
第
44
卷第
7
期
2010
年
7
月
浙江大学学报{工学版)
lournal
of
Zhejiang University (Engineering Science)
DOI:
10.
3785/j. issn. 1008-973X. 2010.
07.
031
于多源数据特征融合的球磨机负荷软测
汤
健
1
,赵立杰
l
,
3
,岳
恒
2
,柴天佑1.
2
Vo
l.
44
No.7
Ju
l.
2010
(1.东北大学流程工业综合自动化教育部重点实验室,辽宁沈阳
110189;
2.
东北大学自动化研究中心,
辽宁沈阳
110189;
3.
沈阳化工大学信息工程学院,辽宁沈阳
110142)
摘
要:针对磨矿过程球磨机负荷
(M
L)难以实时检测,生产中主要依靠人工经验判断负荷状态的难题,依据磨机
简体振动、振声、电流等信号与磨机负荷间存在相关性、信息互补与冗余的现象,提出基于多源数据特征融合的球
磨机负荷软测量新方法.该方法由时域滤波、时频转换、特征提取、特征选择及软测量模型
5
部分组成.采用快速傅
里叶变换
(FF
T)将滤波后的简体振动及振声时域信号转换成频域信号,根据研磨机理将频域信号划分为低、中、高
3
个频段,采用核主元分析
(KPCA)
分别提取各个频段的非线性特征,选择振动、振声频域特征与电流时域特征的
融合信号作为模型输入,建立基于最小二乘支持向量机
(LSSVI
峭的磨机负荷软测量模型.实验结果表明,该方法
与基于主元分析最小二乘支持向量机
(PCA-LSSVM)
方法和单传感器方法相比,磨机负荷参数预测精度较高.
关键词:磨机负荷
(M
L)
;特征提取;特征选择;核主元分析
(KPCA)
;最小二乘支持向量机
(LSSVM)
中图分类号:
TP
29
文献标志码
:A
文章编号:
1008
- 973X(2010)07
-1406
-
08
Soft sensor
for
ball mill load based on
mu
It
i-source
data
feature fusion
TANG
lian
1
,
ZHAO
Li-ji
e
1.
3
,
YUE
Heng
2
,
CHAI
Tian-you
l.
2
(1.
Key
Laboratory
of
Integrated
Automation
for
Process
Industry
,
儿
1inistry
of
Education
,
Northeastern
University
,
Shenyang
110189 ,
China;
2.
Resea
γ
ch
Center
of
Automation
,
Northeastern
University
,
Shenyang
110189
,
China;
3.
College
of
Information
Engineering
,
Shenyang
University
of
Chemical
Technology
,
Shenyang
110142 ,
China)
Abstract:
The
real-time
measurement
of
ball mill load
(ML)
in
grinding
process is difficult to realize,
and
the
states
of
ML
are
identified
mainly
by
the
experience
of
the
operator.
Aiming
at
the
problems
, a
new
soft-sensor
approach
of
ML
based
on
the
multi-source
data
feature fusion
was
proposed
according to
the
relativity,
the
information
complementation
and
redundancy
among
shell
vibration
, acoustic, electricity signal
and
M
L.
The
approach
consisted of five
parts
which
were
data
filter,
time/frequency
transform
, feature
extraction
, feature
selection
and
soft
sensor
mode
l.
The
shell
vibration
and
acoustic signal in
the
time
domain
was
transformed
into
the
frequency domain using fast
Fourier
transform
(FFT).
The
spectral signals
were
partitioned
into
three
parts
which
were
low
,
medium
and
high
frequency
bands
according
to
the
grinding
mechanism.
The
kernel
principal
component
analysis
(KPCA)
was
used
to
extract
the
nonlinear
feature
of
each part.
The
fused signals,
which
consisted of
the
frequency domain feature
of
vibration
and
acoustic signal,
and
the
time
domain
feature
of electricity signal,
were
selected
as
the
input
variables
of
the
soft
sensor
mode
l.
The
soft
sensor
model of
ML
was
conducted based
on
the
least
square
support
vector machine
(LSSVM).
Experimental
results
show
that
the
approach
has
better
prediction
accuracy for
ML
parameters
than
the
PCA-
LSSVM
and
the
single
sensor
approaches.
Key
words:
mill
load
(ML);
feature
extraction;
feature
selection;
kennel
principal
component
analysis
(KPCA);
least
square
support
vector
machine
(LSSVM)
收稿日期:
2010
-
04
-
20.
浙江大学学报(工学版)网址:
www.
journals.
勾
u.
edu.
cn/eng
基金项目:国家
"863"
高技术研究发展计划资助项目
CZ006A
A0
60Z0Z).
作者简介:汤健
0974-)
,男,辽宁北票人,博士生.从事综合自动化系统及基于数据驱动技术的软测量建模研究.
E-mail:
tjian001@126.com
通信联系人:柴天佑,男,教授,院士.
E-mail:
tychai@mai
l. neu.edu.cn