105results about How to "Strongly nonlinear" patented technology

The invention discloses a traffic flow prediction method, a traffic flow prediction model generation method and a traffic flow prediction device. The prediction method includes the following steps that: historical traffic flow data of a road section to be predicted in a period before a current time point are obtained; a traffic flow prediction model corresponding to the road section to be tested is obtained from preset corresponding relations between roads and traffic flow prediction models; and the historical traffic flow data in the period before the current time point are inputted to the traffic flow prediction model corresponding to the road section to be tested, so that traffic flow data in a period after current time point can be obtained. Since traffic flow data have high nonlinearity and uncertainty, and a neural network model has a high nonlinear prediction ability, and therefore, the neural network model can be trained according to the historical traffic flow data of the road, and the traffic flow prediction model obtained through the training can accurately predict the traffic flow data in the period after the current time point according to the traffic flow data in the period before the current time point.

The invention discloses a failure diagnosis and health evaluation method based on wavelet power, manifold dimension reduction and dynamic time warping, and aims to improve the feature separability of bearing failure, impeller failure and the mixed failures of a centrifugal pump and realize diagnosis and health evaluation of various states. The method comprises the following steps: firstly, decomposing collected vibration signals of the centrifugal pump into 8 wavelet components by applying wavelet packet conversion; extracting wavelet energy of each component to be taken as a failure feature to obtain an eight-dimensional failure feature vector; then conducting dimension reduction on the eight-dimensional feature by applying a manifold learning method to obtain a three-dimensional feature vector with better separability, simplicity and stability; finally, based on the feature vector, measuring the distance of test data and training data by applying a dynamic time normalization method so as to determine the current failure state and realize failure diagnosis of a bearing. The distance value can also reflect the health degree of the current state, can realize evaluation of the health state of the centrifugal pump, and has the excellent practical engineering application value.

Disclosed is a high-precision data fusion method oriented to high-dynamic non-Gaussian-model robustness measurement. The method includes: taking sampling period fluctuations of a hardware sensor in a high-dynamic system as system random uncertainty for consideration, establishing a filtering model set comprising a UKF (unscented Kalman filter) model and a fuzzy inference system, computing match probability of the UKF model and a current high-dynamic system state according to the Bayes' theorem, updating the match probability in real time, taking the updated match probability as input of the fuzzy inference system, obtaining adaptive estimation probability by the fuzzy inference system, and finally, fusing multiple states based on the adaptive estimation probability to estimate to obtain a final mean value and covariance estimate of state variables of the high-dynamic system. The high-precision data fusion method oriented to high-dynamic non-Gaussian-model robustness measurement has the advantages that not only can data fusion for a high-dynamic, strong-nonlinearity and non-Gaussian-model combined system be achieved, but also the number of pre-stored model sets can be decreased; meanwhile, computational efficiency for model probability updating and measurement robustness of the high-dynamic system are improved.

The invention discloses an aircraft intelligent trajectory reconstruction reentry guidance method, which comprises the steps of generating an offline reference trajectory by adopting a numerical optimization method according to an aircraft reentry initial state and various constraint conditions; setting waypoints, and calculating corresponding reachable domains; setting five normalization parameters to establish a relationship between an optimal track and a roll angle through the change characteristic of the roll angle of the optimal track, proposing an optimization method for solving the corresponding roll angle parameter by using a search algorithm, then selecting sample data points in an accessible domain, calculating training data offline, and carrying out training based on a BP neuralnetwork; tracking a reference trajectory by adopting a guidance method; when reaching the height of the waypoint, generating a new track by using the trained neural network model; and completing trajectory reconstruction, tracking a new trajectory by the aircraft, and continuing to complete reentry guidance. According to the method, the calculation time of trajectory reconstruction is greatly shortened, and the influence of optimization time on guidance performance is reduced, so that the guidance precision of the reentry vehicle in the reentry stage is further improved.

