CN114971024A - Fan state prediction method and device - Google Patents

Fan state prediction method and device Download PDF

Info

Publication number
CN114971024A
CN114971024A CN202210594394.2A CN202210594394A CN114971024A CN 114971024 A CN114971024 A CN 114971024A CN 202210594394 A CN202210594394 A CN 202210594394A CN 114971024 A CN114971024 A CN 114971024A
Authority
CN
China
Prior art keywords
longicorn
value
data
parameter
fan
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202210594394.2A
Other languages
Chinese (zh)
Inventor
李治
王帝
方超
汪勇
丁刚
邓志成
孙猛
陈荣泽
谷朋泰
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shanghai Power Equipment Research Institute Co Ltd
Original Assignee
Shanghai Power Equipment Research Institute Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shanghai Power Equipment Research Institute Co Ltd filed Critical Shanghai Power Equipment Research Institute Co Ltd
Priority to CN202210594394.2A priority Critical patent/CN114971024A/en
Publication of CN114971024A publication Critical patent/CN114971024A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/20—Administration of product repair or maintenance
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Electricity, gas or water supply
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00—Systems supporting electrical power generation, transmission or distribution
    • Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications

Abstract

The invention discloses a method and a device for predicting a fan state. The fan state prediction method comprises the following steps: step S11, sample data is obtained, wherein the sample data comprises integral level data, equipment level data and component level data; step S12, taking the quantity to be predicted in the sample data as an output vector and the residual quantity as an input vector, and establishing a least square support vector machine model; wherein the kernel function of the least squares support vector machine model is a radial basis kernel function; step S13, determining a global optimization initial value of the hyper-parameter in the radial basis kernel function by using a grid search method; and step S14, determining the global optimal value of the hyper-parameter by using the global optimization initial value as the initial longicorn centroid and using a longicorn whisker algorithm. The method and the device can predict the overall running state of the fan, and improve the accuracy of prediction.

