# US 12,002,014 B2 · Soft measurement method for dioxin emission concentration in municipal solid waste incineration process

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 12,002,014 B2
- Document ID: US12002014B2
- Status: Granted
- Kind: B2
- Application number: 16/967,408
- Filed: 2019-12-02
- Published: 2021-07-29
- Granted: 2024-06-04
- Earliest claimed: 2019-03-24
- Assignees: Beijing University of Technology
- Applicants: Beijing University of Technology
- Inventors: Jian Tang; Junfei Qiao; Zihao Guo; Haijun He
- Web page: https://patentgrep.com/patent/US12002014B2

## CPC

- G06Q 10/30: Administration of product recycling or disposal
- G05B 13/027: {using neural networks only}
- G05B 13/042: {in which a parameter or coefficient is automatically adjusted to optimise the performance}
- G05B 13/048: {using a predictor}
- G06N 20/10: using kernel methods, e.g. support vector machines [SVM]
- G06N 20/20: Ensemble learning
- Y02W 90/00: Enabling technologies or technologies with a potential or indirect contribution to greenhouse gas [GHG] emissions mitigation
- 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
- G05B: CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS (systems for controlling or regulating non-electric variables G05D; systems for regulating electric or magnetic variables G05F; control devices or systems insofar as characterised by mechanical features only G05G)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- Y02W: CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO WASTEWATER TREATMENT OR WASTE MANAGEMENT

## Abstract

Disclosed is a soft measurement method of DXN emission concentration based on multi-source latent feature selective ensemble (SEN) modeling. First, MSWI process data is divided into subsystems of different sources according to industrial processes, and principal component analysis (PCA) is used to separately extract the subsystems' latent features and conduct multi-source latent feature primary selection according to the threshold value of the principal component contribution rate preset by experience. Using mutual information (MI) to evaluate the correlation between the latent features of the primary selection and DXN, and adaptively determine the upper and lower limits and thresholds of the latent feature reselection; finally, based on the reselected latent features, a least squares-support vector machine (LS-SVM) algorithm with a hyperparameter adaptive selection mechanism is used to establish DXN emission concentration sub-models for different subsystems, and based on branch and bound (BB) and prediction error information entropy weighting algorithm to optimize the selection of sub-models and calculation weights coefficient, a SEN soft measurement model of DXN emission concentration is constructed.

## Claims

### Claim 1

A soft measurement method of dioxin emission concentration in a municipal solid waste incineration (MSWI) process, comprising steps of:
using a latent feature extraction and primary selection module to divide data associated with the municipal solid waste incineration process into subsystems of different sources according to industrial processes;
using a principal component analysis (PCA) to separately extract latent features of the subsystems;
conducting a primary selection of the latent features according to a threshold value of a principal component contribution rate preset by experience;
using a latent feature evaluation and reselection module, measuring, based on mutual information (MI), a correlation between the latent features selected via the primary selection and dioxin (DXN);
adaptively determining an upper limit, a lower limit and thresholds of reselection of the latent features; and
using an adaptive selective ensemble modeling module, further comprising:
based on latent feature reselection, using a least squares-support vector machine (LS-SVM) algorithm with a hyperparameter adaptive selection mechanism to establish DXN emission concentration sub-models for different subsystems; and
adopting a strategy optimization based on a branch and bound (BB) and a prediction error information entropy weighting algorithm to select the DXN emission concentration sub-models and calculating weight coefficients to construct a DXN emission concentration selective ensemble soft measurement model, wherein the DXN emission concentration selective ensemble (SEN) soft measurement model is used to optimize the municipal solid waste incineration process to reduce emission of the DXN by the municipal solid waste incineration process.