The invention relates to a precipitation prediction method and device based on multi-LSTM model fusion, electronic equipment and a storage medium. The method comprises the following steps: acquiring historical meteorological data of daily granularity in a target area, and segmenting the historical precipitation in the historical meteorological data into a plurality of data sets according to a preset interval; respectively establishing LSTM models of the plurality of data sets and fusing the LSTM models, and combining the LSTM models with other factor prediction models to generate a precipitation prediction model; setting training process characteristics of the precipitation prediction model to complete the model training; optimizing a hyper-parameter of the precipitation prediction model on a verification set through a k-fold cross validation; and inputting a new sampling moment to the precipitation prediction model at the future time to obtain a corresponding precipitation predictionvalue. According to the precipitation prediction method based on the multi-LSTM model fusion provided by the invention, a plurality of sets of historical meteorological data are introduced through themulti-LSTM model fusion to realize the prediction of the precipitation, the perceptual ability of the historical data is improved, and a prediction result is accurate.

The invention discloses a water supply prediction method and a water supply prediction system. The water supply prediction method is characterized in that historical data is analyzed, and a water supply prediction model is established, and a functional relation between an input layer parameter and a water supply quantity output by an output layer is acquired; real-time water supply quantity prediction is carried out according to the established water supply prediction model and required prediction date; the processes of analyzing the historical data and establishing the water supply prediction model comprise but not limited to the process of preprocessing the historical data by adopting a moving average method, the process of determining the network topological structure of the water supply prediction model by adopting the neural network algorithm, and the process of determining the water supply prediction model according to the preprocessed historical data and the corresponding network topological structure. The moving average method and the neural network algorithm are used for the water supply quantity prediction, and therefore calculation complexity is low, obvious periodicity and strong non-linearity are provided, real-time performance is good, the influences of various factors on the water supply quantity are described accurately, the convergence speed of the neural network model is accelerated, and the water supply prediction method and the water supply prediction system are widely used for the water supply analysis field.

The invention relates to a method for predicating an underground fault of a sucker-rod pump oil pumping well on the basis of a multivariable grey model. The method comprises the following steps of: collecting a ground indicating diagram by an indicating diagram remote data collecting system; converting the ground indicating diagram into an underground pump indicating diagram; extracting a curve moment feature vector capable of representing the graphic feature of the underground pump indicating diagram according to the curve moment theory; predicating each feature vector by the multivariable grey model; and calculating the grey correlation degree of each fault type in predicating samples and standards sets, wherein the fault type corresponding to the maximum grey correlation degree is the fault type of the predicating samples. The working system of the oil pumping well can be reasonably regulated, so that the production efficiency of the oil pumping well is improved, and the fault predication on the underground work condition of the sucker-rod pump oil pumping well can be realized; the calculation is simple, the required samples are few, and the model is applicable to data sets in any distribution; and the prediction accuracy is high, and particularly, the prediction precision on the gradually varied fault type is high.

A controlling system of the dynamic matrix for the air separation tower, it includes the local intelligent meter which connects to the tower, data storage device for the history of the data, and the position machine. The intellective meter, data storage device and position machine connect to each other in sequence. The talked position machine is the controlling machine for the dynamic matrix which consists of the controlling part of the ingredient judging and dynamic matrix. The ingredient judging part includes the module of the testing meter, I / O element module and ingredient judging module. The dynamic matrix includes I / O element module, forecasting model module, feedback revising module, rotating optimizing module and output controlling module. It also provides a controlling method for the realization of the dynamic matrix controlling for the tower. This invention can be suitable to the dynamic characteristic, strongly non linear and coupling of the circuit for the air separation, and the integrated controlling effectiveness is good.

The invention discloses a vehicle sideslip angle estimation method based on a second-order sliding-mode observer and belongs to the field of vehicle control. The vehicle sideslip angle estimation method comprises steps as follows: step 1, a vehicle state is detected by a sensor, and specifically, the velocity vx of a running longitudinal wheel of the vehicle is detected by a velocity sensor; the yaw velocity omega in the vehicle running process is detected by a yaw velocity sensor; a front wheel rotating angle delta in the vehicle running process is detected by a steering wheel angle sensor; step 2, the second-order sliding-mode observer is designed, the step 2 comprises a step 2.1 and a step 2.2, in the step 2.1, a kinetic model of a vehicle system is established firstly, and in the step 2.2, the second-order sliding-mode observer is designed according to the kinetic model of the vehicle; step 3, the information about the vehicle state detected in the step 1 is transmitted to the second-order sliding-mode observer designed in the step 2, and a sideslip angle is obtained through operation estimation. A two-degree-of-freedom kinetic model of the vehicle is used, the designed observer has a fractional power item, the steady-state error can be reduced, the observer has a sign function item, and the robustness of the system can be improved.