Description

Fan state prediction method and device
Technical Field
The invention relates to the technical field of fan state prediction, in particular to a fan state prediction method and device.
Background
Fans have important applications in many areas of modern society, such as thermal power plants. The fan is used as an important auxiliary machine of a thermal power plant, the running state of the fan is closely related to the safety and the economical efficiency of a generator set, but the attention degree of the fan is still low compared with that of a main machine.
At present, most scholars at home and abroad make some state prediction researches on fan equipment based on analytical models, expert systems and data driving methods, but the researches mainly focus on the vibration state and the fault of the fan, but do not predict the overall state of the fan, and are lack of coupling researches with the states of a generator set and a boiler. The state prediction of the power station fan mainly has the following difficulties that (1) the fan has multiple types, complex structure, severe operation condition and various states; (2) various measuring points of the fan are limited, so that historical operating data of the fan is incomplete; (3) the relationship among the operating parameters of the fan is complex, and the nonlinear relationship is strong.
Disclosure of Invention
The invention provides a method and a device for predicting the state of a fan, which are used for predicting the overall operation state of the fan and improving the accuracy of prediction.
According to an aspect of the present invention, a method for predicting a fan state is provided, including:
step S11, sample data is obtained, wherein the sample data comprises integral level data, equipment level data and component level data;
step S12, taking the quantity to be predicted in the sample data as an output vector and the residual quantity as an input vector, and establishing a least square support vector machine model; wherein the kernel function of the least squares support vector machine model is a radial basis kernel function;
step S13, determining a global optimization initial value of the hyper-parameter in the radial basis kernel function by using a grid search method;
and step S14, determining the global optimal value of the hyper-parameter by using the global optimization initial value as the initial longicorn centroid and using a longicorn whisker algorithm.
Optionally, the determining, by using the initial global optimization value as an initial longicorn centroid, a global optimal value of the hyper-parameter by using a longicorn whisker algorithm includes:
step S141, updating the random step length, the longicorn stigma orientation and the fitness value of the current position, wherein the longicorn stigma orientation comprises a longicorn left stigma, a longicorn right stigma and a longicorn middle stigma;
step S142, calculating the fitness value of the longicorn after the longicorn moves the random step length along the left tassel, the right tassel and the middle tassel of the longicorn to determine the next position of the longicorn;
step S143, determining whether the longicorn is moved to the next position according to the fitness value of the longicorn at the current position and the next position, if the fitness value of the longicorn at the next position is smaller than the fitness value of the longicorn at the current position, moving the longicorn to the next position, and executing step S144; if the fitness value of the longicorn at the next position is larger than or equal to the fitness value at the current position, returning to execute the step S141;
step S144, comparing the fitness value of the next position obtained in step S143 with a set minimum fitness value, and if the fitness value is smaller than the set minimum fitness value or reaches the maximum iteration number, ending the iteration to obtain a global optimal value of the hyper-parameter; if not, the step S141 is executed again.
Optionally, the updating the fitness values of the random step size, the skynet whisker orientation and the current position includes:
defining the orientation of long horns as random direction vectors
Figure BDA0003667175670000021
Figure BDA0003667175670000022
Wherein k is an optimization dimension, and rand (.) is a random function;
establishing space coordinates of the longicorn left tassel, the longicorn right tassel and the longicorn middle tassel;
Figure BDA0003667175670000023
wherein x is 0 Is the coordinate of the initial centroid, x l Is a coordinate, x, of crawling along the left beard of a longicorn r Is a coordinate, x, of crawling along the right beard of the longicorn m Coordinates for crawling along the longicorn, d 1 、d 2 、d 3 The distances from the corresponding longicorn left beard, longicorn right beard and longicorn middle beard to the mass center respectively, t is the iteration frequency,
Figure BDA0003667175670000031
respectively are coordinates of the longicorn after crawling along the left tassel, the right tassel and the middle tassel of the longicorn in the t iteration,
Figure BDA0003667175670000032
the position of the barnyard centre of mass at the t iteration and R (theta) is a random direction vector
Figure BDA0003667175670000035
The rotation matrix of (a);
taking the minimum mean square error of the predicted value and the measured value of the quantity to be predicted as a fitness function f (x);
respectively obtaining the fitness values of the left direction of the longicorn, the right direction of the longicorn and the direction of the central hair of the longicorn through fitness functions f (x)
Figure BDA0003667175670000033
To determine the fitness value of the longicorn at the current position;
introducing a random step size x of reduced scale t Iteratively updating the step value of the longicorn whisker algorithm, establishing a longicorn movement model,
Figure BDA0003667175670000034
δ t is the step-size factor at the t-th iteration, sign (.) is the optimization direction function.
Optionally, the step S13 includes:
setting the range of the hyper-parameter to be 0-mxn, setting the number of grid searching nodes to be m, setting the step length to be n, and obtaining the global optimization initial value of the hyper-parameter.
Optionally, before the step S12, the method further includes:
and step S21, performing data cleaning on the sample data.
Optionally, the data cleaning of the sample data includes:
cleaning characteristic dimensions, and removing switching value information in the sample data;
cleaning sample dimensions, and removing abnormal data in the sample data;
and cleaning noise data, and performing variance analysis on the data of the residual state parameters in the sample data to eliminate the noise data.
Optionally, between step S21 and step S13, further comprising:
and step S22, carrying out normalization processing on the sample data.
Optionally, before the step S12, the method further includes:
step S23, taking the predicted quantity in the sample data as an output vector, calculating the correlation strength between the residual quantity in the sample data and the output vector, and eliminating the quantity of the sample data, the correlation coefficient of which with the predicted quantity is smaller than a preset value.
Optionally, after the step S14, the method further includes:
and step S15, constructing a single-parameter prediction model, and traversing each parameter to obtain a prediction model of each parameter.
According to another aspect of the present invention, there is provided a fan state prediction apparatus for performing the above-mentioned fan state prediction method, the fan state prediction apparatus comprising:
the system comprises an operation data acquisition module, a data processing module and a data processing module, wherein the operation data acquisition module is configured to acquire sample data, and the sample data comprises integral-level data, equipment-level data and component-level data;
the fan state prediction modeling module is configured to establish a least square support vector machine model by taking the quantity to be predicted in the sample data as an output vector and the residual quantity of the output vector as an input vector; wherein the kernel function of the least squares support vector machine model is a radial basis kernel function;
the prediction model hyper-parameter optimizing module is configured to determine a global optimization initial value of a hyper-parameter in the radial basis kernel function by utilizing a grid search method; and determining the global optimum value of the hyperparameter by using the celestial cow whisker algorithm by taking the global optimization initial value as the initial celestial cow mass center.
