9 results about "Parameter distribution" patented technology

The invention provides a ship lock structure simulation parameter correction method and device and electronic equipment, and the method comprises the steps: obtaining historical displacement actual measurement data of a ship lock structure, and forming standardized actual measurement data; determining to-be-corrected parameters according to the ship lock structure simulation model; performing global search on the to-be-corrected parameters by using a covariance matrix adaptive evolution algorithm, and obtaining distribution characteristic parameters which enable the matching degree between a displacement response predicted value output by the model and standardized measured data to be optimal through adaptive learning of the correlation between the parameters and iterative optimization so as to determine initial parameter distribution of the to-be-corrected parameters; constructing an augmented state space model in combination with the to-be-corrected parameters and the displacement response predicted value; and taking real-time ship lock displacement data as an observation value, and adopting a Kalman filtering algorithm to perform recursive assimilation on augmented state variables in the spatial model so as to update and correct to-be-corrected parameters. The real-time ship lock displacement data can be assimilated, and the to-be-corrected parameters in the ship lock structure model can be dynamically updated and corrected.

The present application belongs to the field of urban underground pipeline damage sonar evaluation, and discloses a pipeline cracking damage sonar intelligent evaluation method, which extracts two characteristic parameters of sonar echo signal correlation and energy, analyzes the parameter distribution under different damage states by using GMM, and constructs a probability density distribution benchmark library; in the test application stage, the GMM probability density distribution of the characteristic parameters is calculated, the Frechet similarity with the benchmark library is calculated, and the state with the largest similarity is taken as the damage state. The present application uses the training data to construct the GMM probability density distribution benchmark library, uses the similarity between the test data and the training data to carry out probabilistic damage evaluation, and with the increase of the monitoring data, the accuracy of the model will be higher and higher.

The invention discloses a non-parametric method-based product random degradation mechanism inspection method, which comprises the following steps that: degradation amount sample data is collected and expressed as xijl, l is a stress environment sequence number, j is a time point sequence number, and i is a product sequence number; constructing a sample random distribution model according to the sample data, calculating a variance estimated value # imgabs0 # and a mean value estimated value # imgabs1 # of the sample at each moment, and constructing a parameter distribution model; the method comprises the following steps: calculating the reliability R1 (t) of a product at a specified moment according to a scale characteristic # imgabs2 # of a product degradation amount distribution mean value changing along with time in each stress environment, a shape characteristic # imgabs3 # Gamma distribution shape parameter # imgabs4 # and a rate parameter # imgabs5 # of the product degradation amount distribution mean value changing along with time in each stress environment, and calculating the reliability R1 (t) of the product at the specified moment according to the size characteristic # imgabs2 # and the rate parameter # imgabs5 # of the product degradation amount distribution mean value changing along with time in each stress environment; and determining a degradation mechanism consistency test condition of the measurement environment stress condition k and the sample environment stress condition l, and judging whether the product under the measurement environment stress condition k has a degradation mechanism change or not. According to the technical scheme, the degradation condition of the product can be flexibly and visually analyzed and judged.

This invention discloses a method and system for constructing the probabilistic stability domain and inverting parameter uncertainties in milling chatter, belonging to the field of CNC machining and cutting dynamics prediction technology. The method first solves for a given dynamic parameter vector using a fully discrete method. i Stability boundary of the spindle speed-axial depth of cut plane; then use multiple sets i Boundary training agent model to quickly predict arbitrary i The boundary. Given i When distributing parameters, a proxy model is invoked via Monte Carlo sampling to generate a flutter probability distribution map and a preset probability. p The stability boundary is determined. For a limited amount of experimental data, a hierarchical Bayesian model is constructed, with and as hyperparameters as priors. A likelihood function is built using experimental stability / instability labels. The posterior distribution is obtained through Markov chain Monte Carlo sampling, and the probabilistic stability region is updated. Variance is estimated by reducing the likelihood using common random numbers. This invention significantly reduces computational overhead, enabling rapid construction of the probabilistic stability region and parameter distribution inversion, thus aiding in the selection of processing parameters and risk control.