The invention relates to the field of a printing technology and particularly relates to an offset printing ink color matching method based on a least square support vector machine. The offset printing ink color matching method comprises the following steps of: sampling each foundation ink according to different concentration gradients, mixing foundation inks with m concentration ratios selected within a (1-100)% range of the concentration gradients with a reducer, and then measuring an XYZ value corresponding to each standard color sample obtained by sampling; obtaining an XYZ value of a target sample to be matched, and selecting N inks for color matching, so as to obtain m*N XYZ values and corresponding concentration ratio training sample data; training an LS-SVM (Least Square-Support Vector Machine) function model, and establishing relation of the XYZ values and an ink matching formula, so as to convert the XYZ values into the ink concentration ratios; and inputting the XYZ value of the target sample to be matched into the trained LS-SVM function model, so as to obtain the corresponding ink formula through calculation. With the adoption of the offset printing ink color matching method provided by the invention, the precision requirement on a foundation ink database is small, the color matching is simple and convenient and reliable, and the efficiency is high.

The invention discloses an evolutionary computation-based method for diagnosing compressor fault. The method comprises the following steps: data about the working condition of a compressor is collected and acquired; an observation point is taken as a node, and the correlation between observation points produced by the actual compressor physics; a weighting undirected graph G(V,A) is set up for measured data; the distance between fault data points is gauged by means of a relational graph instead of the traditional Euclidean distance method; the data about the fault of the compressor is converted into a problem with the partitioning of the weighting undirected graph G; the Pareto Principle structural map is adopted for partitioning cost function; an evolutionary calculation chart partitioning computation based on the improvement on the complexity system theory is designed to partitioning the relational graph; and the clustering / classification of the compressor fault data is realized through the optimal computation of the cost function, so that the purpose of fault diagnosis can be achieved.

The invention discloses a hypersonic flight vehicle jumping gliding trajectory analysis and solution method. The hypersonic flight vehicle jumping gliding trajectory analysis solution under the GER coordinate system is converted into the solution of the hypersonic flight vehicle jumping gliding trajectory analysis under the AGI coordinate system; and then the analytic solution under the AGI coordinate system is resloved into a GER coordinate system to obtain a high-precision analytic solution of the longitudinal range, the transverse range, the speed, the height and the trajectory inclinationangle of the hypersonic flight vehicle jumping gliding trajectory, thereby providing support for rapid trajectory planning and trajectory forecasting.

The invention discloses a rodless cylinder servo system position precision control method. The method comprises the following steps: 1, a mathematical model for a pneumatic position servo system is built; 2, based on the mathematical model built in the first step, a tracking differentiator is designed according to an active disturbance rejection control technology; 3, a linear extended state observer is designed, and real-time estimation and compensation are carried out on the total disturbance of the system; and 4, a linear state error feedback controller which is a general PID controller is designed, the total disturbance of the system is compensated in the controller, and thus, the original nonlinear pneumatic servo control system is converted to a linear integral series control system. The method of the invention does not rely on the precise mathematical model for the pneumatic servo system, no pressure detection device needs to be added, all uncertain factors and external interference during an experiment process can be overcome, quick and precise position control on the rodless cylinder is finally realized, and the control precision can reach 0.005 mm.

