39 results about "Polynomial regression" patented technology

The invention provides a transformer anti-short circuit capability quantitative evaluation method and system based on multi-physics field simulation mapping, and belongs to the technical field of transformer anti-short circuit capability evaluation. Comprising the following steps: obtaining structure parameters of a to-be-tested type transformer, and constructing a multi-physics field coupling simulation model; running a preset defect working condition by using the multi-physics field coupling simulation model, and calibrating an anti-short-circuit capability limit value corresponding to the defect working condition according to the running data; applying a unit pulse voltage excitation signal, and obtaining a feature vector corresponding to the defect working condition based on the frequency domain response curve; analyzing the mapping relation by using a polynomial regression algorithm to obtain an anti-short-circuit capability quantitative evaluation function; and in a power failure off-line state of the transformer, injecting a unit pulse voltage excitation signal into the winding by using a pulse frequency response tester, calculating an anti-short-circuit capability limit value corresponding to an actually measured operating temperature value according to the anti-short-circuit capability quantitative evaluation function, and evaluating the anti-short-circuit capability of the transformer in combination with a rated thermally stable current.

A computer device includes at least one processor in communication with at least one memory device. The at least one processor is programmed to: a) receive, from the image capture device, a plurality of images for a continuous scan phase shift interferometry (PSI), wherein each image of the plurality of images includes a first plurality of pixels of a first item and a second plurality of pixels of a partial cavity; b) perform intensity scans of each pixel in the second plurality of pixels; c) identify zero transitions for each of the second plurality of pixels; and d) statistically evaluate the identified zero transitions by polynomial regression analysis.

The invention relates to the technical field of power system dynamic safety analysis, and discloses a mechanism-data fusion new energy power system minimum inertia demand rapid evaluation method, which comprises the following steps: in an offline stage, establishing an expansion system frequency response model containing new energy; generating a plurality of sample scenes in a given disturbance power range and a new energy permeability range; a model minimum inertia demand and a simulation minimum inertia demand are obtained through expansion of a system frequency response model and stepping type time domain simulation calculation, a polynomial regression method and a polynomial regression equation are utilized, and the coefficient and order of the polynomial regression equation are determined by adopting an optimization algorithm; during online evaluation, real-time disturbance power and new energy permeability of a power grid are obtained; and obtaining a model minimum inertia demand and a corresponding minimum inertia demand deviation compensation value, and outputting a final system minimum inertia demand evaluation result. According to the method, the final evaluation result has the calculation efficiency of the extended SFR model and the evaluation precision of stepping simulation.

The application relates to the technical field of fire rescue equipment monitoring, and discloses a firefighter air respirator residual amount prediction method and system based on dynamic pressure compensation. The method comprises the following steps: acquiring historical consumption time data of firefighters, using polynomial regression fitting to establish a theoretical pressure consumption rate function corresponding to an individual and establishing a theoretical prediction logic; receiving a combat organization instruction to obtain an initial full load pressure, starting timing and calculating and displaying a current theoretical prediction residual pressure; capturing an artificial interactive correction instruction, extracting a correction time node and an actual residual pressure, and solving to generate a dynamic compensation coefficient; applying the dynamic compensation coefficient to reconstruct and solve an actual prediction residual pressure and control and update display; comparing the actual prediction residual pressure with an alarm critical pressure value to trigger an alarm, and extracting actual combat data to iteratively update the theoretical function after the task is completed. The application realizes individualized benchmark prediction of air consumption and dynamic adaptive error correction of the working environment, and improves residual amount early warning accuracy.

The application discloses an online roll temperature and thermal crown prediction method based on finite difference and polynomial regression, aiming at the technical pain point that the accuracy of the traditional analytical model is insufficient and numerical simulation cannot meet the online real-time calculation demand, a plurality of groups of working conditions are designed through an orthogonal experiment method, and a finite difference method is used to calculate roll steady-state temperature field data to construct a training data set, based on the data set, a polynomial regression algorithm with elastic network regularization is used to fit an axial temperature distribution general model with the roll temperature field as an input variable, after obtaining the online surface axial temperature distribution of the roll, the general model is substituted into the internal temperature field to restore the internal temperature field and calculate the radial thermal expansion amount at each position, so that compensation information is provided for a shape control system of a rolling mill. The application simplifies complex physical field simulation into lightweight algebraic formula operation, retains the high analytical accuracy of the finite difference method, realizes real-time prediction, and can significantly improve the rolling quality of silicon steel products.