According to the technical scheme of the embodiment of the invention, the accuracy of the fan state prediction model can be improved by the adoption of the fan state prediction method through two-stage optimization of the hyper-parameters and sample data of diversity; and the system is combined with a state maintenance system, so that the running states of various fans of the thermal power plant can be predicted in real time, trend graphs and relative deviations of predicted values and measured values are displayed, operation and maintenance personnel can visually check and analyze fan state data, data support is provided for health degree evaluation, fault early warning and diagnosis of the fans, and a control strategy can be provided for the optimized running of the follow-up fans. Meanwhile, the accurate prediction of the state of the fan can reduce the unplanned shutdown of the thermal power generating unit, reduce the operation and maintenance cost of the thermal power plant and ensure the stability and the safety of the operation of the power plant or a power grid. In addition, the superparameter of the LSSVMR model is optimized in two stages globally through a gridding search algorithm and a longicorn algorithm, and the optimization iteration times of the superparameter are reduced.
It should be understood that the statements in this section do not necessarily identify key or critical features of the embodiments of the present invention, nor do they necessarily limit the scope of the invention. Other features of the present invention will become apparent from the following description.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
Fig. 1 is a flowchart of a method for predicting a fan state according to an embodiment of the present invention;
fig. 2 is a flowchart of another method for predicting a fan state according to an embodiment of the present disclosure;
fig. 3 is a flowchart of another method for predicting a fan state according to an embodiment of the present disclosure;
FIG. 4 is a graph of primary air fan outlet pressure prediction results output using a primary air fan state prediction model;
FIG. 5 is a primary air fan outlet pressure prediction relative error plot output using a primary air fan state prediction model;
fig. 6 is a schematic structural diagram of a fan state prediction apparatus according to an embodiment of the present invention;
FIG. 7 illustrates a block diagram of an electronic device 10 that may be used to implement an embodiment of the invention.
Detailed Description
In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the invention described herein are capable of operation in sequences other than those illustrated or described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps S or elements is not necessarily limited to those steps S or elements expressly listed, but may include other steps S or elements not expressly listed or inherent to such process, method, article, or apparatus.
Fig. 1 is a flowchart of a fan state prediction method according to an embodiment of the present invention, and referring to fig. 1, the fan state prediction method includes:
step S11, sample data is obtained, wherein the sample data comprises integral level data, equipment level data and component level data;
specifically, the fans may be, for example, three fans of a thermal power plant, such as a blower, a primary fan, and an induced fan. The primary fan and the air feeder are mainly used for providing primary air quantity, secondary air quantity and air temperature containing pulverized coal required by combustion for the boiler, so that stable combustion of a hearth is ensured; the induced draft fan mainly takes out the flue gas of boiler, plays the effect of maintaining boiler negative pressure. The system comprises an integral level, an equipment level and a component level, wherein the integral level is a generator set, the equipment level is a fan system, and the component level is a fan, a motor, an oil system, a frequency converter and the like; the overall level data is used for representing the overall operation state of the power plant generator set and the boiler; the main parameters of the integral data are mapped to the single fan equipment layer, and the integral data mainly comprise actual power of the generator set, total combustion quantity of the boiler, main steam flow, total air quantity and the like. The equipment level data is used for representing the running state of a fan system of the power plant, and mapping main parameters of the fan system to a single fan equipment level, wherein the parameters of a primary fan comprise primary air volume, air pressure, air temperature, opening degree of a damper and the like; the parameters of the secondary fan comprise secondary air quantity, air pressure, air temperature, opening degree of a damper and the like; the induced draft fan system comprises flue gas flow, wind pressure, wind temperature, air door baffle plate opening degree and the like. The component level data is used for representing the operating states of all components of the single fan device, for example, the parameters of the fan comprise movable blade angles, inlet pressure, inlet air temperature, outlet pressure, outlet air temperature, bearing temperature, oil temperature, bearing vibration and the like; the parameters of the motor comprise current, stator winding temperature, bearing vibration and the like; the parameters of the oil system comprise oil temperature, oil pressure of a lubricating oil pump, oil pressure of an adjusting oil pump, oil level of an oil tank, oil temperature of the oil tank and the like; if the fan adopts a frequency conversion system, the parameters of the frequency conversion device comprise frequency, output current and the like.
Step S12, using the quantity to be predicted in the sample data as an output vector and the allowance as an input vector, and establishing a least square support vector machine model; wherein the kernel function of the least squares support vector machine model is a radial basis function.
Specifically, the quantity to be predicted in the sample data may be a certain parameter in the device-level data or a certain parameter in the component-level data; after the quantity to be predicted is determined (for example, the quantity to be predicted is determined to be wind pressure), the rest quantity except the quantity to be predicted in the sample data is used as an input vector, and a least square support vector machine model is trained through the input vector and the output vector, so that a prediction model of the quantity to be predicted is determined. The Least Square Support Vector Machine Regression (LSSVMR) algorithm has strong nonlinear fitting capability and is very suitable for nonlinear modeling and prediction of the primary fan state with complex working conditions and multivariable strong coupling. By introducing a kernel function idea, modeling is carried out by adopting an LSSVMR algorithm based on historical operating data of the fan, and the prediction precision of the model depends on the optimization results of two super-parameter values of a standardized parameter sigma and a penalty factor gamma in the algorithm. In this embodiment, the establishing the least squares support vector machine model may specifically include:
first a set of sample sets is given:
S＝{(x i ，y i )，x i ∈R n ，y i ∈R}，i＝1，2，…，N
x i ∈R n the feature before or after dimensionality reduction (the dimensionality reduction process will be described later) is used as the pre-dimensionality reductionMeasuring an input vector of the model, wherein n is the dimension of the input vector; y is i E is R is the output vector of the prediction model; n is the size of the sample. The sample set may be randomly divided into training sets S in a certain ratio (e.g., 4: 1) train And test set S test 。
The goal of the LSSVMR algorithm is to construct a prediction function of the form:
Figure BDA0003667175670000071
omega is a weight vector;
Figure BDA0003667175670000072
is the kernel function of LSSVMR, and can convert the input sample x i Mapping to a high-dimensional feature space to perform linear regression solution; b is a deviation amount.
According to the structural risk minimization principle, the objective function of the LSSVMR optimization problem is described as:
Figure BDA0003667175670000081
Figure BDA0003667175670000082
e i is the fitting error; gamma is a penalty factor and is used for controlling the penalty degree of the error. For the optimization problem of equality constraint, the method can adopt a structured Lagrangian function to solve, and introduce a Lagrangian multiplier lambda i The method comprises the following steps:
Figure BDA0003667175670000083
according to the KKT (Karush-Kuhn-Tucker) condition, solving the partial derivative optimization of the formula, and eliminating the weight vector omega and the error vector e of the feature space i Converting quadratic programming problem into dual problem by using optimization theory and solving optimal solution of omega and bObtaining a regression function of the fan state prediction:
Figure BDA0003667175670000084
Figure BDA0003667175670000085
for the kernel function, a non-linear mapping of the input space to the high-dimensional feature space is represented. The selection of the kernel function has great influence on the prediction performance of the LSSVMR, and the selection of the proper kernel function can improve the efficiency of a prediction algorithm, reduce the prediction time and improve the prediction performance. The invention selects a radial basis kernel function as a kernel function of a prediction model, and the expression is as follows:
