36results about How to "Guaranteed numerical stability" patented technology

The invention discloses a lithium-ion battery SoC (State-of-Charge) estimation method based on improved and self-adaptive dual unscented-Kalman-filters. The method mainly comprises the following steps: 1) establishing a state filter and a parameter filter of a lithium ion power battery based on an equivalent circuit model; 2) coupling the state filter and the parameter filter of the lithium ion power battery based on the equivalent circuit model, and building dual unscented-Kalman-filters; and 3) inputting operation parameters of a to-be-detected lithium ion power battery into the dual unscented-Kalman-filters, and carrying out parameter correction and stage-of-charge (SoC) estimation on a lithium ion power battery equivalent circuit model. According to the method in the invention, the effectiveness of cholesky decomposition is guaranteed, the problem that iteration is stopped due to the fact that a covariance matrix is indefinite caused by reasons such as an initial value error, noisedisturbance, calculation module floating point error and the like is overcome, and the numerical stability of the filtering process and the robustness of the algorithm are enhanced.

The invention provides a strong tracking UKF filter method based on sampling point changing. The strong tracking UKF filter method based on the sampling point changing comprises the steps that (1) initial parameter setting is carried out on a system; (2) Sigma points are sampled according to an orthogonal transformation sampling point method, a corresponding prediction equation is determined, and time and measurement are updated; (3) fading factors are calculated; (4) new one-step prediction covariance is calculated by using the fading factors, the Sigma points are recalculated, and auto-covariance and cross covariance after the fading factors are introduced are obtained through nonlinear measurement function propagation; (5) filter updating is carried out to the end. According to the strong tracking UKF filter method based on the sampling point changing, the problem of non-local sampling of the system is solved effectively, the precision of the system is improved, and the system is made to have certain strong tracking capability. The strong tracking UKF filter method based on the sampling point changing can be used for solving the problems of poor robustness and filtering divergence when a model of the system is uncertain, solves the problem of the non-local sampling in the high-dimensional system, and expands the application range of strong tracking filter. In an MEMS/GPS combined navigation system, the positioning and attitude determination performance of the MEMS/GPS combined navigation system can be improved through the method.

The invention discloses a prediction method of gas-assisted injection molding gas penetration based on a finite volume method. The method includes the following steps that (1) a three-dimensional node control body is constructed by cavity entity grids to serve as a finite volume; (2) forming process parameters are set; (3) material parameters are set; (4) the flow velocity, pressure and the temperature of plastic melt are calculated; (5) the flow leading edge of the plastic melt is updated, and the time step length is determined; (6) whether the filler content of the plastic melt reaches an appointed volume or not is judged; (7) the flow velocity, pressure and the temperature of the plastic melt and the flow velocity, pressure and the temperature of gas are calculated; (8) the flow leading edge of the plastic melt and the penetration leading edge of gas are updated, and time steps are determined; (9) whether a cavity is filled or not is judged. The method is based on the three-dimensional finite volume method, a traditional finite volume method is improved, coupled flowing of the plastic melt and gas is calculated, consideration is given to strong nonlinearity among viscosity, the velocity, pressure and the temperature, and distribution of the gas-assisted injection molding gas penetration leading edge and various physical quantity fields can be accurately predicted.

A low-voltage active distribution network congestion management method based on node marginal price, Firstly, according to the basic principle of positive semidefinite programming algorithm, The nonlinear constraint in the low voltage active distribution network congestion management model based on intelligent soft switching is transformed into the positive semi-definite constraint by the positivesemi-definite relaxation technique, The original non-convex nonlinear model is transformed into a positive semi-definite programming model, and the symmetric component method is used to reduce the mutual impedance coupling between phases, so that the numerical stability of positive semi-definite relaxation can be improved by increasing the numerical difference of the elements in the positive semi-definite matrix. Then the sensitivity factors of branch active power, branch reactive power, node voltage variation and network loss to node injection power are calculated based on the linearizationapproximation principle. Based on the sensitivity factor, a node marginal pricing model is established. The invention can effectively reduce the difficulty of problem solving and ensure the calculation accuracy while improving the calculation speed. It can effectively solve the congestion management problem of low-voltage active distribution network.

