91results about How to "Good practical engineering application value" patented technology

The invention relates to a method for determining the flexural capacity of a high-strength steel-UHPC composite beam. The method is characterized by comprising the following steps that (1) the corresponding section curvature k when UHPC reaches crushing serves as the section curvature of the steel-concrete composite beam, and the strain distribution rule of the section of the composite beam is determined; 2) aiming at two conditions that a neutralizing shaft is positioned in a UHPC plate and the neutralizing shaft is positioned in a steel beam, simplifying a UHPC stress diagram of a pressed area into a triangle, and by considering the strain hardening characteristic of high-strength steel, simplifying a steel beam web stress diagram into a trapezoid, and determining internal force distribution of the section of the main beam under the two conditions; and (3) according to the bending moment balance equation of the section of the main beam, the method for determining the flexural capacity of the high-strength steel-UHPC composite beam is established. According to the method, the problem that a method for calculating the flexural capacity of the high-strength steel-UHPC composite beam is lacked is solved, and technical support is provided for promoting application of the high-strength steel-UHPC composite beam in bridge engineering.

The invention discloses abrasion-resistant concrete based on industrial waste residues and a preparation method of the abrasion-resistant concrete. The abrasion-resistant concrete is prepared from 1050 to 1200 parts by weight of coarse aggregate, 480 to 550 parts by weight of coarse sand, 135 to 210 parts by weight of tungsten slag, 300 to 360 parts by weight of Portland cement, 75 to 90 parts by weight of superfine tungsten slag powder, 1 to 1.5 parts by weight of high-performance water reducing agent and 140 to 160 parts by weight of water, the tungsten slag is obtained by drying industrial waste tungsten slag generated by producing ammonium paratungstate at 80 DEG C to constant weight, the superfine tungsten slag powder is obtained by grinding the tungsten slag, and the fineness of the superfine tungsten slag powder is 1%-5% of screen residue of a 45-micron screen; the coarse aggregate is formed by mixing small stones with the particle size of 5-20 mm and medium stones with the particle size of 20-40 mm. Through the combined action of the tungsten slag, the superfine tungsten slag powder and the Portland cement, the overall abrasion resistance of the concrete is improved, the problem of cyclic utilization of the tungsten slag industrial waste residues is solved, and meanwhile it is guaranteed that the workability of the concrete is not reduced.

The invention discloses a time delay estimation model optimization method and system based on kernel density, a medium and equipment, and the method comprises the steps: obtaining time delay related data of a power distribution network communication network, and the time delay related data comprises the number of access subjects of the power distribution network communication network, the kernel density function bandwidth and a time delay measurement value; constructing a time delay estimation model of the power distribution network communication network by adopting a non-parametric kernel density estimation method according to the time delay related data; and optimizing the fixed bandwidth of the time delay estimation model by adopting a Gaussian kernel function and a geometric distance to obtain a time delay estimation optimization model. Under the condition of multivariate main body access, the adaptive nonparametric kernel density estimation method is utilized to adjust the time delay estimation model, the probability density function is estimated according to the characteristics of the data, different characteristics in the data are conveniently captured, and the obtained time delay estimation model of the communication network of the power distribution network is high in data processing precision and high in reliability. The method has good engineering application value.

The invention discloses a synthesis method and application of rare earth silicate nano composite powder for an environmental barrier coating, and aims to solve the problems of silicon volatilization and poor thermal stability in the thermal spraying process of the rare earth silicate environmental barrier coating. The synthesis method comprises the following steps: 1, preparing asymmetric mesoporous silica based on a nano silicon source; 2, co-precipitating the asymmetric mesoporous nano silicon dioxide serving as a nucleation center and rare earth ions, and calcining to prepare high-activity nano precursor powder; and 3, uniformly mixing the nano precursor with potassium sulfate and sodium sulfate, and calcining to obtain the rare earth silicate nano powder. The asymmetric mesoporous silica is selected as a silicon source, a silicon-rich end is supplemented from the nanoscale, silicon loss caused in the spraying process is made up, meanwhile, the rare earth silicate nano-powder is prepared by combining a chemical coprecipitation method with a molten salt method, the effect of optimizing the purity of the nano-powder is achieved, preparation of a high-quality environmental barrier coating material is achieved, and the application prospect is wide. And the thermal stability of the environmental barrier coating is improved.

