Kalman filtering parameter self-adaptive updating method based on extreme learning machine

Abstract

The invention discloses a Kalman filtering parameter self-adaptive updating method based on an extreme learning machine. The method generally includes three parts: a first part, a learning part of anextreme learning machine according to recursive least squares; a second part, carrying out iterative updating algorithm analysis on the Kalman filtering algorithm; and a third part, a learning part which learns according to the limit of Kalman filtering. According to the method, the classification precision of the steam turbine in extreme learning is improved by updating the connection weight fromthe hidden layer to the output layer on line.

Description

technical field [0001] The invention belongs to the field of fault diagnosis, and relates to a method for updating adaptive parameters of a Kalman filter based on an extreme learning machine, which is applied to the fault diagnosis of a steam turbine. Background technique [0002] With the development of the power industry, the degree of automation of power equipment has been continuously improved, and more and more high-parameter and large-capacity steam turbines have played a key role in modern thermal power generation. With the continuous optimization of power equipment, its structure is becoming more and more complex and there are more and more unsafe factors. Therefore, an important task facing the current power industry is how to effectively improve the accuracy of steam turbine fault diagnosis and at the same time ensure the safe and effective operation of its equipment. [0003] In recent years, methods such as neural network and support vector machine have been app...

AI Technical Summary

Problems solved by technology

On the other hand, there are also some problems. First, the training speed based on the gradient learning algorithm is very slow. Second, all the parameters in the neural network need to be solved, so it will take a lot of time. unacceptable for applications

In practical applications, support vector machines also face the difficulty of selecting multiple parameters, and it takes a lot of time to adopt the method of optimizing parameters.

At the same time, the complex structure and various fault mechanisms of steam turbines lead to the need to further improve the recognition accuracy of fault diagnosis and analysis methods

[0004] Extreme learning machine is used to solve the problem of single hidden layer feed-forward neural network. It has the characteristics of fast speed and strong generalization ability, and can use a variety of non-differentiable functions. However, in practice, the data information reflecting the essence of the model may not have been To deal with this problem, the extreme learning machine based on recursive least squares is proposed. The basic idea is to fit the data through least squares without bias, and the recursive least squares method is stable. Good convergence and minimum mean square error in the environment, due to the use of a fixed forgetting factor, it is not able to provide satisfactory results in time-varying and unstable systems

Images