The invention relates to a soft sensing method for SO2 emission of a power plant based on variable compression BP neural network, comprising the steps of S1, collecting sample data about time variation in a wet flue gas desulfurization system, and determining neurons of an input layer and an output layer of the dynamic neural network according to the collected sample data; 2, performing correlation analysis on that sample data, compressing the original sample variable to remove sample data whose correlation with SO2 in the flue gas at the outlet of the wet flue gas desulfurization system is less than 0.2; Step S3: Compressing by using the variable-BP neural network model wet flue gas desulfurization system, establish wet flue gas desulfurization system prediction model; Step S4, calculating the predicted value of SO2 concentration at the flue gas outlet of the wet flue gas desulfurization system by using the wet flue gas desulfurization prediction model established in the step S3; StepS5: controlling The slurry spraying amount of the wet flue gas desulfurization system by using the predicted value of step S4.

The invention discloses a heat exchanger dirt growth prediction method based on PCA-GA-RBF. The method comprises the following steps: step 1, constructing an optimal heat exchanger dirt training sample; step 2, constructing a GA-RBF-based artificial neural network prediction model, and determining the number of input layer neurons, the number of hidden layer neurons and the number of output neurons of the RBF artificial neural network prediction model; step 3, optimizing the initial parameter value of the RBF artificial neural network prediction model by using a genetic algorithm to obtain theinitial parameter value of the RBF artificial neural network prediction model; optimizing hidden node center and width parameters of the RBF network by using a genetic algorithm; and step 4, substituting the optimized initial parameter value of the RBF artificial neural prediction model into the RBF artificial neural network prediction model, and adjusting the central position and the weight of the hidden layer node of the RBF by adopting an orthogonal least square method. The method is high in learning speed and high in precision.

The invention discloses a four-rotor aircraft attitude control method and controller. For the disadvantage that a four-rotor aircraft system is difficult to establish a precise physical model, the nonlinear dynamic characteristic of the system is described by establishing a RBF-ARX model, which is locally linear and globally nonlinear, by adopting a system identification method; for the characteristic of the model and to satisfy quick response characteristic and relatively high control performance requirements of the system, the model is converted into a non-minimum state space model having anintegration link at first; then, a prediction controller is designed; furthermore, input is parameterized by using a laguerre function; therefore, predictive control of the online optimization variable of the system is reduced; furthermore, a given reference signal can be precisely tracked; and, for the numerical stability problem due to increasing of a prediction time domain, an exponential decay weight is introduced into a target function, so that a closed-loop system has the specified stability. The scheme has the characteristics of being short in online optimization time and steady in numerical value in a relatively large prediction time domain condition, and has relatively high practical value and application prospect.

The present invention is a novel, high resolution, color, three-dimensional (3-D) volumetric display system for dynamic images—the video cube. The video cube consists of an air-tight glass cube filled with a gas mixture and multiple planes of thin wires arranged in alternating orthogonal layers. These wires may be set at voltage potentials capable of producing a glow discharge at the intersection of pairs of wires. Using a computer capable of storing dynamic image data and electronic controllers capable of energizing pairs of wires appropriately at the proper time 3-D dynamic images may be formed from multiple glows between excited wire pairs. The video cube may be used to display complex real-time information from computers and other digital processors with high accuracy for unlimited number of simultaneous unaided observers.

The invention discloses a differential geometry nonlinear control method of spatial mechanical arm anti-interference attitude stabilization. A differential geometry theory is used to carry out linearization conversion on a nonlinear attitude dynamics model so as to acquire a kind of linear space form. And then LQR is used to carry out controller design on a linearized attitude stabilization system so as to acquire a non-linear control law. A spatial mechanical arm attitude stabilization controller based on a differential geometry theory and an LQR control method can guarantee robustness stability of a system to various kinds of uncertainties and a non-cooperative interference, a simple linearity control theory is adopted, a control structure is simple, a calculated amount is reduced and resources on a satellite is saved so that the method is especially suitable for a spatial mechanical arm attitude stabilization task which needs real-time calculation and has a limited calculating capability.