This invention discloses a bridge construction deflection sensitivity analysis method, comprising the following steps: Step S1, establishing a historical database; Step S2, standardizing each parameter in the historical database; Step S3, based on the standardized parameter data, taking the average of all data for that parameter, and generating random numbers with a sample size of N from the parameter distribution; Step S4, integrating a hybrid learner; Step S5, establishing a three-dimensional deflection model of the main beam using the hybrid learner; Step S6, obtaining multiple deflection values ​​as sample output values ​​using the arithmetic mean median method; Step S7, calculating the conditional expectation of each parameter with respect to the sample output values; Step S8, calculating the magnitude of the influence of the parameters on the main beam deflection using the Sobol sensitivity analysis system. This invention, by analyzing the influence of each parameter on the main beam deflection, guides the strict observation and control of parameters with a significant impact during construction, thereby avoiding resource waste and reduced construction efficiency.

The invention belongs to the technical field of integrated circuits, and relates to a flow model-based CMOS intensive model generation method, which specifically comprises the following steps of S1, key parameter extraction; s2, carrying out distributed modeling; and S3, generating parameters. Constructing a reversible transformation network by using a flow model, directly learning joint probability distribution of an extracted intensive model basic parameter group, and capturing a nonlinear dependency relationship and complex distribution characteristics among parameters by maximizing a log-likelihood function; according to the trained flow model, a real-time generation module is established through potential spatial disturbance and reversible mapping, the module can accurately control the intensity of statistical magnitude such as fluctuation and aging in a generated parameter group by adjusting distribution parameters (such as injected Gaussian noise or offset) of potential variables, and meanwhile, an original correlation structure between the parameters is kept. According to the method, the problem of repeated sampling caused by limited number of models in an intensive model sample library in large-scale integrated circuit design is solved, and the parameter group generated in real time can keep original parameter distribution and correlation among parameters.

To obtain arithmetic results of a plurality of parameters having a distribution with higher accuracy compared to the prior art.SOLUTION: An information processing apparatus comprises an acquisition unit, a matrix generation unit, and an arithmetic unit. The acquisition unit acquires, as operation objects, a plurality of parameters whose values have variations and for which probabilities of respective values are given as a distribution. The matrix generation unit generates a plurality of parameter matrices respectively corresponding to the plurality of parameters, wherein each of the plurality of parameter matrices has the number of parameters as the number of dimensions, a width of the distribution of each parameter as a length of the dimension corresponding to each parameter, and each parameter matrix has each value obtained by equally dividing the distribution of the corresponding parameter as a value of each element of the parameter matrix. The arithmetic unit updates each element of the plurality of parameter matrices by performing an operation on each element of the plurality of matrices.SELECTED DRAWING: Figure 1

The invention relates to a method (100) for identifying a parameter distribution for modeling a technical system (50), comprising the following steps: - determining (101) a nominal parameter estimate on the basis of provided measurement data and at least one pre-given system dynamics model of the technical system (50); - determining (102) a distribution of parameter variations on the basis of a quantification of the uncertainty due to parameters of the nominal parameter estimate; - providing (103) the nominal parameter estimate and the distribution of parameter variations as a basis for modeling the technical system (50) taking into account the uncertainty due to parameters.

This invention relates to a method for extrapolating the extreme values ​​of time-domain wave loads based on truncated samples and a three-parameter distribution. The method comprises the following steps: extracting amplitude samples from the original time history; removing a predetermined number of smaller amplitude samples to obtain truncated samples (x1, x2, ... x...). m The cutoff sample (x1, x2, ... x) obtained by fitting the three-parameter generalized extreme value distribution G(x; σ, γ, μ) is selected. m The method involves estimating the scale parameter σ, shape parameter γ, and position parameter μ using the probability weighted moments method, and obtaining the corresponding extrapolated extreme values ​​X of the wave load. This invention provides an extrapolation method for wave load extreme values ​​that balances rationality, stability, and accuracy. Compared with existing technologies, the positive advancements of this invention are: it has the advantages of reasonable amplitude extraction, appropriate distribution function selection, and high extrapolation accuracy. While ensuring accurate and stable fitting of wave load samples, it can conveniently perform extreme value extrapolation.