53results about How to "Forecast stability" patented technology

The invention discloses an online transient stability analysis method based on a concise expression form of electromagnetic power of a single generator in a multi-machine power system. Electromagnetic output power of any single generator in the power system is a multivariable nonlinear function of all generator load angles or a multivariable nonlinear function of a slip between all generator load angles and an inductor rotor, which is known by analyzing the traditional power system. A concise expression form of the electromagnetic output power of any single generator in the complex power system is obtained through expanding a Taylor series; a load angle swinging curve of a multi-machine power system and an electromagnetic power curve are obtained through time domain simulation, and a coefficient in a high-order polynomial function is obtained by using data fitting; the load angle curve obtained through the fitting is extended to obtain a predictive load angle curve of a future moment, and the stability margin of the single generator in the multi-machine system and the system can be obtained by using the obtained concise expression form of the single generator load angle curve. The invention can be used for the online transient stability analysis of the large-scale power system.

The invention belongs to the field of thermal automatic control, and discloses a power station boiler SCR denitration modeling method based on a selective integration model library. The method comprises the following steps: (a) acquiring parameter values of an inlet and an outlet of a power station SCR reactor for multiple days, dividing the parameter values into a training set and a test set, selecting a learner, and training by using data of the training set to obtain a plurality of models to form a model library; (b) predicting a predicted value at each moment in one day by using the modelin the model library; and (c) t = t + 1, returning to the step (b), after prediction of all moments of one day is completed, updating the model library, taking the updated model library as the currentmodel library, and returning to the step (b) until updating of the model library of the last day in the test set is completed, thereby obtaining the final required model library, namely completing modeling of power station SCR denitration. Through the method, the problems of data screening and parameter adjustment caused by model updating are greatly reduced, the manual action is reduced, and theintelligent degree is higher.

The present invention provides a method and device for predicting business traffic, comprising: S100 acquiring user consumption data and feature attribute information affecting customer traffic; S200 constructing a decision tree model respectively according to the user consumption data and the feature attribute information And the target ARIMA model to predict the future traffic of each merchant. The invention organically combines the characteristics and advantages of the decision tree model and the ARIMA model, provides a scientific and accurate model selection method, and more accurately predicts the customer flow of merchants.

The invention relates to a method for obtaining a prediction model of knowledge points of text-type education resources. The method comprises the following steps: on the basis of a big-data analysis method, designing a wholly-new characteristic engineering; after collecting the text-type education resources with enough quantity, analyzing the contents and the related knowledge points of the text-type education resources, forming a learning model, optimizing gradually and perfecting prediction; newly defining the selecting and optimizing process of the characteristic, and integrating transformation between specific formulas and related contents, thereby obtaining the prediction model of the knowledge points of the text-type education resources. The invention also discloses an application method for the prediction model of the knowledge points of the text-type education resources. On the basis of the prediction model of the knowledge points of the text-type education resources designed by the invention, the prediction of the knowledge points to which to-be-predicted text-type education resources belong can be effectively carried out by the contents of the text-type education resources, the prediction process is stable and effective and the accuracy of the prediction result is high.

The invention provides an island microgrid distributed coordination control method considering a communication problem. Based on distributed information and in consideration of the actual communication problem encountered in the information transmission process, the stability control problem of the frequency and the voltage of the AC microgrid under off-network autonomy is effectively solved, andthe system instability caused by excessive fluctuation of centralized communication pressure and frequency voltage and communication interruption is avoided. A droop controller is designed, and the dynamic performance of the frequency and the voltage is optimized. In secondary control, a centralized event trigger consistency protocol is introduced, communication resources are saved, irregular switching of a network topology structure is considered, a data compensation mechanism is designed under a communication interruption fault, and a scheme of combining neighbor data compensation and centralized data compensation is adopted; and on the aspect of data prediction, an improved exponential smoothing model based on a time sequence is provided, and finally secondary adjustment of the frequency and the voltage is achieved.