The invention relates to the technical field of sewage pipe network optimization, and provides an information entropy-based sewage pipe network inflow and seepage detection sensor optimization arrangement method, which comprises the following steps of: performing random abnormal event simulation of which the total number is M, and obtaining an information content vector of each node of a sewage pipe network; a joint entropy matrix EM of the whole sewage pipe network is constructed by calculating the joint entropy between every two nodes; performing first-stage clustering on the basis of the joint entropy, clustering the number of nodes into an NS class according to the number of sensors needing to be arranged, and performing second-stage re-clustering on the basis of a sewage pipe network topological structure to cluster the number of the nodes into an NC class; nC class nodes are ranked according to ADR values, a node is selected from each of NS classes with the highest ADR value to arrange a sensor, and a combination with the maximum multivariate joint entropy serves as a final optimal arrangement scheme. The application method of the information entropy in sewage pipe network abnormal event detection is provided for the first time, and a new thought is provided for optimized arrangement of sewage pipe network sensors.

The invention relates to the technical field of weld joint dynamic modeling, in particular to a thin-wall shell weld joint dynamic modeling method based on a common shell unit, which comprises the following steps of: S1, dispersing the middle surface of a welding part by a global shell grid, and simulating a weld joint by using a CSE instead of a rigid connection unit; s2, simplifying the modeling process of the CSE by adopting a grid common node processing method; s3, constructing an interface shell unit to replace part of GSM on the two sides of the welding area, and achieving the grid consistency requirement; s4, the CSE welding seam modeling method is applied to welding of the thin-wall shell for dynamic simulation analysis; and S5, performing result verification and adjustment on the data obtained by simulation in the step S4. According to the method, the ISE unit is introduced to meet the grid consistency requirement of the non-matched connection interface; building a weld joint dynamic model based on a CSE and RE modeling method, and carrying out dynamic simulation analysis and experimental research; the problem that the finite element analysis accuracy is affected due to the fact that an existing connecting unit cannot accurately simulate the real characteristics of weld joints is solved.

The invention discloses a rapid rendezvous and docking orbit design method for a multi-task spacecraft, and the method comprises the steps: obtaining a phase adaptation lower limit and a phase adaptation upper limit through employing a rapid rendezvous and docking orbit transfer strategy according to a preset target aircraft orbit height and a preset active aircraft injection initial orbit; obtaining the earliest moment when the spacecraft performs orbital maneuver; analyzing a phase deviation maximum value and a phase deviation minimum value caused by the initial orbit deviation of the active aircraft at the earliest moment of orbit maneuver of the spacecraft according to the initial orbit deviation of the active aircraft in orbit injection; according to the fact that the minimum value of phase deviation caused by initial orbit deviation of the active aircraft at the earliest moment of orbit maneuver of the spacecraft is not smaller than the rapid rendezvous and docking phase adaptation lower limit, the orbit phase value at the orbit injection moment is obtained. According to the method, the orbital transfer interval is prolonged, the task requirement of an on-orbit test is met, rapid rendezvous and docking are also considered, and the adaptability of the spacecraft to the fault working condition is enhanced.

The invention relates to a geological disaster prediction method and system based on an optimized PU-XGBoost model, and belongs to the technical field of geological disaster monitoring and risk assessment, and the method comprises the steps: obtaining multi-source data including InSAR deformation monitoring data and historical geological disaster point data, taking the historical geological disaster points and InSAR significant deformation points in the InSAR deformation monitoring data as a positive sample set P, and carrying out the prediction of the InSAR significant deformation points in the historical geological disaster points; a PU-Learning strategy of a Spy sample is introduced to construct a reliable negative sample system, an XGBoost hyper-parameter is optimized and searched in combination with introduction of a BTO optimization algorithm, an optimal hyper-parameter combination is obtained, and an optimal model is obtained through training of a positive sample set P and a negative sample set RN; and performing geological disaster susceptibility prediction and hierarchical mapping on the research area based on the optimal model, thereby realizing high-precision and high-reliability evaluation of the geological disaster susceptibility of the research area.