The present application relates to the field of strain detection, and particularly to a self-adaptive strain detection method, system and medium. The method comprises: inputting grid point coordinates, noisy displacement data and a maximum smoothing half window; calculating an overall noise level sigma through local linear regression analysis; extracting a local strain curvature characteristic quantity through cubic polynomial regression; establishing a noise and curvature balance relationship based on sigma and the curvature characteristic quantity, and calculating an adaptive smoothing half window; obtaining strain estimation values of each grid point through linear polynomial regression, and finally obtaining overall strain distribution. The method can automatically adjust the optimal smoothing window without manual intervention, balances between noise suppression and detail preservation, solves the problem of insufficient accuracy of traditional fixed window methods in non-uniform deformation areas, realizes high-precision full-field strain detection of noisy displacement fields, and is suitable for derivative calculation of digital image correlation systems and various noisy continuous signals.

The invention belongs to the technical field of simulation optimization, and particularly relates to a centrifugal pump impeller multi-physics field coupling simulation optimization method which comprises the following steps: constructing a geometric model of a centrifugal pump by adopting three-dimensional drawing software according to an actual structure of the centrifugal pump; finite element modeling is conducted, and a finite element analysis model of the centrifugal pump is obtained; hexahedral grids are divided for the impeller, a shape variable space of the impeller is built, and the impeller deformation serves as a design variable; solving the finite element analysis model after grid division through a solver; an experimental design sample is generated by adopting a Latin hypercube sampling method, a global response surface model is constructed through quadratic polynomial regression fitting, an optimal impeller deformation amount is iteratively solved by adopting a global response surface optimization algorithm, a finite element analysis model is adjusted, and an optimization model is formed. According to the method, fluid-solid coupling simulation, a global response surface method and an automatic iteration process are fused, and the efficiency of the centrifugal pump impeller is maximized and optimized on the basis of accurately matching actual working conditions.

The embodiment of the invention discloses an on-line detection method and device for the loss of a distribution transformer, equipment and a storage medium. The method comprises the steps that the current load rate and the current load current of the transformer are acquired; determining a target order total loss model according to the current load rate and a preset corresponding relation between the load rate and the order of the total loss model; and determining the current total loss of the transformer by using the current load current and the target order total loss model. The polynomial function relation between the loss and the load current is obtained through theoretical derivation, and the loss of the transformer is calculated through the polynomial regression algorithm. In order to compensate the influence of the load on loss calculation, a polynomial dynamic order adjustment mechanism is introduced, and the loss calculation accuracy is further improved. Meanwhile, on-line loss detection only needs to detect primary and secondary side voltage and current effective values during normal operation of the transformer through a sensor, power failure is not needed, power supply reliability is greatly improved, and waste of manpower and material resources is avoided.

The present application provides a kind of carbon emission trend prediction method based on polynomial regression and neural network, comprising: S11, respectively constructs the carbon emission model and flow chart model of whole life cycle;S12, data cleaning and pretreatment are carried out to carbon emission model and flow chart model, and high-quality data are output;S13, high-quality data are updated by dynamic updating mechanism;S14, a carbon emission prediction model is established using polynomial regression analysis method, carbon emission data are input into the carbon emission prediction model, and carbon emission trend prediction result is output.The present application can improve the accuracy and timeliness of high-carbon emission data, shorten the evaluation period, reduce the cost, and provide a strong basis for environmental regulation and protection decision.

The invention discloses a powder grinding multi-target cooperative control method and system, and belongs to the field of industrial process automation control, and the method comprises the steps: extracting powder concentrator load and mill main motor power time sequence data, separating internal model components based on variational mode decomposition, and combining with residence time distribution to generate a working condition quality mapping matrix; performing least square polynomial regression and numerical control comparison on the working condition quality mapping matrix to construct candidate control parameters; calculating an internal model component instantaneous frequency variance to construct a dynamic weight, selecting a candidate vector with the minimum weighted sum, and executing amplitude truncation to generate a control instruction; collecting new data to calculate a prediction residual vector, and executing least square iteration update to generate a correction regression equation; and eliminating abnormal lines of the working condition quality mapping matrix by utilizing a quantile boundary on the prediction residual error sequence, adding new data and executing secondary calibration. According to the method, data-driven closed-loop control is adopted, and the adaptive capacity and control precision of the model to nonlinear working conditions can be improved.