The invention belongs to the technical field of electric welding quality control, and particularly relates to a resistance spot welding quality prediction method based on ensemble learning, which comprises the following steps: collecting welding process data of a welding spot sample according to process parameters measured by a sensor in a welding process; constructing a database; preprocessing the input data set through features; establishing an ensemble learning model for solder joint governance prediction, wherein each classifier outputs a quality prediction result of a to-be-detected sample; and according to output results of different classifiers and a voting mode, integrating results of the different classifiers for predicting the quality of the welding spot sample to be tested, andtaking most of judgment results as final prediction output. According to the method, the problems of high loss and low efficiency in the traditional welding spot quality detection process can be effectively solved, the welding spot quality can be quickly and accurately identified and predicted based on the welding process parameters, the welding spot quality analysis efficiency of electronic components is greatly improved, and the production cost is saved. The method is used for predicting the resistance spot welding quality.

The invention belongs to the technical field of early warning of lane changing of straight roads, and discloses a prediction system and a prediction method for line cross moment in a lane changing process of a straight road. The prediction system for the line cross moment in the lane changing process of the straight road comprises a vehicle-mounted CAN (controller area network) bus, a data processing unit, a vehicle speed sensor, a visual sensor and a gyroscope, wherein the vehicle speed sensor is arranged on a vehicle transmission; the visual sensor is used for measuring the distance between a vehicle and a lane line; the gyroscope is arranged in the center of a chassis of the vehicle; the visual sensor is arranged at the front end of the outside of the vehicle; the output end of the vehicle speed sensor is electrically connected with the vehicle-mounted CAN bus; and a signal input end of the data processing unit is respectively electrically connected with a signal output end of the gyroscope, a signal output end of the visual sensor and the vehicle-mounted CAN bus.

The invention provides a road traffic flow condition prediction method in a data sparse time period. When traffic flow data are insufficient, a transformation rule of a to-be-predicted road traffic flow in a time sequence is explored by using a time dynamic sequence supplementing method; a road traffic condition of a data sparse time period is restored by environment information feature extractionbased on a conditional random domain, so that a time evolution sequence of the road traffic flow in a period of time is obtained; matching with a historical time sequence of the road traffic flow isperformed to find out a time sequence fragment with a similar evolution trend and a traffic flow condition of a prediction time point is deduced. The test and on-site detection show that the prediction result of the traffic flow data sparse time period is basically accurate and reliable, so that defects of road traffic flow condition prediction in the data sparse time period in the prior art are effectively overcome, the traffic flow condition prediction weakness is eliminated, and the overall quality of the road traffic flow condition prediction method is improved.

The invention discloses a bridge temperature prediction method based on a long-term and short-term memory network, a medium and equipment. The method comprises the steps of dividing a bridge temperature data set into a training set, a verification set and a test set in proportion; constructing a network model based on Keras, and training the network model by using the training set and the verification set to obtain a prediction model; and sending the test set data into a prediction model, and carrying out early warning processing when an abnormal value is found in a prediction window. According to the method, the time window is divided, sufficient early warning time is provided to solve the discovered problem, and the neural network prediction model is more accurate and stable in peak prediction.

The invention discloses a surrounding rock deformation monitoring method and prediction method suitable for a double-shield TBM. The monitoring method mainly comprises three parts of advanced borehole measuring tube design and deformation monitoring based on a quasi-distributed FBG sensing principle, monitoring value correction based on construction process numerical simulation and inversion analysis, and surrounding rock convergence deformation numerical simulation analysis and advanced prediction based on monitoring inversion parameters. The monitoring method and the prediction method have the advantages that the quasi-distributed FBG advanced drilling measurement pipe is utilized, the problem that surrounding rock deformation of a double-shield TBM tunnel is difficult to monitor due to shield and segment shielding is effectively solved, displacement correction is conducted on an imaginary fixed point of the measurement pipe through refined numerical simulation, the creep constitutive parameters of surrounding rock are inverted according to monitoring data, and advanced prediction analysis of the convergence deformation of the surrounding rock is realized based on the numerical model, so that a guidance basis is provided for construction tunneling and support design of a tunnel, and safe and efficient construction of the double-shield TBM is ensured.

The invention relates to a method for determining a rack force (Fr, estcomplrack) for a steer-by-wire steering system (1) for a motor vehicle, wherein the rack force (Fr, estcomplrack) is determined from two components, wherein a first component of the rack force (Fr, estvehicle) is generated in a module for the vehicle model-based estimation of the rack force (15) by means of a vehicle model, anda second component of the rack force (Fr, estrack) is generated in a module for the steering-mechanism-model-based estimation of the rack force (16) by means of a steering-mechanism model.