### Claim 2 (depends on 1)

The soft measurement method of dioxin emission concentration in the municipal solid waste incineration process according to claim 1, wherein the step of using the latent feature extraction and primary selection module further comprises the steps of:
for an ith subsystem, using the PCA to extract the latent features of high-dimensional input process variables comprising normalizing input data Xi to mean=0 and variance=1, and decomposing Xi into:
Xi=t1FeAlli(p1FeAlli)T+L+tmFeAllii(pmFeAllii)T+L+tMFeAllii(pMFeAlli)T (3);
wherein, tmFeAllii and pmFeAllii represent a score and load vector of mFeAllith principal component (PC), T represents a transpose, and mFeAllith represents a number of the latent features extracted for the ith subsystem, with a calculation formula as follows:
MFeAlli=rank(Xi) (4);
wherein based on the calculation formula, all the latent features extracted from the input data Xi are expressed as:
Ti=[t1FeAlli,L, tmFeAllii,L, tMFeAllii] (5)
Wherein, Ti∈RN×MFeAlli represents a score matrix, which is an orthogonal mapping of the input data Xi in a direction of a load matrix Pi, wherein Pi is expressed by formula:
Pi=[p1FeAlli,L, pmFeAllii,L, pMFeAllii] (6);
wherein Pi∈RM×MFeAlli,
wherein the latent features extracted from the input data Xi are further expressed as:
Z
FeAll
i
=
T
i
=
X
i
⁢
P
i
⁢
=
[
z
1
FeAll
i
,
L
,
z
m
FeAll
i
i
,
L
,
z
M
FeAll
i
i
]
⁢
=
[
{
(
z
1
FeAll
i
)
n
}
n
=
1
N
,
L
,
{
(
z
m
FeAll
i
i
)
n
}
n
=
1
N
,
L
,
(
z
M
FeAll
i
i
)
n
}
n
=
1
N
]
;
⁢
=
{
(
z
FeAll
i
)
n
}
n
=
1
N
(
7
)
wherein ZFeAlli∈RN×MFeAlli,
wherein all the latent features are further expressed as:
ZFeAll=[ZFeAll1,L, ZFeAlli,L, ZFeAllI]={ZFeAlli}i=1I (8);
a feature vector corresponding to the mFeAllith load vector pmFeAllii is recorded as λmFeAllii, a contribution rate θmFeAllii corresponding to mFeAllith latent feature zmFeAllii is calculated using formula:
θ
m
FeAll
i
i
=
λ
m
FeAll
i
i
∑
m
FeAll
i
=
1
M
FeAll
i
⁢
λ
m
FeAll
i
i
×
100
;
(
9
)
wherein a threshold preset by experience is recorded as θContri, with a default value being 1, where the primary selection of the latent features is performed using rules:
ξ
m
FeAll
i
i
=
{
1
,
if
⁢
θ
m
FeAll
i
i
≥
θ
Contri
0
,
else
⁢
θ
m
FeAll
i
i
<
θ
Contri
;
(
10
)
wherein, ξmFeAllii represents a mark value denoting whether the mFeAllith latent feature is selected, and a value of 1 indicates that the latent feature is selected for the primary selection; wherein, the latent features of the primary selection for the ith subsystem are expressed as:
Z
FeSe
⁢
1
⁢
st
i
=
[
z
1
FeSe
⁢
1
⁢
st
i
,
L
,
z
m
FeSe
⁢
1
⁢
st
i
i
,
L
,
z
M
FeSe
⁢
1
⁢
st
i
i
]
=
[
{
(
z
1
FeSe
⁢
1
⁢
st
i
)
n
}
n
=
1
N
,
L
,
{
(
z
m
FeSe
⁢
1
⁢
st
i
i
)
n
}
n
=
1
N
,
L
,
{
(
z
M
FeSe
⁢
1
⁢
st
i
i
)
n
}
n
=
1
N
]
=
{
(
z
FeSe
⁢
1
⁢
st
i
)
n
}
n
=
1
N
;
(
11
)
wherein, all the latent features from the primary selection ZFeSe1st are expressed as:
ZFeSe1st=[ZFeSe1st1,L, ZFeSe1sti,L, ZFeSe1stI]={ZFeSe1sti}i=1I (12).

### Claim 3 (depends on 1)

The soft measurement method of dioxin emission concentration in municipal solid waste incineration process according to claim 1, wherein using the latent feature evaluation and reselection module further comprises the steps of:
for an ith subsystem, a mutual information (MI) value between each primary feature zmFeSe1stii and the DXN emission concentration is marked as ξMImFeSe1sti and calculated using formula:
ξ
MI
m
FeSe
⁢
1
⁢
st
i
=
∫
∫
p
prob
(
z
m
FeSe
⁢
1
⁢
st
i
,
y
)
⁢
log
⁡
(
p
prob
(
z
m
FeSe
⁢
1
⁢
st
i
,
y
)
p
prob
(
z
m
FeSe
⁢
1
⁢
st
i
)
⁢
p
p
⁢
r
⁢
o
⁢
b
(
y
)
)
⁢
d
⁡
(
z
m
FeSe
⁢
1
⁢
st
i
)
⁢
dy
=
H
⁡
(
y
)
-
H
⁡
(
y
⁢
❘
"\[LeftBracketingBar]"
z
m
FeSe
⁢
1
⁢
st
i
)
;
(
13
)
wherein, pprob(zmFeSe1sti) and pprob(y) represent a marginal probability density of zprob(zmFeSe1sti)and y, pprob(zmFeSe1sti,y) represents a joint probability density, H(y|zprob(zmFeSe1sti) represents a conditional entropy, and H(y) represents an information entropy;
adaptively determining a threshold of reselection of the latent features according to a prediction performance of the DXN emission concentration SEN soft measurement model, which further comprises the steps of:
calculating the upper limit of the threshold θContriUplimit, the lower limit of the threshold θContriDownlimit and a fixed step size θContriStep using formulas:
θ
Contri
Uplimit
=
min
⁡
(
max
⁡
(
ξ
MI
m
FeSe
⁢
1
⁢
st
i
)
,
L
,
max
⁡
(
ξ
MI
m
FeSe
⁢
1
⁢
st
i
)
,
L
,
max
⁡
(
ξ
MI
m
FeSe
⁢
1
⁢
st
i
)
)
;
(
14
)
θ
Contri
Downlimit
=
max
⁡
(
min
⁡
(
ξ
MI
m
FeSe
⁢
1
⁢
st
i
)
,
L
,
min
⁡
(
ξ
MI
m
FeSe
⁢
1
⁢
st
i
)
,
L
,
min
⁡
(
ξ
MI
m
FeSe
⁢
1
⁢
st
i
)
)
;
(
15
)
θ
Contri
Step
=
θ
Contri
Uplimit
-
θ
Contri
Downlimit
N
Contri
Step
;
(
16
)
wherein functions max(⋅) and min(⋅) represent taking maximum and minimum values respectively; NContriStep represents a number of candidate thresholds preset by experience, and a default value is 10; and
recording the determined threshold as θContri, and adaptively selecting between θContriUplimit and θContriDownlimit based on the prediction performance of the DXN emission concentration SEN soft measurement model;
using following rules to reselect the latent features of the primary selection:
ξ
m
FeSe
⁢
1
⁢
st
i
i
=
{
1
,
if
⁢
ξ
MI
m
FeSe
⁢
1
⁢
st
i
≥
θ
Contri
0
,
else
⁢
ξ
MI
m
FeSe
⁢
1
⁢
st
i
<
θ
Contri
;
(
17
)
wherein, ξmFeSe1stii represents a mark value of whether the mFeSe1stith latent feature is selected, and a value of 1 indicates that the latent feature is selected again;
expressing the reselected latent features for the ith subsystem as:
Z
FeSe
⁢
2
⁢
nd
i
=
[
z
1
FeSe
⁢
2
⁢
nd
i
,
L
,
z
m
FeSe
⁢
2
⁢
nd
i
i
,
L
,
z
M
FeSe
⁢
2
⁢
nd
i
i
]
=
[
{
(
z
1
FeSe
⁢
2
⁢
nd
i
)
n
}
n
=
1
N
,
L
,
{
z
m
FeSe
⁢
2
⁢
nd
i
i
)
n
}
n
=
1
N
,
L
,
{
(
z
M
FeSe
⁢
2
⁢
nd
i
i
)
n
}
n
=
1
N
]
=
{
(
z
FeSe
⁢
2
⁢
nd
i
)
n
}
n
=
1
N
;
and
(
18
)
expressing all the reselected latent features ZFeSe2nd as:
ZFeSe2nd=[ZFeSe2nd1,L, ZFeSe2ndi, L, ZFeSe2ndI]={ZFeSe2ndi}i=1I (19).