Figure BDA0003667175670000086
σ is a normalized parameter of the kernel function.
In an LSSVM prediction model of a radial basis function, a standardized parameter sigma reflects the distribution characteristic of training sample data, a penalty factor gamma determines the magnitude of a training error and the strength of generalization capability, and the two hyper-parameters directly influence the prediction effect of the model. In the traditional LSSVMR prediction, the two hyper-parameters are usually selected according to experience, and in order to improve the performance of model prediction, an intelligent optimization algorithm is selected to optimize the 2 hyper-parameters.
And step S13, determining the global optimization initial value of the hyper-parameter in the radial basis kernel function by using a grid search method.
Specifically, the grid search method is an exhaustive search method for specifying parameter values, and obtains an optimal learning algorithm by optimizing parameters of an estimation function through a cross validation method. That is, the possible values of each parameter are arranged and combined, and all possible combination results are listed to generate a 'grid'. Each combination was then used for LSSVMR training and performance was assessed using cross-validation. In this embodiment, the global optimal initial value of the hyper-parameter may be determined first by the gridding search algorithm, so that the situation that the hyper-parameter falls into a locally optimal trap due to the random setting of the initial value of the hyper-parameter is avoided.
And step S14, determining the global optimum value of the hyper-parameter by using the celestial cow whisker algorithm with the global optimum initial value as the initial celestial cow centroid.
Specifically, after the global optimization initial value is determined, the optimal value of the hyper-parameter can be searched by using a longicorn algorithm, and the optimal value of the hyper-parameter is substituted into a least square support vector machine model, so that the fan state prediction model can be obtained.
In this embodiment, after determining the global optimal value of the hyper-parameter, step S15 may be executed to construct a single-parameter prediction model, and traverse each parameter to obtain a prediction model of each parameter; the training set data S can be obtained by taking two hyperparameters sigma and gamma obtained by optimization as set values of the LSSVMR algorithm train And inputting the LSSVMR prediction model of the fan for training to obtain a certain parameter prediction model of the fan. Then testing set data S test Inputting the trained prediction model to obtain the predicted value of a certain parameter. And comparing the predicted value and the true value of the parameter, and preferably selecting an evaluation index to evaluate the effect of the prediction model. The method can select the average absolute percentage error MAPE and the root mean square error RMSE as the evaluation indexes of the model prediction effect, wherein the MAPE of the primary air fan outlet pressure is less than 1 percent and the RMSE is<0.1 (typically around 1% of the range of measurements).
According to the technical scheme of the embodiment, the fan state prediction method is adopted, and the accuracy of the fan state prediction model can be improved through two-stage optimization of the super-parameters and sample data of diversity; and the system is combined with a state maintenance system, so that the running states of various fans of the thermal power plant can be predicted in real time, trend graphs and relative deviations of predicted values and measured values are displayed, operation and maintenance personnel can visually check and analyze fan state data, data support is provided for health degree evaluation, fault early warning and diagnosis of the fans, and a control strategy can be provided for the optimized running of the follow-up fans. Meanwhile, the accurate prediction of the state of the fan can reduce the unplanned shutdown of the thermal power generating unit, reduce the operation and maintenance cost of the thermal power plant and ensure the stability and the safety of the operation of the power plant or a power grid. In addition, the superparameter of the LSSVMR model is optimized in two stages globally through a gridding search algorithm and a longicorn algorithm, and the optimization iteration times of the superparameter are reduced.
Preferably, fig. 2 is a flowchart of another wind turbine state prediction method provided by an embodiment of the present invention, and referring to fig. 2, taking a global optimization initial value as an initial longicorn centroid, and determining a global optimal value of a hyper-parameter by using a longicorn whisker algorithm includes:
and step S141, updating the fitness values of the random step length, the anoplophora chinensis whisker orientation and the current position, wherein the anoplophora chinensis whisker orientation comprises an anoplophora chinensis whisker left whisker, an anoplophora chinensis whisker right whisker and an anoplophora chinensis whisker.
Specifically, the longicorn stigma may be first defined as a random direction vector
Figure BDA0003667175670000101
Figure BDA0003667175670000102
Wherein k is an optimization dimension, and rand (.) is a random function; in this embodiment, the value of k is 2. Then establishing space coordinates of the longicorn left beard, the longicorn right beard and the longicorn middle beard:
Figure BDA0003667175670000103
in this embodiment, the longicorn is abstracted as a centroid, x 0 As coordinates of the initial centroid, x l Is a coordinate, x, of crawling along the left beard of a longicorn r Is a coordinate, x, of crawling along the right beard of the longicorn m Coordinates for crawling along the longicorn, d 1 、d 2 、d 3 The distances from the corresponding longicorn left beard, longicorn right beard and longicorn middle beard to the mass center respectively, t is the iteration frequency,
Figure BDA0003667175670000104
respectively are coordinates of the longicorn after crawling along the left tassel, the right tassel and the middle tassel of the longicorn in the t iteration,
Figure BDA0003667175670000105
the position of the barnyard centre of mass at the t iteration and R (theta) is a random direction vector
Figure BDA0003667175670000108
The rotation matrix of (a); in this embodiment, θ is equal to 90 °, and of course, in other embodiments, θ may have other values.
Then, the minimum mean square error of the predicted value and the measured value of the to-be-predicted quantity is used as a fitness function f (x), and the fitness values of the left beard direction of the longicorn, the right beard direction of the longicorn and the central beard direction of the longicorn are respectively obtained through the fitness function f (x)
Figure BDA0003667175670000106
To determine the fitness value of the longicorn at the current position; introducing a random step size x of reduced scale t Iteratively updating the step value of the longicorn whisker algorithm, establishing a longicorn movement model,
Figure BDA0003667175670000107
δ t is the step-size factor at the t-th iteration, sign (.) is the optimization direction function.
Step S142, calculating fitness values of the longicorn left beard, the longicorn right beard and the longicorn middle beard after moving by random step length to determine the next position of the longicorn;
specifically, after the value of the random step is determined in step S141, the fitness values of the longicorn after the longicorn moves the random step along the longicorn left whisker, the longicorn right whisker and the longicorn middle whisker are calculated, respectively, and the position with the minimum fitness value is used as the next position of the longicorn.
Step S143, determining whether the longicorn is moved to the next position according to the fitness value of the longicorn at the current position and the next position, if the fitness value of the longicorn at the next position is smaller than the fitness value of the longicorn at the current position, moving the longicorn to the next position, and executing step S144; if the fitness value of the longicorn at the next position is greater than or equal to the fitness value at the current position, the process returns to step S141.