The invention discloses a product regional pricing method based on a gray neural network model. The method comprises the steps of obtaining consumer price index data over the years and product regional price indexes over the years; performing interval average processing on the product regional price indexes over the years to obtain an average product price index; performing the best translationaltransformation on the consumer price index data over the years and data fitting with a continuous gray model to construct a continuous consumer price index GM (1,1) model over the years; obtaining theconsumer price index training data, performing network training on the consumer price index training data through the continuous consumer price index GM(1,1) model over the years and the average product price index and establishing the neural network model; using the continuous consumer price index GM(1,1) model over the years for prediction to obtain predicted consumer price indexes in the nextyear, substituting the predicted consumer price indexes in the next year into the neural network model and performing calculation to obtain product regional price indexes in the next year.

The embodiment of the invention relates to the technical field of big data, and particularly discloses a link prediction method and device and a storage medium, and the method comprises the steps: obtaining any node pair to be subjected to link prediction in a specified network, and taking the node pair as a target node pair; extracting common neighbor sharing influence, common neighbor dispersion influence and inter-node influence of the target node pair, wherein the common neighbor sharing influence is represented by the degree of the common neighbor node of the target node pair, the common neighbor dispersion influence is represented by the shortest path number of the common neighbor nodes of the target node pair, and the inter-node influence is represented by the similarity of common neighbor nodes of the target node pair; and performing link prediction on the target node pair according to the common neighbor sharing influence, the common neighbor dispersion influence and the inter-node influence of the target node pair so as to determine the link relationship between the node pairs in the specified network based on a link prediction result, thereby improving the accuracy and stability of network link prediction.

The invention discloses a traffic flow prediction method based on a feature reconstruction error, and belongs to the technical field of machine learning. The method comprises the steps of: (1) selecting a target machine learning network, and initializing the parameters of the target machine learning network; (2) constructing a training data set of traffic flow, and initializing parameters of a feature correction weight matrix; (3) training the feature correction weight matrix by using the training data set, and using a random gradient descent algorithm and a feature reconstruction error loss function in the training process; (4) fixing feature correction weight matrix parameters, training a target machine learning network, and using a random gradient descent algorithm in the training process; (5) repeating the step (3) and the step (4) until the loss function converges or reaches a maximum training step number; and (6) after the training is finished, inputting the traffic flow data tobe predicted into the trained network model to obtain the predicted traffic flow. According to the invention, the stability of the model in traffic flow prediction can be enhanced.

The invention provides a model fusion method and device, a prediction method and device, equipment and a storage medium, and relates to the technical field of data processing. The method comprises thesteps of obtaining a plurality of prediction models; predicting target sample data by adopting each prediction model to obtain a sample prediction value of each prediction model; determining a prediction error of each prediction model according to the sample prediction value of each prediction model and a standard prediction value corresponding to the target sample data; and calculating the weight of each prediction model according to the prediction error of each prediction model and the prediction errors of the plurality of prediction models, wherein the weights of the plurality of prediction models are prediction weights of the plurality of prediction models for the input data in the process of predicting the input data in the hybrid application scene. Therefore, the prediction based onplurality of prediction models is more stable, the prediction result can be output in combination with the weight of each prediction model, the processing effect on the input data is improved, and the prediction result is more accurate.

The invention discloses a welding structure fatigue performance analysis method based on a data driving method, and relates to the field of welding structure fatigue performance analys.The welding structure fatigue performance analysis method comprises the following steps that fatigue performance data are obtained, a multi-scale fatigue performance database is established, and the data in the database is divided into training data and test data; analyzing a linear correlation degree and a non-linear correlation degree between the fatigue performance and the influence factors by using a Pearson correlation coefficient and a maximum information coefficient respectively; based on an optimized gradient lifting algorithm, carrying out quantitative analysis on the weights of the fatigue performance influence factors; training a deep convolutional neural network by using the training data; inputting the fatigue performance influence factors into the trained convolutional neural network to obtain a prediction parameter curve; and extracting the fatigue life according to the predicted parameter curve. According to the invention, accurate and stable fatigue life prediction of the complex welding structural member under different materials, shapes, sizes, processing technologies and service conditions can be realized.