The invention belongs to the technical field of semantic communication, and particularly relates to a hybrid semantic communication system based on a pulse neural network. The method comprises the following steps: performing semantic extraction on a server side by using an ANN model; then, the semantic information is converted into a data format suitable for SNN model processing after being subjected to time-space dimension semantic discretization; the received semantic information is processed at the edge end by utilizing an SNN model which is easy to deploy and low in power consumption, so that efficient and high-precision semantic understanding is realized. And different frequency characteristic optimization distribution modes of the intelligent distribution sub-module are matched with corresponding pulse neurons to trigger calculation only when a pulse event is detected, and the threshold condition of multiple signal triggering of a single pulse neuron can be set, so that as many as three levels of triggering type power consumption optimization can be realized. Under the condition that the FPGA is combined and used, high-efficiency performance is kept when large-scale data are processed. The method is particularly suitable for low-power-consumption and efficient communication application scenes.

The invention relates to a high-precision steel surface defect detection method suitable for a complex industrial environment, and aims to solve the problems of feature coupling and background interference caused by data scarcity, long tail distribution and multi-defect coexistence. According to the method, a conditional generative adversarial network (CGAN) and a convolutional neural network (CNN) are combined, an SE-Net channel attention mechanism is integrated, and the method is used for automatic detection and classification of multiple defects on the steel surface. By introducing a physical constraint CGAN data generation method, a multi-defect coexistence composite image conforming to industrial reality can be generated, so that the diversity and accuracy of a training data set are effectively enhanced. The model optimizes the expression of defect features through multi-level feature extraction and an attention mechanism, and enhances the distinguishing ability and recognition precision of multiple defects. Experiments show that the method is excellent in performance in a multi-defect identification task, the accuracy rate of 98.89% and the F1 score of 99.72% are achieved, the method is remarkably superior to a traditional detection method and other deep learning models, and the method has high robustness and practical value in a complex industrial environment.

The invention discloses an internal model control-based network construction type full-power wind power converter control method. The method comprises the following steps of: acquiring a control quantity of a network construction type full-power wind power converter; determining an active power current given value by using preset internal model control according to the control quantity; determining a modulation voltage of a wind power converter machine side converter according to an active power current given value by utilizing current loop control; utilizing direct-current voltage synchronous control, and outputting a control synchronous angle according to the control quantity; by means of voltage and current double-loop control, output control output voltage is output according to the control quantity; and realizing machine-side converter control of the wind power converter according to the modulation voltage, and realizing grid-side converter control of the wind power converter according to the control synchronization angle and the control output voltage.

The invention discloses a design method and device for a new energy automobile battery area temperature monitoring system. According to the actual situation of a new energy automobile, the design key points of a battery area temperature monitoring system and the overall scheme of a battery area multi-channel temperature monitoring system are provided, and on the basis, hardware circuits of a battery area temperature acquisition terminal and a receiving terminal are designed; software programming is carried out according to a hardware circuit, corresponding hardware functions are realized, and a battery area temperature acquisition terminal program and a receiving end terminal program are designed; finally, temperature monitoring and display software based on a Windows terminal and a mobile terminal is designed, real-time display and monitoring of temperature data are achieved, an abnormal temperature alarm function is achieved by setting an alarm threshold value, and the safety of new energy vehicle management and the informatization degree of vehicle monitoring are effectively improved.

