A rectifying tower fault diagnosis method of an improved particle swarm optimization support vector machine

A technology of support vector machines and improved particle swarms, applied in computer parts, instruments, manufacturing computing systems, etc., to improve classification accuracy and improve the effect of falling into local optimum

CN109886462AActive Publication Date: 2019-06-14HANGZHOU DIANZI UNIV
3 Cites 1 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Publication Date
2019-06-14

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

The invention provides a rectifying tower fault diagnosis method of an improved particle swarm optimization support vector machine, which is optimized according to the following steps of (1) setting an initial value of a system subgroup, and giving a range of a parameter C; And (2) randomly generating positions and speeds of the particles, and evaluating a fitness value of each particle accordingto a fitness function, and (3) updating the speed and the position according to a corresponding improved formula, and (4) checking whether a termination condition is met or not, if so, ending searching, and if not, returning to the step (2) for recalculation, and (5) obtaining parameters of the optimized support vector machine, and establishing the improved particle swarm optimization support vector machine model. According to the method, the problem that a common particle swarm optimization support vector falls into local optimum and the like due to the fact that the common particle swarm optimization support vector is too large or too small in speed limit value selection at the beginning is effectively solved.
Need to check novelty before this filing date? Find Prior Art

Description

technical field

[0001] The invention belongs to the technical field of optimization algorithm application, and in particular relates to a method for diagnosing a rectification tower fault by improving a particle swarm optimization support vector machine. Background technique

[0002] In the production process of the petrochemical industry, the distillation column occupies an important position. The rectification tower is a tower-type vapor-liquid contact device that utilizes the different volatilities of each component in the mixture to achieve the purpose of separation, thereby achieving rectification. In the petrochemical industry and other industries, the quality of the distillation column equipment in the operation process is directly related to the economic benefits of the enterprise. Therefore, fault diagnosis of distillation columns can improve process safety and product quality.

[0003] In today's rapidly developing technological society, fault diagnosis begins to...

Examples

Embodiment Construction

[0036] Such as figure 1 As shown, a rectification column fault diagnosis method for improved particle swarm optimization support vector machine, the method specifically includes the following steps:

[0037] Step 1: Setting of initial value of particle swarm, given input data X={X 1 ,...,X n} and learning objective y={y 1 ,...,y n} are all derived from the fault data of rectification tower, where T max The maximum number of iterations is 300, set w as the inertia weight of 0.9, and the acceleration factor c 1 is 1.6, the acceleration factor c 2 1.5, V max The initial maximum set speed is 120, X max Set the position to 180 for the initial maximum. Given the parameter C, the range of σ is [0,100], C is the penalty coefficient, and σ is the selected RBF function (K(x i ,x j ) = exp(||x i -x j || 2 / σ 2 )) As a kernel, a parameter of this function, i=1,2,...n,j=1,2,...n,x i =[x 1 ,...,x n ]∈X represents the multiple feature space contained in each sample of the in...