The invention discloses a quick evaluation method and system for the crispness of sauced cucumbers based on hyperspectral imaging, and the method comprises the steps: describing the sensory crispness score of to-be-detected sauced cucumbers, and collecting the texture indexes of the to-be-detected sauced cucumbers; modeling indexes are determined as hardness and elasticity through correlation analysis, and a quadratic polynomial regression equation among the brittleness, the hardness and the elasticity is established; for each index, acquiring spectral data of the sauced cucumber to be detected through a hyperspectral imaging method, performing preprocessing and characteristic wavelength screening on the spectral data, and then establishing a machine learning prediction model; and according to a hardness prediction value and an elasticity prediction value obtained by prediction of the machine learning model, inputting the hardness prediction value and the elasticity prediction value into the quadratic polynomial regression equation to obtain a brittleness evaluation result. Compared with a traditional sensory evaluation method, an instrument analysis method and other existing methods, the method has the advantages of high efficiency, safety, no damage, cost saving and the like.

The invention discloses a method for forecasting the ultimate strength of a multi-layer egg-shaped pressure-resistant shell. The method comprises the following steps: multiplying a plastic attenuation factor, a defect attenuation factor and a linear buckling formula to establish an ultimate strength forecasting formula of the multi-layer egg-shaped pressure-resistant shell; determining the relationship between the ultimate strength of the multi-layer egg-shaped pressure-resistant shell and the linear buckling load, the plastic attenuation factor and the defect attenuation factor; determining influence factors and value ranges of the plastic attenuation factor and the defect attenuation factor, and designing a response surface test; establishing a three-dimensional finite element model of the multi-layer egg-shaped pressure-resistant shell; fitting based on finite element numerical simulation data to obtain a polynomial regression equation of a plastic attenuation factor and a defect attenuation factor; designing an orthogonal test according to the influence factors of the plastic attenuation factor and the defect attenuation factor, and verifying an ultimate strength forecasting formula; and calculating the ultimate strength of the multi-layer egg-shaped pressure-resistant shell according to actually measured geometry, materials and defect parameters. According to the method, the ultimate strength of the multi-layer egg-shaped pressure-resistant shell can be efficiently and accurately predicted.

The application discloses a kind of powder grinding multi-objective collaborative control method and system, belong to industrial process automation control field, it includes extracting powder concentrator load and mill main motor power time series data, based on variation mode decomposition separates internal mode component and generates working condition quality mapping matrix in combination with residence time distribution;Least square polynomial regression and numerical domination comparison are executed to working condition quality mapping matrix to construct candidate control parameter;The instantaneous frequency variance of internal mode component is calculated to construct dynamic weight, select the candidate vector of minimum weighted sum and execute amplitude truncation to generate control instruction;New data is collected to calculate prediction residual vector, and least square iteration update is executed to generate revised regression equation;The abnormal row of working condition quality mapping matrix is removed using the upper quantile boundary of prediction residual sequence, new data is added and secondary calibration is executed.The application uses data-driven closed-loop control, and the adaptability and control precision of model to nonlinear working condition can be improved.

The invention relates to a vision measurement technology, and discloses a basketball projection detection method, equipment system and storage medium, and the method comprises the steps: obtaining an RGB video stream and a depth map in a process that a basketball flies towards a basket based on a depth camera; carrying out basketball target detection according to the RGB video stream, and aligning the depth map with a corresponding image frame in the RGB video stream; based on the depth map, obtaining point cloud data of a basketball target in each image frame; fitting the point cloud data of the basketball target in each image frame to generate a three-dimensional model of a sphere, and calculating the coordinates of the center of sphere; fitting the center coordinates of the continuous image frames to generate a basketball flight path; and based on the basketball flight path after coordinate conversion, combining polynomial regression and extended Kalman filtering to predict a projection result. According to the method, the falling point coordinates and the incident angle of the basketball are accurately calculated, more detailed and accurate shooting result information is provided, and the requirement for accurate analysis of basketball shooting is met.

This invention relates to the field of short-term meteorological climate prediction, specifically disclosing a seasonal adaptive integration and dynamic anomaly correction temperature prediction system and method. The system includes a data acquisition module, a modeling and calculation module, and an evaluation and visualization module. The method includes: S1, fusing multi-source data to construct time-coded, lag, and cross-feature features, and generating a dynamic climate benchmark with variable weights; S2, based on the Stacking integration framework, using Ridge+LightGBM in winter, SVR and multinomial regression in summer, weighting during the transition season, and residual calibration in winter; S3, three-level anomaly evaluation, combined with benchmark calculation and visualization, with fine-tuning when the matching rate is low. The seasonal adaptive integration and dynamic anomaly correction temperature prediction system and method proposed in this invention solves the problems of difficulty in nonlinear capture, benchmark rigidity, and poor seasonal adaptation, effectively improving the accuracy of seasonal temperature prediction.