This invention discloses a multi-frequency, multi-target range-velocity deambiguation method and system based on phased array fire control radar, mainly addressing the low accuracy of existing methods in handling multi-target ambiguity. The implementation scheme is as follows: acquiring measurement points from the phased array fire control radar and generating a full-space candidate point set; discretizing the range-velocity plane into a grid, counting the number of candidate points and non-repeating frequency groups within each grid block; dividing the grid into blocks based on block-level neighborhood statistics to generate the candidate point set; calculating the unique frequency support number of the candidate point set and performing clustering to obtain candidate point clusters; statistically estimating the candidate point clusters, outputting target range and velocity estimates, forming the deambigued target set for the current frame. This invention can reduce the probability of false clustering structures formed by erroneous unfolding points, clutter points, and false alarm points, improving the accuracy and real-time processing efficiency of multi-target range-velocity joint deambiguation in complex scenarios, and can be used for multi-target signal processing and target tracking in phased array fire control radar.

The application discloses a kind of millimeter wave broadband filtering antennas with wide-angle beam scanning function and array, from top to bottom include broadband radiator, metal ground layer and microstrip line feed layer in sequence, the broadband radiator includes first rectangular patch, second rectangular patch and driving patch;By controlling the size of first metal patch, second metal patch and driving patch, the resonant frequency of broadband radiator is regulated;By adjusting the distance between first rectangular patch, second rectangular patch and driving patch, and the spacing of adjacent first metal patch and adjacent second metal patch, the coupling amount of broadband radiator is regulated, and the bandwidth of broadband radiator is further regulated.The application has simple and compact structure, wide impedance bandwidth, and can cover four millimeter wave frequency bands.

The invention discloses a muon imaging detection system and method based on scintillation fiber and SiPM reading. Comprising at least one detection unit, the detection unit comprises a scintillation optical fiber array and SiPM array support plates, the scintillation optical fiber array is formed by arranging a plurality of scintillation optical fibers in parallel, the two ends of the scintillation optical fiber array are each provided with one SiPM array support plate, SiPM arrays are integrated on the SiPM array support plates, the number of the SiPM units is the same as that of the scintillation optical fibers, and the SiPM units and the scintillation optical fibers are in one-to-one correspondence; the SiPM unit is optically coupled with one end of a corresponding scintillation optical fiber signal, receives a scintillation optical signal transmitted in the corresponding scintillation optical fiber and generates an electric signal; the signal processing unit is used for preprocessing the electric signal output by the SiPM to obtain a processed signal; and the positioning reconstruction unit is used for reconstructing the space motion trail of the muon according to the processed signal. The system can realize three-dimensional high-precision positioning and track reconstruction of muons, and has the advantages of high integration level, modular design and the like.

A very low frequency circuit frequency response parameter calibration system is suitable for high-precision frequency response calibration of a signal processing circuit working in a frequency range of 300 Hz to 100 kHz. The system comprises a signal acquisition module and an FPGA. The signal acquisition module comprises three synchronous acquisition channels, two output channels of a to-be-calibrated circuit, and a GPIO outputting a calibration signal of the FPGA. The FPGA is configured to generate a calibration signal after receiving a configuration instruction of an upper computer system. The calibration signal is divided into three paths: one path enters a very low frequency signal acquisition module as a reference signal for calibration comparison; the other two paths are injected into two input channels of the to-be-calibrated circuit to drive the to-be-calibrated circuit to generate corresponding outputs; the three sampling channels of the very low frequency signal acquisition module are synchronously controlled to start data acquisition and complete isochronous sampling tasks, ensuring the synchronization of the three paths of data in the time domain; after sampling is completed, the three-channel data is preprocessed, and then the preprocessed data is uploaded to the upper computer system for visual processing.

The application discloses a computer-implemented multi-tissue biomechanics simulation method, system and medium, and the method comprises the following steps: S1, receiving patient medical image data, generating tissue labels through tissue segmentation, and generating a discrete graph representing biological tissues; S2, inputting reference coordinates, tissue types, boundary conditions, loads and material parameters to nodes; S3, generating displacement, velocity, acceleration or state increment by using a graph neural network containing an antisymmetric message passing layer; S4, mapping the decoding output to the physical displacement scale through a material scaling factor; S5, calculating the stress field and the strain field according to the physical displacement field and outputting deformation visualization, risk maps or surgery planning suggestions; and the graph neural network is pre-trained by minimizing the total potential energy containing a volume-weighted average Jacobian volume energy item and a cell isochoric deviatoric energy item. The application solves the problems of cross-tissue material scale imbalance, volume locking, cell inversion and dependence on finite element labels.