Specifically, if the fitness value of the longicorn at the current position is greater than the fitness value of the longicorn at the next position, the situation that the longicorn at the next position is closer to the target value is shown, so that the longicorn can be moved to the next position; if the fitness value of the longicorn at the current position is less than or equal to the fitness value at the next position, the current position is closer to the target, and therefore the movement is not required, and the step S143 is returned to, and the random step length, the orientation of the longicorn and the fitness value of the current position are updated. It should be noted that, at this time, the fitness value of the skyhook orientation and the current position is determined, the step size of the skyhook may be updated only by using the skyhook movement model, and step S142 and step S141 are sequentially executed again by using the new step size until the fitness value of the skyhook at the next position is smaller than the fitness value of the skyhook at the current position.
Step S144, comparing the fitness value of the next position obtained in the step S143 with a set minimum fitness value, and if the fitness value is smaller than the set minimum fitness value or reaches the maximum iteration times, ending the iteration to obtain a global optimal value of the hyper-parameter; if not, the step S141 is executed again.
Specifically, the minimum fitness value is set to be about 1% of the measurement value range, for example, the fan outlet pressure may be 0.1, the maximum number of iterations may be 80, for example, if the fitness value of the next position is smaller than the set minimum fitness value, which indicates that the error between the longicorn and the target is already small, it may be considered that the longicorn has reached the target, and thus the position of the longicorn at this time may be used as the result of the optimization. If the iteration times exceed the maximum iteration times, the loop can be forcibly ended, so that the program is prevented from circulating indefinitely. If the fitness value of the next position obtained in step S143 is greater than or equal to the set minimum fitness value and does not reach the maximum iteration number, which indicates that the longicorn is farther from the target, the position of the longicorn is updated to the current position, and step S141 is executed again, so that the longicorn is controlled to continue to move toward the target.
In this embodiment, the number of the movement directions of the longicorn is three, and the current position of the longicorn is the longicorn left whisker, the longicorn right whisker and the longicorn middle whisker, respectively, and the search direction is increased by one middle whisker relative to the conventional longicorn whisker algorithm, so that the operation speed can be greatly improved, the calculation is performed by adopting a random mixed step length, the iteration times of the search is greatly reduced, the calculation time is shortened, and the convergence speed and the prediction accuracy are further improved.
Preferably, the determining the global optimization initial value of the hyperparameter in the radial basis function by using the grid search method comprises the following steps:
the LSSVMR and gridding search parameters may be initialized first, for example with the range of the hyper-parameters set to 0-30 and the step size set to 5. And then calculating the calculation accuracy of each grid point in turn, and obtaining a global optimization initial value of the hyper-parameter by comparing the calculation accuracy of each grid point. Such as taking the grid point with the highest calculation precision as the global optimization initial value of the hyper-parameter.
Optionally, fig. 3 is a flowchart of another fan state prediction method according to an embodiment of the present invention, and with reference to fig. 3, before step S12, the method further includes: step S21, data cleaning is performed on the sample data.
Specifically, the sample data screens historical sample data from SIS or DCS systems and other monitoring systems, which may exceed characteristic parameters related to the hyperparameters in the characteristic dimension, and may also exceed normal operating conditions of the wind turbine to which the present invention is directed in the sample number dimension, for example, fault data, noise data, shutdown data, and the like may be included. Therefore, the effect of removing redundancy and error quantity is achieved by cleaning the sample data; the prediction precision is effectively improved, and the method is suitable for complex and changeable severe operating environments of power plants. The data cleaning of the sample data may specifically include:
firstly, cleaning characteristic dimensions; the collected fan sample data dimension not only comprises the state characteristics concerned by the invention, but also comprises characteristics irrelevant to the invention, such as on-off information of fan component states, operation modes, alarm signals and the like, and the data is mainly used for control logics of the fan and the components thereof. For continuous fan state prediction, the discrete switching value data not only increases the redundancy of training data, but also reduces the accuracy and efficiency of a prediction model and is not beneficial to the training of the model, so the first step of data cleaning is to eliminate the switching value data. In addition, for a plurality of measured values caused by faults of some measuring points are lost or overrun, data lose value, and characteristic data of the measuring points can be directly removed.
Secondly, dimensional cleaning of a sample; in the actual production of a power plant, a large amount of useless sample data can be recorded in a sample data set during the size repair and temporary stop of a unit and the temporary repair of a single fan, and the sample data cannot be used for establishing a fan state prediction model and is regarded as abnormal data to be cleaned. And in sample data, the overhaul test data of a single fan is also required to be cleaned. And for a small number of missing values in the sample data, performing data filling by adopting a mean value filling method or a hot card filling method.
Thirdly, cleaning noise data; and carrying out variance analysis on residual sample data of each state parameter of the fan, and smoothing the noise data exceeding the limit by adopting a box median method.
Optionally, with continued reference to fig. 3, step S12 is preceded by: step S22, normalization processing is performed on the sample data.
Specifically, the sample data related to this embodiment is more and the dimensions are also not uniform, the influence of the data value size and the dimensions on the sample analysis and the algorithm optimization convergence can be reduced through normalization, and a standard deviation normalization method or a linear normalization method is selected to perform normalization processing on the data according to the distribution characteristics of the data, where the formula of the standard deviation normalization method is as follows:
Figure BDA0003667175670000131
Figure BDA0003667175670000132
the value is a numerical value after normalization processing of each input parameter, and the value is between 0 and 1; x is the original value of each input parameter;
Figure BDA0003667175670000133
an average value input for the parameter; σ is the value of the parameter inputStandard deviation of (2).
The formula of the linear normalization method is:
Figure BDA0003667175670000134
x min a minimum value input for the parameter; x is the number of max The maximum value entered for this parameter.
Optionally, with continued reference to fig. 3, step S12 is preceded by: and (3) taking the predicted quantity in the sample data as an output vector, calculating the correlation strength of the residual quantity and the output vector in the sample data, and rejecting the quantity of which the correlation coefficient with the predicted quantity is smaller than a preset value in the sample data.
Specifically, for data after sample cleaning, a pearson correlation coefficient method can be adopted to calculate the correlation strength between the quantity to be predicted and other quantities, if the correlation coefficient is smaller than a preset value, it is indicated that the correlation between the quantity and the quantity to be predicted is low, and the data corresponding to the quantity can be removed, so that the effect of reducing the dimension is achieved. Of course, it should be noted that the data to be removed corresponding to different quantities to be predicted may also be different.
In this embodiment, a prediction model of each parameter may be obtained by traversing each parameter, and only the parameter to be predicted is selected as the quantity to be predicted, and the fan state prediction method provided by the embodiment of the present invention is sequentially executed.
For example, fig. 4 is a primary air fan outlet pressure prediction result graph output by using a primary air fan state prediction model, and fig. 5 is a primary air fan outlet pressure prediction relative error graph output by using a primary air fan state prediction model; with reference to fig. 4 and 5, a primary air fan outlet pressure prediction model is constructed by taking the primary air fan outlet pressure as the quantity to be predicted, and comparison with an actual value shows that the fan state prediction method provided by the invention has high prediction accuracy.