The invention discloses a vehicle model verification system and method for guardrail safety performance simulation evaluation. By using finite element method computer simulation as a technical means,the method simplifies an actual vehicle and establishes and verifies a vehicle model in a computer analogue simulation system; therefore, whether the highway guardrail structure reaches a corresponding protection level can be accurately reflected and evaluated. Specifically, a sub-component model, a whole vehicle model and a vehicle model before and after the whole vehicle collision are tested andverified through a sub-component test module, a whole vehicle test module and a whole vehicle collision test module respectively, so that the accuracy, numerical stability and effectiveness of the vehicle model are ensured. The test verification of the vehicle model is beneficial to realizing standardization and normalization of the guardrail safety performance simulation system, is a basic guarantee for scientific research technicians to accurately determine the guardrail safety performance, and is also a basic basis for owners and supervisors to check and accept the guardrail design and usesafety performance.

The invention discloses an active power distribution network dynamic simulation method based on a multistage optimization pursuit variational iteration method. The method comprises the steps offirstlycarrying out the improvement of a variational iteration algorithm in two aspects: on one hand, considering a solved approximate analytical solution at the right side of a correction functional,and onthe other hand, preferentially calculating a differential equation with higher coupled frequency such that an optimization pursuit variational iteration method with faster off-line convergence is provided,secondly, solving approximate analytical solutions corresponding to dynamic elements in a photovoltaic system and a synchronous generator system in the active power distribution network based onthe method,and then, introducing a multi-stage mechanism is introduced.An approximate analytical solution is applied to continuous time intervals, and a limited convergence domain of an original variational iteration method is broken through, so that the method can be applied to active power distribution network simulation requiring an accurate solution in a larger time span. Compared with a traditional numerical integration method, the method is higher in solving speed.

The invention relates to a method for mixed explicit implicit integration applicable to random dynamic simulation of an active power distribution network and belongs to the field of active power distribution network simulation. Random characteristics and multi-time scale characteristics in the process of operating the active power distribution network are considered. The method is based on a mixed explicit implicit solving idea, an explicit integration algorithm is adopted to integrate a random disturbance term forming an active power distribution network random dynamic model, and an implicit integration algorithm is adopted to integrate a deterministic model for forming the active power distribution network random dynamic model. According to the method, numerical solution advantages of the explicit integration algorithm and the implicit integration algorithm can be brought into full play, and the method is applicable to random dynamic simulation calculation of a system under malfunction or operation conditions. Compared with a conventional explicit integration algorithm, simulation calculation efficiency and numerical stability are improved while numerical precision needed for simulation is met, and the random dynamic simulation of the active power distribution network under the conditions of considering integrated access of photovoltaic and wind power generation systems is implemented.

A nuclear thermal coupling iteration method of supercritical water reactor core comprises: after carrying out m times of neutronics calculation, carrying out one time of thermotechnical hydraulic calculation, wherein m is set neutronics calculation iteration times, m is greater than 5; in the process of carrying out the neutronics calculation, introducing a self-adaption relaxation factor omega and utilizing a power density distribution iteration function to adjust the power density distribution obtained by carrying out the second to the mth neutronics calculation, wherein the adjusted power density distribution is used in the next time of neutronics calculation or thermotechnical hydraulic calculation. The invention solves the problem that iteration calculation in the nuclear thermal coupling iteration calculation of the SCWR core is difficult to converge and improves the calculation efficiency and numerical stability of the nuclear thermal coupling calculation.

The invention discloses a state estimation method based on a tensor method. According to the state estimation method, through calculation of a tensor equation, then a nonlinear correction is obtained. Calculation and storage of a Hessian matrix are avoided. The calculating efficiency of nonlinear state estimating is better improved, and meanwhile, robustness and especially numerical stability of the estimation of the nonlinear state are kept.