This invention discloses a vehicle trust assessment method for cooperative decision-making at unsignalized intersections, belonging to the field of intelligent transportation and autonomous driving cooperative control technology. The method first acquires multi-source state and interaction information of the target vehicle and conflicting vehicles, constructing reference position estimates and physical interaction characteristics. Then, based on multi-source observation consistency analysis, it calculates the information reliability of the target vehicle and, combined with the spatiotemporal interaction relationship between vehicles, calculates the physical interaction trust of the conflicting vehicle pairs. Further, it determines the comprehensive trust based on the information reliability and physical interaction trust, and generates a safe time interval between vehicles based on the comprehensive trust. This interval is then used as a conflict constraint parameter input into the intersection cooperative scheduling model to achieve adaptive optimization control of vehicle passage order and entry time. This method can adaptively adjust the safe time interval between vehicles under conditions of uncertainty in vehicle state information, differences in vehicle execution consistency, and complex interaction relationships, improving the safety, stability, and traffic efficiency of cooperative decision-making at unsignalized intersections.

The application provides a variable driving parameter silicon carbide MOSFET active driving method, circuit and device. The active driving method comprises: in the on process, a power supply is controlled to apply a first weak driving level in a first time period and a first strong driving level in other time periods, the starting time of the first time period is the time when the silicon carbide MOSFET enters the Miller platform, and the ending time is the time when the drain current reaches the current overshoot peak value; in the judgment process, the power supply is controlled to apply a second weak driving level in a third time period and a second strong driving level in other time periods, the starting time of the third time period is the time when the drain-source voltage is equal to the DC bus voltage, and the ending time is the time when the drain-source voltage reaches the voltage overshoot peak value. The method has good versatility and high engineering application value, reduces the additional switching loss caused by weak driving, maintains high switching speed while ensuring the overshoot suppression effect, and thus realizes a better compromise between overshoot, loss and switching time.

The application discloses a temperature compensation method and system suitable for insulation resistance sensors. The application aims to solve the problem that it is difficult to simultaneously consider single-board compensation accuracy and temperature calibration cost in a wide temperature and wide resistance range. The method comprises the following steps: selecting a plurality of sample sensors to calculate sample error data and converting the sample error data into an error feature vector; based on the error feature vector, the sample sensors are clustered and divided to form a plurality of sensor categories; a basic compensation model library is constructed according to the categories; the error feature vector of a sensor to be calibrated, except the sample sensors, is calculated; the error feature vector is compared with the feature center of each category in the basic compensation model library, and a matching category is determined according to the comparison result; a single-board correction parameter is solved in combination with the basic compensation model; and based on the basic compensation amount and the single-board correction parameter, an insulation resistance value after compensation is output. The application can realize effective compensation of single-board temperature drift difference in mass production only at a small number of calibration points.

A non-intrusive multi-load load monitoring method and system based on shared time-series coding and load-specific refinement is disclosed, relating to the fields of smart grid, non-intrusive load monitoring, and energy data analysis. This method addresses the lack of existing non-intrusive multi-load load monitoring methods that can simultaneously address shared modeling, load difference representation, training stability, and engineering deployment feasibility. The process involves generating a fixed-length input window; after the input window is compressed to form latent space temporal features, multiple shared time-series coding modules extract the shared time-series representations for the corresponding multiple loads; these shared time-series representations are then fed into multiple independent load-specific refinement heads, which generate power predictions and state predictions for the target loads; residual gating is applied to the power predictions of the target loads using the state predictions, yielding the final decomposition results for the multiple target loads.

The invention discloses a device and a method for testing anisotropic mechanical properties of strong heterogeneity rocks, the device comprises a rotatable rock sample fixing device, a workbench and a sliding track, and a scratching cutter control and data monitoring system, the workbench and the sliding track device main body are fixed on a table top and are convenient to support and move; the slidable scratching system moving track supports the slidable scratching system to move left and right; the rotatable rock sample fixing device is located on the left side above the table top and can fix the position of a rock sample and change the rock scratch angle. The scratching tool control and data monitoring system is located on the right side of the table top and connected with the sliding rail, the slidable scratching system can scratch rocks, and the data monitoring system can monitor and collect tangential force data borne by a tool bit and displacement data of the tool bit in real time; the device and the method for testing the anisotropy of the shale based on the scratches can be researched so as to judge the anisotropy mechanical property of the strong heterogeneity rock.

The application discloses a high-parallelism storage-computation integrated polar code decoding method based on coset decoding, and belongs to the technical field of channel coding and decoding. The method comprises the following steps: constructing an error pattern group and a check constraint subgroup; performing coset division on the error pattern group to obtain a set of mutually disjoint cosets; screening a limited number of minimum cosets based on the minimum Hamming weight of the cosets; and performing parallel decoding processing on the paths corresponding to the minimum cosets in the storage-computation integrated architecture. k The number of paths in the traditional multi-bit decoding is reduced from 2 to a preset number L through coset division, and the parallel execution of the likelihood information updating and the path metric calculation is realized in combination with the calculation-type storage array, so that the decoding performance is ensured, the calculation complexity and the hardware resource consumption are significantly reduced, and the decoding throughput and the energy efficiency are improved.

This invention discloses a multidimensional indexed modulation chaotic communication method and apparatus based on neural network-assisted detection, belonging to the field of wireless communication and chaotic spread spectrum communication technology. The method includes: firstly, segmenting the input bitstream to obtain modulation bits, domain index bits, codeword index bits, and reference index bits; then, selecting the corresponding transform domain codebook based on the domain index and the corresponding codeword basis vector based on the codeword index, while simultaneously generating an orthogonal chaotic sequence pool and selecting the corresponding reference chaotic sequence; next, performing MPSK modulation on the modulation bits, and combining the codeword basis vector and the reference chaotic sequence to generate a multidimensional indexed modulation chaotic signal, which is then transmitted through a wireless channel after multi-carrier modulation; the receiving end demodulates, filters, and samples the received signal, and inputs it into a multi-task convolutional neural network detector to achieve joint detection of modulation symbols, domain index, codeword index, and reference index, thereby recovering the transmitted bitstream.

The invention provides a pumped storage unit multi-source signal abnormal information extraction and fault diagnosis method based on an enhanced hierarchical Poincare map index, and relates to the field of intelligent operation and maintenance of a hydroelectric power generation system. A pumped storage unit multi-source vibration signal feature extraction tool fusing multi-frequency nonlinear dynamic indexes and a feature level fusion strategy is developed, a pumped storage unit intelligent fault diagnosis model integrating signal acquisition, feature extraction and mode recognition is constructed, and accurate recognition of the abnormal state of the unit is achieved. Rich state information contained in multi-source vibration signals of the pumped storage unit can be deeply excavated, on the premise that normal production and power generation of the hydropower unit are not affected, efficient recognition of the hydropower station on the early abnormal state of equipment is improved, shutdown and maintenance cost caused by equipment faults is reduced, and the working efficiency of the hydropower station is improved. And an effective technical means is provided for improving the operation reliability of the shaft system and ensuring the safe and stable operation of the hydroelectric generating set.

