Generator bar vibration optimization method based on modified long-short term memory neural network

A long-short-term memory, neural network technology, applied in the field of generator bar vibration optimization based on modified long-short-term memory neural network, can solve the problem of high cost, lack of historical data level mathematical support, unable to clearly give temperature adjustment direction, etc. question

Active Publication Date: 2021-08-13
浙江浙能数字科技有限公司 +1
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] At present, the treatment methods for excessive vibration at the end of the generator mainly include: overhaul dismantling inspection method, that is, dismantling during overhaul to check the fixing condition of the wire rods, whether there is wear or not, rebinding, reinforcing, and adjusting the structure of the wire rods to make the stator The natural frequency at the end avoids the single and double frequency areas, which has a certain effect, but this method takes a long time, consumes a lot of money, and needs to be shut down, which affects the power generation plan of the unit
The method of adjusting the vibration of the wire rod by trying to increase or decrease the cooling water temperature and the hydrogen temperature according to the feedback has a certain effect, but this method lacks the mathematical support of the historical data level, and cannot clearly give the direction of temperature adjustment. The feedback adjustment mechanism is only suitable for monotone functions, and it is easy to be limited to the local optimal solution for complex generator bar vibration and cannot achieve a good suppression effect

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Generator bar vibration optimization method based on modified long-short term memory neural network
  • Generator bar vibration optimization method based on modified long-short term memory neural network
  • Generator bar vibration optimization method based on modified long-short term memory neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0053] Using data from a 1050MW generator in a power plant, the radial vibration of the B-phase wire rod of the generator often exceeds the standard, and the highest value has reached 383um, which is much higher than the alarm value of 250um. Taking the vibration at the end of the B-phase wire rod of the generator as the research and verification object, a method for optimizing the vibration of the generator wire rod based on the modified long-short-term memory neural network, such as figure 1 shown, including steps:

[0054] Step 1. Train the generator bar vibration LM_LSTM (long short-term memory neural network) model;

[0055] Step 1.1. Obtain historical data for 3 months with an interval of 15 seconds; through mechanism analysis, select the measuring points that have a strong correlation with the vibration of the generator bar as the feature vector. The measuring points include: generator A-phase current, power generation Generator B-phase current, generator C-phase curre...

Embodiment 2

[0083] On the basis of the generator bar vibration LM_LSTM (long short-term memory neural network) model trained in Example 1, the generator bar vibration optimization method is as follows:

[0084] Taking the active power of the generator in the historical data obtained by the plant as the abscissa, and the B-phase radial vibration at the end of the wire rod on the turbine side as the ordinate, draw a scatter diagram to obtain Figure 4 , it can be seen from the figure that when the generator load of the power plant is between 420MW and 630MW, and near 780MW, the B-phase radial vibration at the end of the turbine side bar of the generator easily exceeds the standard value of 250um. Therefore, in this embodiment, the data of three working conditions with unit loads of 480MW, 520MW and 780MW are selected as analysis objects, and the data of working condition 1, working condition 2 and working condition 3 are respectively recorded for optimization analysis.

[0085] Dimensionall...

Embodiment 3

[0095] The method in the above-mentioned embodiment 1 and embodiment 2 can also be used to construct the optimization model of the vibration of the wire rod end of the generator phase A and C, and this method can also input real-time data and output the vibration of the wire rod end of the generator in real time 3D optimization control chart.

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention relates to a generator bar vibration optimization method based on a modified long-short term memory neural network. The method comprises the following steps: acquiring and processing historical data; and taking the feature vector and the label vector as input of a generator bar vibration LMLSTM model, and training the generator bar vibration LMLSTM model by using data in the training set and the verification set. The method has the beneficial effects that: the influence of each parameter of the generator on the vibration of the coil bar is excavated from a method of combining a mechanism and a mathematical theory; a generator bar vibration model based on the corrected long-short term memory neural network is established, the influences of the average temperature of the generator bar and the average temperature of the generator iron core on the vibration of the generator bar are analyzed, and a three-dimensional relation control chart is made; and the overall and local adjustment directions of the stator cooling water inlet temperature of the generator and the cold hydrogen temperature of the generator are clearly pointed out, so that the excessive vibration of the end part of the generator bar in the operation process is suppressed by adjusting the related temperature, and the requirement of safe and stable operation is met.

Description

technical field [0001] The invention belongs to the technical field of generator wire rod vibration optimization, and in particular relates to a generator wire rod vibration optimization method based on a corrected long-short-term memory neural network. Background technique [0002] When the generator is running normally, the stator bar of the generator is subjected to the alternating electromagnetic force of single and double frequency. With the continuous expansion of the unit capacity of the generator, the electromagnetic force also increases. Due to the low stiffness of the end of the generator bar suspended outside the stator core, the problem of excessive vibration at the end of the generator bar caused by the influence of electromagnetic force is becoming more and more prominent. [0003] When the natural vibration frequency of the end of the generator is close to 100HZ, it is easy to resonate with the electromagnetic force of twice the power frequency, causing the v...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06F30/27G06K9/62G06N3/04G06N3/08H02K3/50G06F111/06G06F119/08
CPCG06F30/27G06N3/08H02K3/50G06F2111/06G06F2119/08G06N3/044G06F18/214G06F18/241
Inventor 戴程鹏解剑波范海东郭鼎王豆傅骏伟杨勤金泱王展宏孟瑜炜
Owner 浙江浙能数字科技有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products