Fig. 6 is a schematic structural diagram of a fan state prediction apparatus according to an embodiment of the present invention, and referring to fig. 6, the fan state prediction apparatus includes:
the operating data acquisition module 601 is configured to acquire sample data, wherein the sample data includes integral-level data, equipment-level data and component-level data;
the fan state prediction modeling module 602 is configured to establish a least square support vector machine model by using the quantity to be predicted in the sample data as an output vector and the residual quantity as an input vector; the kernel function of the least square support vector machine model is a radial basis kernel function;
the prediction model hyper-parameter optimizing module 603 is configured to determine a global optimization initial value of a hyper-parameter in the radial basis kernel function by using a grid search method; and determining the global optimal value of the hyper-parameter by using a longicorn whisker algorithm by taking the global optimal initial value as the initial longicorn centroid.
The fan state prediction apparatus provided in this embodiment is configured to execute the fan state prediction method provided in any embodiment of the present invention, and the working principle of the fan state prediction apparatus may refer to the description of the fan state prediction method part in the embodiment of the present invention, which is not described herein again. According to the technical scheme of the embodiment, the fan state prediction method is adopted, and the accuracy of the fan state prediction model can be improved through two-stage optimization of the super-parameters and sample data of diversity; and the system is combined with a state maintenance system, so that the running states of various fans of the thermal power plant can be predicted in real time, trend graphs and relative deviations of predicted values and measured values are displayed, operation and maintenance personnel can visually check and analyze fan state data, data support is provided for health degree evaluation, fault early warning and diagnosis of the fans, and a control strategy can be provided for the optimized running of the follow-up fans. Meanwhile, the state of the fan is accurately predicted, so that the unplanned shutdown of the thermal power generating unit is reduced, the operation and maintenance cost of the thermal power plant is reduced, and the stability and the safety of the operation of the power plant or a power grid are ensured. In addition, the superparameter of the LSSVMR model is optimized in two stages globally through a gridding search algorithm and a longicorn algorithm, and the optimization iteration times of the superparameter are reduced.
It should be noted that the fan state prediction apparatus may further include: and the data preprocessing module is configured to clean and normalize the sample data. And the data characteristic selection module is configured to perform dimension reduction processing on the sample data. And the fan overall prediction model module is configured to construct a prediction model of each state parameter of the fan, set the optimal hyper-parameter of each model, input training set data of the operation data into the fan state prediction model for training, and obtain the optimized fan state prediction model. And inputting the test set data of the operating data into a model for model verification test, and evaluating the model prediction effect by adopting indexes of Mean Absolute Percentage Error (MAPE) and Root Mean Square Error (RMSE). And the overall fan state prediction and display module is configured to calculate the predicted value of each state parameter in real time based on the fan state prediction model, and display the measured value and the predicted value of the fan state, the relative error of the measured value and the predicted value and the trend chart of the fan state.
FIG. 7 illustrates a schematic diagram of an electronic device 10 that may be used to implement an embodiment of the invention. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the inventions described and/or claimed herein.
As shown in fig. 7, the electronic device 10 includes at least one processor 11, and a memory communicatively connected to the at least one processor 11, such as a Read Only Memory (ROM)12, a Random Access Memory (RAM)13, and the like, wherein the memory stores a computer program executable by the at least one processor, and the processor 11 can perform various suitable actions and processes according to the computer program stored in the Read Only Memory (ROM)12 or the computer program loaded from a storage unit 18 into the Random Access Memory (RAM) 13. In the RAM 13, various programs and data necessary for the operation of the electronic apparatus 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other via a bus 14. An input/output (I/O) interface 15 is also connected to bus 14.
A number of components in the electronic device 10 are connected to the I/O interface 15, including: an input unit 16 such as a keyboard, a mouse, or the like; an output unit 17 such as various types of displays, speakers, and the like; a storage unit 18 such as a magnetic disk, an optical disk, or the like; and a communication unit 19 such as a network card, modem, wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information/data with other devices via a computer network such as the internet and/or various telecommunication networks.
The processor 11 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various processors running machine learning model algorithms, a Digital Signal Processor (DSP), and any suitable processor, controller, microcontroller, or the like. The processor 11 performs the various methods and processes described above, such as a fan state prediction method.
In some embodiments, the fan condition prediction method may be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as the storage unit 18. In some embodiments, part or all of the computer program may be loaded and/or installed onto the electronic device 10 via the ROM 12 and/or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more of the steps S of the fan status prediction method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to perform the wind turbine state prediction method in any other suitable manner (e.g., by means of firmware).
Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuitry, Field Programmable Gate Arrays (FPGAs), Application Specific Integrated Circuits (ASICs), Application Specific Standard Products (ASSPs), system on a chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which may be special or general purpose, receiving data and instructions from, and transmitting data and instructions to, a storage system, at least one input device, and at least one output device.
A computer program for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the computer programs, when executed by the processor, cause the functions/acts specified in the flowchart and/or block diagram block or blocks to be performed. A computer program can execute entirely on a machine, partly on a machine, as a stand-alone software package partly on a machine and partly on a remote machine or entirely on a remote machine or server.
In the context of the present invention, a computer-readable storage medium may be a tangible medium that can contain, or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. A computer readable storage medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. Alternatively, the computer readable storage medium may be a machine readable signal medium. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
To provide for interaction with a user, the systems and techniques described here can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which a user can provide input to the electronic device. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic, speech, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a back-end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), Wide Area Networks (WANs), blockchain networks, and the internet.
The computing system may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, also called a cloud computing server or a cloud host, and is a host product in a cloud computing service system, so that the defects of high management difficulty and weak service expansibility in the traditional physical host and VPS service are overcome.
It should be understood that steps S may be reordered, added, or deleted using various forms of the flow shown above. For example, the steps S described in the present invention may be executed in parallel, may be executed sequentially, or may be executed in different orders, as long as the desired result of the technical solution of the present invention can be achieved, and the present invention is not limited thereto.
The above-described embodiments should not be construed as limiting the scope of the invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may be made in accordance with design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (10)