This invention provides an optimization method for the mathematical model of industrial film characteristic curves, belonging to the field of mathematical modeling technology for film characteristic curves. It solves the problems of slow speed, low accuracy, poor repeatability, and difficulty in querying film characteristic parameters in industrial X-ray inspection, which relies on manual table lookup and calculation. This results in long testing cycles and low work efficiency. The method includes the following steps: S1: Collect and organize data corresponding to exposure and density of characteristic curves for different types of industrial films; S2: Analyze the characteristics of the industrial film characteristic curves and adopt... n The mathematical model of the industrial film characteristic curve is obtained by fitting the polynomial regression method; S3: The parameters in the polynomial of the industrial film characteristic curve are adjusted by the least squares method and regression analysis to obtain the optimal mathematical model of the industrial film characteristic curve; This invention is applied to the mathematical modeling of film characteristic curves.

The invention provides a BMS battery fault prediction method and system, and belongs to the technical field of fault prediction and health management. The method comprises the following steps: acquiring battery operation condition data, and dynamically adjusting a parameter weight vector to enable the weight to adapt to different operation scenes; a nonlinear relation mapping method is adopted to determine an initial predicted value of a health state parameter, a nonlinear change boundary of the parameter is effectively adapted, and prediction deviation of a risk critical region is avoided; capturing a nonlinear fluctuation rule of the initial predicted value through a quadratic polynomial regression algorithm, and constructing a fluctuation boundary feature sequence; the prediction step length is adjusted in real time by means of an accumulation and control chart algorithm, accurate matching of the step length and the actual risk level is achieved, missing of the intervention opportunity when the risk rises is avoided, redundant calculation when the risk is stable is reduced, and the problem that the step length and the risk are disjointed is solved. Finally, the fault probability is determined based on the adjusted predicted value sequence, timely and accurate early warning is guaranteed, serious consequences such as thermal runaway are avoided, and the reliability and accuracy of battery fault prediction are greatly improved.

The application discloses a mathematical model for identifying the age of dried orange peel, and a regression equation is obtained by taking X as the horizontal coordinate, Y as the vertical coordinate and fitting a polynomial regression curve, so that the mathematical model is constructed, wherein X is the mass fraction of esterified phenolic acid in total phenolic acid in dried orange peel of each year, and Y is the age of dried orange peel. The application further discloses a method for establishing the mathematical model for identifying the age of dried orange peel. The application further discloses a method for identifying the age of dried orange peel. The mass fraction of esterified phenolic acid in total phenolic acid is used as the independent variable, which conforms to the change rule of the existing form of phenolic acid in the aging process of dried orange peel, and can reduce the deviation caused by individual differences of samples and improve the accuracy of the identification of the age of dried orange peel.

This invention belongs to the field of numerical analysis technology, specifically relating to a method and system for polynomial regression analysis of discrete data. The method includes the following steps: Step S1, data loading and matrix construction; Step S2, polynomial feature expansion; Step S3, order and coefficient initialization; Step S4, matrix solving; Step S5, residual evaluation and order optimization; Step S6, optimal model determination. This technical solution overcomes the shortcomings of existing technologies in terms of efficiency, accuracy, and intelligence by combining the nonlinear fitting capability of polynomials, the optimal solution of matrix solving, and RSS-driven adaptive model selection. It provides an efficient, high-precision, and adaptive discrete data analysis solution, which is of great significance for promoting the discrete data analysis and intelligent upgrading of industrial equipment.

The invention provides a method and a system for detecting abnormal vibration of a nuclear power cold source pump. The method comprises the following steps: acquiring an original vibration signal; in the start-stop stage of the cold source pump, the original vibration signal is processed through a Savitzky-Golay smooth function and dynamic polynomial regression, and a trend prediction value is obtained; calculating a fitting residual sequence according to the trend prediction value and the original vibration signal; calculating a dynamic standard deviation and a corresponding adaptive confidence interval; performing anomaly detection according to the adaptive confidence interval; after the cold source pump finishes the start-stop process and enters a relatively stable operation stage, performing dynamic weighted fusion on the predicted value of the time sequence feature, the predicted value of the multi-source collaborative feature and the theoretical predicted value to obtain a final predicted value; calculating a residual error according to the original vibration signal and the final predicted value; performing anomaly detection according to the dynamic threshold interval to obtain an anomaly judgment result; and real-time abnormity identification and intelligent early warning of the cold source pump in the start-stop and stable operation processes are realized.

The present application relates to the field of data processing, and particularly relates to a kind of intelligent deviation rectification method and system for strip cutting, first by real-time acquisition and smooth processing deviation profile data of strip;Then, by calculating the local variation energy of the second derivative of deviation profile, the region where the morphology changes significantly is identified, and its characteristic scale is further calculated;Then, according to the characteristic scale of macro deviation, an adaptive analysis window is dynamically generated;Finally, in the window, using low-order polynomial regression, the complex deviation form is accurately decoupled into basic parameters such as translation, tilt, C-shaped bending and S-shaped bending, and is transmitted to the controller to realize fast and accurate differentiated deviation rectification.The present application identifies macro deviation by calculating local variation energy and characteristic scale, constructs adaptive analysis window and decouples deviation form using local polynomial regression, to realize accurate control.

The present application relates to a system for adaptively determining a read threshold voltage using meta information. Embodiments of the present application adaptively determine a read retry threshold voltage for a next read operation using meta information collected from a previous failed read data. A controller obtains meta information associated with a read operation on a selected page, the meta information including a set of read threshold voltages. The controller determines a mathematical model for estimating a checksum value associated with data of a next read operation using a set function of the set of read threshold voltages and a set checksum value. The controller determines a set of parameters by performing a polynomial regression on the mathematical model. The controller estimates a next read threshold voltage for the next read operation based on the set of parameters.

The invention provides a semiconductor parametric modeling method based on a polynomial regression algorithm, belongs to the technical field of semiconductor modeling, and aims at solving the problems that traditional TCAD fitting calculation amount is large and the requirement for machine learning data is high, the method utilizes a small amount of TCAD simulation data to construct an agent model, and global rapid optimization is achieved in combination with PDK exact parameters and empirical values. Taking a FinFET device as an example, the scheme not only realizes high-precision fitting consistent with actual measurement data in a wide temperature range from-40 DEG C to 120 DEG C, but also has excellent extremely low temperature characteristic prediction capability. According to the method, through cooperation of agent model optimization and parameter sensitivity analysis, the efficiency and precision of semiconductor device modeling are effectively improved while the computing power cost and the sample collection workload are remarkably reduced.

The invention provides a lane detection method and system based on a millimeter wave radar, and the method comprises the steps: S1, carrying out the statistics of a track transverse distribution histogram, detecting a peak value, and determining the number of lanes and the initial position of a center; s2, fitting the track point set to generate a global center curve, and translating to obtain an initial center line of each lane; s3, calculating a subsequence dynamic time warping distance between the trajectory and each center line, distributing the trajectory to the lane clusters according to the minimum distance, carrying out joint fitting on each cluster by using a random sampling consensus algorithm and polynomial regression to update the center line, carrying out iteration until convergence, and outputting a final center line set; and S4, constructing a common reference center line based on the final center line set, calculating the normal global offset of each center line relative to the reference line so as to determine the lane boundary offset, and translating along the normal direction of the reference line to generate a parallel and smooth lane boundary. According to the invention, all-weather automatic lane detection independent of vision and manual marking is realized.

The present application is a method for predicting the time series of mycelium concentration, ethanol concentration and glucose concentration in the fermentation process of fuel ethanol. The method constructs a fermentation end time prediction model based on a polynomial regression algorithm, and establishes a mathematical relationship between the fermentation end time and the initial conditions. Through the scaling of time scale, the time series of different batch data is transformed to achieve a unified standard end time. The method establishes a mycelium concentration prediction model based on the XGBoost algorithm. The obtained mycelium concentration and growth rate are added as supplementary mechanism knowledge to the prediction model of ethanol and glucose concentration. The method establishes a dynamic parameter model of artificial neural network, which can provide dynamic parameters according to time variables and environmental variables. By integrating it into the previous mycelium concentration, ethanol concentration and glucose concentration prediction model, the time series prediction of mycelium concentration, ethanol concentration and glucose concentration under different initial conditions can be realized.

The invention discloses a water guide shoe temperature trend analysis and early warning method based on a multi-model dynamic threshold, and relates to the technical field of state monitoring and intelligent early warning of a hydroelectric generating set, and the method comprises the steps: recognizing the operation condition of the set according to the power, the rotating speed and a water head; gaussian filtering processing is carried out on the identified non-thermal-stability working condition data; polynomial regression fitting is carried out to generate a non-thermal-stability working condition dynamic threshold model; clustering by utilizing a Gaussian mixture model, and dividing the thermal stability working condition into a plurality of subclasses; establishing a 3 delta dynamic threshold model to generate a dynamic threshold curve of a corresponding working condition; selecting a corresponding threshold model according to an identification result, and loading an upper and lower limit threshold curve; and calculating the deviation between the real-time bearing temperature curve and the dynamic threshold, and when the real-time bearing temperature exceeds the upper limit and the lower limit of the corresponding dynamic threshold, generating a trend anomaly judgment result. According to the method, working condition identification, multi-model modeling and dynamic threshold generation mechanisms are introduced, so that the reliability and the intelligent level of hydroelectric generating set operation monitoring are improved.

The application discloses a soft rock tunnel multilayer support parameter optimization method, comprising the following steps: S1, obtaining soft rock tunnel basic information; S2, according to the soft rock tunnel basic information, combining engineering regional geological environment and topography and geomorphology, preliminarily selecting construction and support parameters; S3, according to the preliminarily selected construction and support parameters, obtaining support effect information; S4, using a Latin hypercube sampling method, sampling and combining the construction and support parameters, and extracting the support effect corresponding to the combined working condition; S5, using a quadratic polynomial fitting, performing quadratic polynomial regression analysis on the support effect; S6, using a variance analysis method, analyzing the support effect information, and obtaining the contribution rate of each construction and support parameter to the support effect; and S7, using a genetic algorithm to obtain the optimal construction and support parameter combination. According to the actual support effect of construction, the application can select ideal and economic construction and support parameters, and realizes dynamic feedback adjustment of the construction and support parameters.

The invention discloses a gas storage injection and production well casing damage prediction method and system, electronic equipment and a computer readable storage medium. The method comprises the steps that an influence factor system of casing damage is established; establishing a numerical calculation model of the casing damage layer section containing the actual measurement data of the influence factor system; comparing the real test data with the simulation test data, verifying the accuracy of the numerical calculation model, carrying out sensitivity analysis, and further determining key control factors of casing damage; according to an orthogonal experiment design method, the determined key control factors of the casing damage are experimented, a discrete relation data sequence of the key control factors and real strain in the casing damage process is formed, a quadratic polynomial regression equation of the correlation between the casing damage and the control factors is extracted by adopting a data statistical analysis method, and the casing damage is obtained. Forming a prediction function of the real strain and the critical strain of the casing; and predicting the damage state of the casing according to the prediction function. According to the method, the damage of the casing can be accurately predicted.

The invention discloses a hybrid modeling method, system and device for improving avalanche photodiode current gain prediction and a medium, and belongs to the technical field of semiconductor device simulation, the core of the hybrid modeling method is to construct a physically guided improved one-dimensional U-Net neural network, a third-order polynomial regression module is embedded at the bottleneck of the U-Net network, and the current gain prediction of the avalanche photodiode is improved. Variable parameters related to the gain are explicitly learned and input, the dependency relationship of the gain on the parameters such as the electric field and the temperature is directly modeled, and the physical rationality is ensured; meanwhile, the inherent jump connection structure of the U-Net is utilized to transmit fine features captured by the encoder to the decoder, so that high-fidelity reconstruction of key details of gain distribution is realized; the method aims at solving the core contradiction between high-precision numerical simulation and calculation efficiency in the design process. Through the design of mechanism driving, the hybrid prediction model is enabled to deeply accord with the physical nature of the avalanche photodiode, and second-level, high-fidelity and physically credible gain prediction is realized.

The invention discloses a circulating water energy-saving dispatching method and system based on model fusion, and belongs to the technical field of condensing steam unit circulating water energy-saving dispatching. According to the method, a back pressure correction model and a neural network model are integrated, a fusion model is constructed, and condenser back pressure which is used for reflecting condenser heat transfer characteristics and has high prediction precision is obtained; by means of polynomial regression fitting of condenser variable working condition characteristics, the relation between turbine slightly-increased output and back pressure and the relation between circulating water flow and circulating water pump power consumption data, optimal operation circulating water flow is obtained through optimization with the maximum turbine energy consumption net income as the target, and circulating water energy-saving dispatching is achieved. Through model fusion, the double advantages of the back pressure correction model and the neural network model are integrated, the condenser back pressure can be accurately predicted, meanwhile, the method has scientific guiding significance for unit circulating water dispatching, the cold source loss of the unit is reduced, and the thermal economic benefit of the unit is improved.