The invention relates to a battery energy storage system scheduling method considering dynamic charge and discharge efficiency. The method comprises the following steps: A, acquiring basic informationof a battery energy storage system, and performing segmented linearization on a charge and discharge efficiency characteristic curve; B, acquiring scheduling information such as running limit, initial and final electricity quantities, a predicated time-sharing electricity price or a price multiplier and the like of the battery energy storage system; C, establishing a scheduling model; D, calculating the dynamic upper limit and lower limit of storing the electric quantity by the battery energy storage system in each time period; E, performing dynamic planning on reverse recursion to obtain anoptimization model of a core sub-problem needing to be repeatedly solved in each time period; F, using geometric features of the core sub-problem optimization model to sequentially solve an optimal path in each horizontal sub-region from bottom to top; G, solving the optimal path in the horizontal sub-regions; H, processing the optimal path jump; I, performing dynamic planning on forward recursion, and calculating the optimal charge and discharge electric quantity and power decision of each time period; and J, entering the next time period along with the schedule development, updating the scheduling information in the step B, and executing the step C. According to the method, an effective way is provided for optimized operation of the battery energy storage system, and an engineering practical value is achieved.

The invention provides a model output post-processing method and device and computer equipment. A recognition system obtains a probability matrix output by a speech recognition model, carries out window search in the probability matrix, carries out rule processing on the probability matrix based on a threshold value, eliminates invalid paths in the probability matrix, the probability calculation of the next window interval of the probability matrix is calculated based on the probability result of the previous window interval, and the probability value corresponding to each window interval is calculated. Finally, the command word corresponding to the window interval with the highest probability value is selected as a speech recognition result. According to the method and device, the numerical stability of the probability matrix is effectively ensured through threshold-based rule processing. Log domain calculation is replaced by a probability accumulation calculation mode, so that the calculation amount is reduced. In the window searching process, the probability calculation of the latter window interval is based on the probability result of the former window interval, so that the repeated calculation process is reduced, the calculation amount is further reduced, and the processing speed is effectively accelerated.

The invention discloses an UKF algorithm-based lithium ion power battery state estimation method. According to the method, on-line estimation and updating are performed on parameters of a time-variable charge state space model and a health state space model by a joint estimation strategy, so that model accuracy under a dynamic test working condition can be effectively ensured, and the lithium ionpower battery state estimation accuracy is improved. Moreover, during unscented transformation, quasi-linear processing is performed on a measurement equation, so that calculation cost during the cycle iterative process is reduced very well; by on-line estimation and correction on statistic characteristic parameter of process noise, and the self-adaptive fault-tolerant capability of the UKF algorithm is greatly improved; and the convergence of filtering is traced in real time, filtering divergence is prevented by correcting covariance when a filtering result is in a divergent trend, and the numerical value stability during the filtering process is ensured.

An order reducing method includes setting up second order system equation to describe the interconnected circuit, carrying out frequency shift and linearization for the secondary system, using second order block Arnoldi algorithm to obtain a set of orthogonal standard base, obtaining order reduced secondary system by orthogonal projection, linearizaing the secondary system to be primary circuit equation with block structure in original secondary equation for obtaining small size RCS equivalent circuit.

The invention belongs to the technical field of perfect matching layer construction, and discloses an M-UFSPML model, a construction method, an intelligent terminal and a server, and the method comprises the steps: carrying out the Fourier transform of a weak form of a time domain second-order linear elastic anisotropy wave equation, and obtaining a corresponding frequency domain equation; on the basis of the concept of a multi-axis complete matching layer, adopting a complex coordinate continuation technology to carry out continuation on space variables in a weak form frequency domain wave equation, and obtaining a newly constructed artificial boundary, namely a non-split convolution multi-axis perfect matching layer. The precision of the newly constructed non-split convolution multi-axis perfect matching layer is improved compared with that of a previous multi-axis perfect matching layer, and meanwhile it can be guaranteed that calculation lasts for a long time and is stable. The newly constructed non-split convolution multi-axis perfect matching layer adopts a non-split form, the direct discrete space second-order partial derivative is avoided, the calculation efficiency is higher, the numerical result can ensure the numerical stability of more than 100,000 step lengths, and the precision is obviously superior to that of a low-order artificial boundary.