The invention relates to a quadrotor unmanned aerial vehicle parameter distinguishing method based on an improved shuffled frog leaping algorithm. The shuffled frog leaping algorithm is taken as a basis, a frog leaping rule in the algorithm is improved, the global convergence rate of the shuffled frog leaping algorithm is improved, a fitness function is designed according to quadrotor unmanned aerial vehicle parameters, the computation complexity of the algorithm is reduced, and quadrotor unmanned aerial vehicle parameter distinguishing efficiency is improved. The method exhibits certain universality, and parameter distinguishing for the majority of nonlinear systems with a high coupling degree can be realized.

The invention discloses a mesoscale vortex trajectory prediction method. The method comprises the steps of downloading and reading satellite height measurement data, extracting mesoscale vortex characteristic parameters and position information, tracking mesoscale vortexes, establishing a mesoscale vortex trajectory prediction model based on a BP neural network, determining input and output data,training the BP neural network, and inputting the initial positions of the mesoscale vortexes into the trained prediction model for prediction. According to the invention, the BP neural network is used as a prediction model, and the tracked mesoscale vortex position information and the initial ground rotation flow velocity anomaly are used as model inputs, so that the short-term prediction of themesoscale vortex trajectory can be realized. Meanwhile, in the mesoscale vortex tracking process, the minimum distance method and the similarity method are combined, limitation on the mesoscale vortexamplitude is added, the error tracking rate is reduced, the tracking accuracy is guaranteed, and therefore the model prediction precision is improved.

The invention discloses a magnetic suspension small sphere control experiment table based on model design and automatic code generation. The magnetic suspension small sphere control experiment table comprises a mechanical part and an electric part, wherein the mechanical part is a magnetic suspension small sphere device; the electric part comprises a PC (Personal Computer) machine, a simulator, a real-time embedded controller, a power amplifier and an eddy current displacement sensor; the PC machine is taken as a host, is based on a physical model design control system and code generation, and is used for downloading into the real-time controller through the simulator; the output of the controller passes through the power amplifier in order to control the coil current of a magnetic suspension small sphere device; the eddy current displacement sensor is used for acquiring the position information of a small sphere, feeding back the position information to the controller, and controlling the small sphere through a closed loop. Through the combination of the magnetic suspension small sphere and a control system development method based on the model and an automatic code generation technology, a good experiment platform is provided for control theory learning, and the magnetic suspension small sphere control experiment table has the characteristics of short control system development time, low cost and the like.

The invention discloses a mechanical vibration state identification method based on statistics in different orders and a support vector machine. The mechanical vibration state identification method includes the following steps of (1) using a vibration measurement device to collect vibration data of a mechanical system and subjecting the vibration data to segmentation and de-meaning pretreatment; (2) calculating third-order statistics and fourth-order statistics of each segment of the vibration data after pretreatment, and using the third-order statistics and the fourth-order statistics as two feature vectors; estimating fraction low-order statistics of each segment of the data after treatment, i.e., a feature index alpha and a dispersion coefficient gamma as another two feature vectors; (3) classifying and judging vibration states of the mechanical system by means of the support vector machine on the basis of the four feature vectors. The mechanical vibration state identification method based on statistics in different orders and the support vector machine has the advantages that under the concept of non-Gaussian signal processing, two kinds of statistical methods of high-order statistics and fracture low-order statistics in a feature extracting method are combined, vibration signal features can be more comprehensively extracted, and the problem of performance degradation of a system under a non-Gaussian condition in traditional second-order statistics based methods is solved.

The invention discloses a method for nondestructive measurement of the content of free hemoglobin in a blood bag. The method comprises the following steps that a modulation apparatus modulates a light source; a displacement platform controls the light source to move to dual-light-path locations of a plurality of positions for transmission and excitation of blood separately; a spectrum receiving apparatus acquires the dual-light-path transmission and fluorescence spectra of each position; the dual-light-path transmission and fluorescence spectra are converted to frequency domains for construction of within-frequency-domain transmission spectra and within-frequency-domain fluorescence spectra; the logarithm of a ratio of transmission spectra in two frequency domains at each position is calculated so as to obtain the absorption spectrum of the blood at the position, the absorption spectra at the plurality of positions and the fluorescence spectra of the plurality of positions in dual-light-path frequency domains are subjected to normalization together, and then a mathematic model is established on the basis of the normalization and chemical examination data; and the absorption spectra and within-frequency-domain of an unknown complex solution at a plurality of positions are acquired by using same methods mentioned above, then subjected to normalization and finally substituted into the mathematic model so as to obtain the content of free hemoglobin.