1. A fan state prediction method is characterized by comprising the following steps:
step S11, sample data is obtained, wherein the sample data comprises integral level data, equipment level data and component level data;
step S12, taking the quantity to be predicted in the sample data as an output vector and the residual quantity as an input vector, and establishing a least square support vector machine model; wherein the kernel function of the least squares support vector machine model is a radial basis kernel function;
step S13, determining a global optimization initial value of the hyper-parameter in the radial basis kernel function by using a grid search method;
and step S14, determining the global optimal value of the hyper-parameter by using the global optimization initial value as the initial longicorn centroid and using a longicorn whisker algorithm.
2. The wind turbine state prediction method according to claim 1, wherein the determining the global optimal value of the hyper-parameter by using the global optimization initial value as an initial skyhook centroid using a skyhook algorithm comprises:
step S141, updating the random step length, the longicorn stigma orientation and the fitness value of the current position, wherein the longicorn stigma orientation comprises a longicorn left stigma, a longicorn right stigma and a longicorn middle stigma;
step S142, calculating the fitness value of the longicorn after the longicorn moves the random step length along the left beard, the right beard and the middle beard of the longicorn to determine the next position of the longicorn;
step S143, determining whether the longicorn is moved to the next position according to the fitness value of the longicorn at the current position and the next position, if the fitness value of the longicorn at the next position is smaller than the fitness value of the longicorn at the current position, moving the longicorn to the next position, and executing step S144; if the fitness value of the longicorn at the next position is larger than or equal to the fitness value at the current position, returning to execute the step S141;
step S144, comparing the fitness value of the next position obtained in step S143 with a set minimum fitness value, and if the fitness value is smaller than the set minimum fitness value or reaches the maximum iteration number, ending the iteration to obtain a global optimal value of the hyper-parameter; if not, the step S141 is executed again.
3. The fan condition prediction method of claim 2,
the updating the fitness values of the random step size, the skynet whisker orientation and the current position comprises:
defining the orientation of the longicorn stigma as a random direction vector
Figure FDA0003667175660000021
Wherein k is an optimization dimension, and rand (.) is a random function;
establishing space coordinates of the longicorn left tassel, the longicorn right tassel and the longicorn middle tassel;
Figure FDA0003667175660000022
wherein x is 0 Is the coordinate of the initial centroid, x l Is a coordinate, x, of crawling along the left beard of a longicorn r Is a coordinate, x, of crawling along the right beard of the longicorn m Coordinates for crawling along the longicorn, d 1 、d 2 、d 3 Respectively the distances from the corresponding longicorn left beard, longicorn right beard and longicorn middle beard to the mass center, t is the iteration times,
Figure FDA0003667175660000023
respectively are coordinates of the longicorn after crawling along the left tassel, the right tassel and the middle tassel of the longicorn in the t iteration,
Figure FDA0003667175660000024
the position of the barnyard centre of mass at the t iteration and R (theta) is a random direction vector
Figure FDA0003667175660000025
The rotation matrix of (a);
taking the minimum mean square error of the predicted value and the measured value of the quantity to be predicted as a fitness function f (x);
respectively obtaining the fitness values of the left direction of the longicorn, the right direction of the longicorn and the middle direction of the longicorn through fitness functions f (x)
Figure FDA0003667175660000026
To determine the fitness value of the longicorn at the current position;
introducing a random step size x of reduced scale t Iteratively updating the step value of the longicorn whisker algorithm, establishing a longicorn movement model,
Figure FDA0003667175660000027
δ t is the step-size factor at the t-th iteration, sign (.) is the optimization direction function.
4. The fan condition prediction method according to claim 1, wherein the step S13 includes:
setting the range of the hyper-parameter to be 0-mxn, setting the number of grid searching nodes to be m, setting the step length to be n, and obtaining the global optimization initial value of the hyper-parameter.
5. The fan condition prediction method according to claim 1, wherein the step S12 is preceded by:
and step S21, performing data cleaning on the sample data.
6. The wind turbine state prediction method according to claim 5, wherein the data cleaning of the sample data comprises:
cleaning characteristic dimensions, and removing switching value information in the sample data;
cleaning sample dimensions, and removing abnormal data in the sample data;
and cleaning noise data, and performing variance analysis on the data of the residual state parameters in the sample data to eliminate the noise data.
7. The fan condition prediction method according to claim 5, further comprising between the step S21 and the step S13:
and step S22, carrying out normalization processing on the sample data.
8. The fan condition prediction method according to claim 1, wherein the step S12 is preceded by:
step S23, taking the predicted quantity in the sample data as an output vector, calculating the correlation strength between the residual quantity in the sample data and the output vector, and eliminating the quantity of the sample data, the correlation coefficient of which with the predicted quantity is smaller than a preset value.
9. The fan condition prediction method according to claim 1, further comprising, after the step S14:
and step S15, constructing a single-parameter prediction model, and traversing each parameter to obtain a prediction model of each parameter.
10. A fan state prediction apparatus for executing the fan state prediction method according to any one of claims 1 to 9, the fan state prediction apparatus comprising:
the operating data acquisition module is configured to acquire sample data, wherein the sample data comprises integral-level data, equipment-level data and component-level data;
the fan state prediction modeling module is configured to establish a least square support vector machine model by taking the quantity to be predicted in the sample data as an output vector and the residual quantity of the output vector as an input vector; wherein the kernel function of the least squares support vector machine model is a radial basis kernel function;
the prediction model hyper-parameter optimizing module is configured to determine a global optimization initial value of a hyper-parameter in the radial basis kernel function by utilizing a grid search method; and determining the global optimal value of the hyper-parameter by using the global optimization initial value as the initial longicorn centroid and using a longicorn whisker algorithm.
CN202210594394.2A 2022-05-27 2022-05-27 Fan state prediction method and device Pending CN114971024A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210594394.2A CN114971024A (en) 2022-05-27 2022-05-27 Fan state prediction method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210594394.2A CN114971024A (en) 2022-05-27 2022-05-27 Fan state prediction method and device