The application discloses a construction method of a drag-reducing isolation structure applied to a precast pile, and belongs to the technical field of pile foundation construction, which comprises a precast pile, a drag-reducing isolation material, an isolation protective layer, a cement-soil pile, an energy circulation channel and an energy medium pump; the drag-reducing isolation material is uniformly distributed on the surface of the precast pile, the isolation protective layer is wrapped outside the precast pile and the drag-reducing isolation material, the cement-soil pile is outside the isolation protective layer and is bonded between the isolation protective layer; the energy circulation channel is arranged inside the precast pile, and the energy circulation channel and the energy medium pump are connected through a pipeline. The construction steps are as follows: precast pile manufacturing; drag-reducing isolation material manufacturing; isolation protective layer manufacturing; drag-reducing isolation material smearing; isolation protective layer wrapping; cement-soil pile construction; implanted precast pile processing; energy circulation reduces and eliminates negative friction drag; and continuous maintenance. The application can reduce the influence of negative friction drag on the precast pile through multiple active drag reduction, and has good engineering application value and popularization value.

The invention relates to alumina toughened zirconia ceramic powder as well as a preparation method and application thereof. The invention develops a modification preparation method which integrates precise proportioning, multi-stage PH regulation and control, composite coating agent modification and controllable aging process, and definitely adopts a solid-to-liquid ratio of (50-200): (100-1500) (the solid-to-liquid mass ratio is (0.5-2): (1-15)) to ensure the stability of the slurry, so that uniform dispersion of powder, strong interface bonding and collaborative optimization of mechanical properties and chemical stability are realized. Compared with the prior art, the invention has the advantages of good powder dispersibility, strong phase interface bonding, high slurry stability, excellent ceramic mechanical properties and chemical stability, and the like.

The application discloses a Mongolian-Chinese multi-expert neural machine translation method and system based on decision layer relative strategy optimization, and the method comprises the following steps: acquiring Mongolian-Chinese bilingual parallel corpus and preprocessing to obtain Mongolian-Chinese translation training data; training a basic translation model based on the Mongolian-Chinese translation training data to obtain a Mongolian-Chinese task adaptation model; performing sensitivity analysis on the Mongolian-Chinese task adaptation model to obtain hierarchical sensitivity results of each network layer; according to the hierarchical sensitivity results and the complexity of an input sentence, different network layers of the Mongolian-Chinese task adaptation model are allocated with different quantization precisions to obtain a light-weight translation model; a course type training sequence is constructed, and an optimized Mongolian-Chinese translation model is obtained based on a composite reward function; a Mongolian sentence to be translated is input into the optimized Mongolian-Chinese translation model to generate a corresponding Chinese initial translation; and a final Chinese translation is determined according to the Mongolian sentence to be translated and the Chinese initial translation.

The application relates to the technical field of wind power equipment decommissioning and recycling, in particular to a decommissioned wind power blade on-site modular efficient disassembling method. First, the decommissioned wind power blade is subjected to structure scanning to obtain internal main beam area and shell structure distribution information; then, an unfoldable temporary bearing framework is introduced into the blade interior to form a temporary bearing system in the blade interior; under the condition that the temporary bearing system continuously supports, the blade shell is subjected to longitudinal unloading cutting and circumferential modular cutting; then, the bonding interface between the main beam and the shell is subjected to softening treatment, and the expansion separation mode is used to realize directional decoupling of the main beam structure and the shell structure; finally, each module formed by disassembling is turned over, transported and classified and recycled. The application can maintain the overall force stability of the blade during the disassembling process, reduce the structural collapse risk, improve the modular disassembling efficiency and the composite material classification and recycling integrity, and has good engineering application value.

The invention particularly relates to a multi-jacket liquid nitrogen ice plug freezing method using a non-uniform gap structure, and belongs to the technical field of nuclear power plant pipeline freezing, and the method comprises the following steps: step 1, determining the total length L1 of jackets and the number N of the jackets of a pipeline freezing section; 2, calculating the value range of the total gap length L2 of the jacket of the frozen section of the pipeline; and step 3, designing and distributing non-uniform gaps. By designing the non-uniform jacket gaps, energy regulation and control are carried out in the ice plug forming process, so that the initial freezing position is accurately controlled, the complete freezing time is shortened, the pressure bearing capacity of the ice plug is improved, and finally energy conservation and efficiency improvement are achieved.