The invention discloses a controller simulating a driver and an automobile simulation system. The controller comprises an accelerating pedal control module, a brake pedal control module and a control switching module, wherein the accelerating pedal control module is used for obtaining and outputting a pedal control signal according to the automobile speed deviation and the required acceleration, and the automobile speed deviation is the difference between the required automobile speed and the current automobile speed of the driver; the brake pedal control module is used for obtaining and outputting a brake pedal control signal according to the automobile speed deviation and the required acceleration; and the control switching module is used for switching and calling the accelerating pedal control module or the brake pedal control module to run according to the automobile speed deviation and the required acceleration. According to the controller simulating the driver, the behavioral intention of the driver can be effectively simulated, the controller can be applied to the automobile simulation system for forward simulating, and the simulation effect of the automobile simulation system can be improved.

The present invention discloses a method for on-line predicting the static voltage stability margin of a parallel sequential limit learning machine. By using a parallel structure to train the sequential limit learning machine parallelly, the purposes of rapidly finding out the optimal hidden layer node number satisfying the precision requirement, the historical data length N0 and the Block value of the numbers of the data blocks newly added every time are achieved. The method of the invention can accurately predict the static voltage stability of the power network, and provides the basis for the operation and planning of the power system.

The invention provides a multifunctional converter based on dual-DSP. The multifunctional converter comprises a main power circuit and a control circuit. The main power circuit comprises a three-phaseuncontrollable rectifier, a power-on flexible charging loop, a filter circuit, an H-bridge inverter circuit, a three-phase full-bridge inverter circuit and a drive circuit; the control circuit comprises a CPU minimum system, an analog acquisition and conditioning circuit, a discrete magnitude input and output circuit and a position detection and decoding circuit. As for characteristics of a three-level type synchronous motor, an exciter in a starter generator and a main electricity generator can be powered to achieve the starting function of an aero-engine, and meanwhile, the device can powerthe permanent-magnet synchronous motor to achieve the electric drive function of the actuation system on the motor; partial sharing of the main power circuit is achieved, the control circuit partially adopts the dual-DSP and CPLD configured CPU minimum system, according to different instructions, different PWM pulse signals are generated to the drive circuit, and then the inverter circuits are driven to output and power different motors.

The invention discloses a high-linearity modulation chip and method based on a dual-output silicon-based series push-pull Mach-Zehnder modulator. According to the invention, a dual-output silicon-based carrier depletion mode modulator is designed, and the silicon-based modulator is arranged at a minimum working point. The optical power input by a laser is divided into two paths through an on-chip optical power divider, one path of optical signal serves as an optical carrier and is modulated by an electric signal through a silicon-based modulator, and the other path of optical signal serves as an optical local oscillation signal and is used for compensating and suppressing modulation nonlinearity in the modulated optical signal. Finally, two paths of complementary output optical signals and optical local oscillation signals of the silicon-based modulator are combined through two silicon-based multimode interferometers, so that the third-order electro-optical nonlinearity of the modulator is suppressed. Meanwhile, a 3dB optical power divider and a silicon optical delay waveguide with a specific length are integrated on a chip, and a receiving end can suppress second-order nonlinearity with a specific frequency by using a balance detector, so that high-linearity modulation of the silicon-based modulator is realized.

A general mode control system of space dividing tower comprises an on-site intelligent device connected with the space dividing tower, a data storage device for storing history data, and a superior machine, wherein the intelligent device, data storage device and superior are serially connected, the superior machine is general mode controller comprising a component controller and a general mode controller. The component controller comprises a checking module, an I / O module, and a component analyzing module. The general mode controller comprises an I / O module, a secondary curvature simulating module, a general mode controlling calculating module and a control output module. The invention also provides a relative control method. The invention can be applicable for the dynamic, non-liner and couple properties of space dividing operation, to obtain better control effect.