Publications (1)

Publication Number Publication Date
CN114971024A true CN114971024A (en) 2022-08-30

Family

ID=82956979

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210594394.2A Pending CN114971024A (en) 2022-05-27 2022-05-27 Fan state prediction method and device

Country Status (1)

Country Link
CN (1) CN114971024A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115795999A (en) * 2022-10-26 2023-03-14 国网新源控股有限公司 Performance abnormity early warning method for long-term service pumped storage unit

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115795999A (en) * 2022-10-26 2023-03-14 国网新源控股有限公司 Performance abnormity early warning method for long-term service pumped storage unit
CN115795999B (en) * 2022-10-26 2023-08-01 国网新源控股有限公司 Early warning method for abnormal performance of long-term service pumped storage unit

Similar Documents

Publication Publication Date Title
US11436395B2 (en) Method for prediction of key performance parameter of an aero-engine transition state acceleration process based on space reconstruction
Wan et al. Optimal prediction intervals of wind power generation
Li et al. Predicting short-term electricity demand by combining the advantages of arma and xgboost in fog computing environment
Wang et al. A compound framework for wind speed forecasting based on comprehensive feature selection, quantile regression incorporated into convolutional simplified long short-term memory network and residual error correction
CN111062508A (en) Method for evaluating real-time running state of wind turbine generator based on big data technology
CN110555058A (en) Power communication equipment state prediction method based on improved decision tree
Browell et al. Forecasting for day-ahead offshore maintenance scheduling under uncertainty
CN113011680A (en) Power load prediction method and system
CN114971024A (en) Fan state prediction method and device
Yang et al. A Fuzzy Comprehensive CS-SVR Model-based health status evaluation of radar
Ahmad et al. Efficient energy planning with decomposition-based evolutionary neural networks
CN109523077B (en) Wind power prediction method
Yang et al. Similarity-based information fusion grey model for remaining useful life prediction of aircraft engines
Wang et al. A novel approach of full state tendency measurement for complex systems based on information causality and PageRank: a case study of a hydropower generation system
Li et al. Short-term load forecasting using support vector regression-based local predictor
CN115238573A (en) Hydroelectric generating set performance degradation trend prediction method and system considering working condition parameters
Ding et al. A statistical upscaling approach of region wind power forecasting based on combination model
Chen et al. Research on prediction algorithm of ship equipment heath condition
Kumar et al. A novel technique of optimization for software metric using PSO
Jiang et al. A hybrid degradation tendency measurement method for mechanical equipment based on moving window and Grey–Markov model
Kolesnikova et al. Method for Evaluating the Steam Turbines Operating Efficiency
CN116702630B (en) Error iterative evaluation method for capacitive voltage transformer CVT
Zhao et al. An interpretable ultra-short-term wind power prediction model based on the feature matrix reconstruction through regression trees
Bai et al. Abnormal Detection Scheme of Substation Equipment based on Intelligent Fusion Terminal
Ping et al. Sparse Bayesian Learning Based Hybrid Method for Load Forecasting in Integrated Energy Systems

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination