coupling lines of the weakly coupled line unit. The strongly coupled line unit is a three-dimensional integrated vertical substrate circuit structure. This application overcomes the contradiction between ultra-wideband, high isolation, low loss, and mass production through a three-dimensional integrated structure, gradually changing coupling lines, and a multi-level bridging resistor network, providing a high-performance power divider solution for high-frequency communication systems.

The invention relates to the technical field of industrial automation control, and discloses an impedance control method and system based on AI Agent and MCP architecture, and the method comprises the steps: obtaining user side multi-modal test demand data and circuit board engineering data; constructing a standardized demand constraint set; constructing a three-level atomic power service architecture; a standardized test context is generated; generating a concurrent candidate action sequence of multiple test heads; security verification is carried out through a multi-sub-agent cross evidence mechanism; executing closed-loop adaptive parameter adjustment; and outputting an impedance test result. The technical problems that in the prior art, a software module is strongly coupled, the artificial data conversion efficiency is low, and especially dynamic safety impedance control cannot be achieved under the multi-head cooperative operation condition are solved. Due to the fact that the layered atomic power architecture is constructed, separate control and self-adaptive optimization of demand understanding, process scheduling and physical execution are achieved, and the impedance testing efficiency and the industrial safety reliability are improved.

The invention belongs to the technical field of software testing, and particularly relates to a method for simplifying a reverse proxy system to acquire multi-dimensional traffic, which comprises the following steps of: constructing a closed-loop self-consistent system integrating traffic generation and response simulation, thoroughly breaking the strong coupling relationship between test logic and physical deployment architecture in the traditional test method, and improving the test efficiency. The complexity of a test environment is greatly simplified, hardware resources and operation and maintenance cost are remarkably reduced, testers are endowed with extremely high flexibility to define, combine and dynamically schedule multi-dimensional service scenes with any complexity, meanwhile, the reproducibility of the test is ensured by means of a deterministic data generation mechanism, and the test efficiency is improved. Therefore, a brand-new and systematic solution is provided for efficient, deep and comprehensive verification of the reverse proxy system.

The invention relates to the technical field of microfluidic control, and discloses a microfluidic droplet forming detection, regulation and control method and system.The method comprises the steps that an ideal concentric annulus geometric model is constructed by obtaining microcapsule target parameters, a force-shape coupling response matrix is established, and a preset flow velocity triad is solved according to the force-shape coupling response matrix to generate initial droplets; then collecting a liquid drop image sequence, fitting a real-time concentric circle group, carrying out geometric superposition on the real-time concentric circle group and the ideal concentric circle geometric model to generate a deviation annular domain, and carrying out hierarchical decomposition to obtain geometric deformation characteristic components; and finally, calculating a collaborative flow velocity regulation vector by combining the force-shape coupling response matrix, constructing a composite fluid correction field, and iteratively executing a deviation ring ablation cycle until geometric convergence is realized. According to the method disclosed by the invention, the problem of oscillation caused by strong coupling of parameters in a multiphase flow field is effectively solved by constructing a quantitative mapping and reverse decoupling mechanism of the fluid and the geometrical morphology, and collaborative precise control of the external dimension and the internal structure of the microcapsule is realized.

This invention discloses a method for predicting the quality of ultrafast laser multi-stage processing of complex microstructures based on multi-task deep learning, belonging to the field of laser processing quality prediction technology. The method first performs data augmentation on sparsely sampled historical processing data to construct a sufficient training sample set. Based on this, a quality prediction model based on multi-stage, multi-task learning is constructed. This model explicitly models the transmission of physical effects between stages by introducing a feature extraction module of a structured state space, and utilizes task clustering and specific feature modulation to achieve the collaborative prediction of multiple strongly coupled quality indicators such as roughness, depth, and perpendicularity. During training, an adaptive loss function based on uncertainty is used to automatically balance multi-task learning. In application, by inputting the process parameters for each stage, the model can output the predicted values ​​of the quality indicators for each stage of the entire process. This invention solves the problem of accurate quality prediction under the conditions of multi-stage cumulative effects and multi-indicator coupling relationships.