### Claim 4 (depends on 1)

The soft measurement method of dioxin emission concentration in the municipal solid waste incineration process according to claim 1, wherein the step of using the adaptive selective ensemble modeling module further comprises the steps of:
for an ith subsystem, describing a process of constructing a DXN emission concentration sub-model based on a reselection of the latent features ZFeSe2ndi and modeling hyperparameter pairs {Keri, Regi}, further comprising:
transforming a reselected latent feature {(zFeSe2ndi)n}n=1N into a high-dimensional feature space by mapping φ(⋅), and subsequently solving a following optimization problem:
{
min
W
i
,
h
i
O
LS
-
SVM
=
1
2
⁢
(
w
i
)
T
⁢
w
i
+
1
2
⁢
R
eg
i
⁢
∑
n
=
1
N
(
ζ
n
i
)
2
s
.
t
.
:
y
^
n
i
=
(
w
i
)
T
⁢
φ
⁢
(
(
z
FeSe
⁢
2
⁢
nd
i
)
n
)
+
b
i
+
ζ
n
i
;
(
20
)
wherein, wi represents a weight coefficient, bi represents an offset, and ζni is a prediction error of an nth sample;
using Lagrangian method to obtain formula:
L
i
(
w
i
,
b
i
,
ζ
i
,
β
i
)
=
1
2
⁢
(
w
i
)
T
⁢
w
i
+
1
2
⁢
∑
n
=
1
N
(
ζ
n
i
)
2
-
∑
n
=
1
N
β
n
i
[
(
w
i
)
T
⁢
φ
⁡
(
(
z
FeSe
⁢
2
⁢
nd
i
)
n
)
+
b
i
+
ζ
n
i
-
y
^
n
i
]
;
(
21
)
wherein, βi=[β1i,L, βni,L, βNi] represents a Lagrange operator vector, and ζi=[ζ1i,L, ζni,L, ζNi] represents a prediction error vector;
solving the formula:
∂
L
i
∂
w
i
=
0
,
∂
L
i
∂
b
i
=
0
,
∂
L
i
∂
ξ
i
=
0
,
∂
L
i
∂
β
i
=
0
,
;
(
22
)
expressing an adopted kernel function as follows:
Ωkeri(zFeSe2ndi,(zFeSe2ndi)n)=<φ(zFeSe2ndi)·φ((zFeSe2ndi)n)< (23);
transforming an LS-SVM problem into solving a linear equation system:
[
0
1
L
1
1
Ω
ker
i
(
(
z
FeSe
⁢
2
⁢
nd
i
)
1
,
(
z
Fe
⁢
Se
⁢
2
⁢
nd
i
)
1
)
+
1
R
eg
i
L
Ω
ker
i
⁢
(
(
z
FeSe
⁢
2
⁢
nd
i
)
1
,
(
z
Fe
⁢
Se
⁢
2
⁢
nd
i
)
N
)
M
M
M
M
1
Ω
ker
i
⁢
(
(
z
FeSe
⁢
2
⁢
nd
i
)
N
,
(
z
Fe
⁢
Se
⁢
2
⁢
nd
i
)
1
)
L
Ω
ker
i
⁢
(
(
z
FeSe
⁢
2
⁢
nd
i
)
N
)
,
(
z
Fe
⁢
Se
⁢
2
⁢
nd
i
)
N
)
)
+
1
R
eg
i
]
·
[
b
i
β
1
i
M
β
N
i
]
=
[
1
y
1
i
M
y
N
i
]
;
(
24
)
solving the linear equation system to obtain βi and bi;
expressing the DXN emission concentration sub-model based on the LS-SVM problem as:
y
ˆ
i
=
∑
n
=
1
N
β
n
i
·
Ω
ker
i
(
z
FeSe
⁢
2
⁢
nd
j
,
(
z
FeSe
⁢
2
⁢
nd
i
)
n
)
+
b
i
;
(
25
)
wherein the hyperparameter adaptive selection mechanism of the DXN emission concentration sub-model is implemented using the following two steps:
step 1: adopting a grid search strategy with a prediction performance of the DXN emission concentration sub-model as an objective function, and adaptively selecting an initial hyperparameter pair {(Kerinitial)i, (Reginitial)i} in a candidate hyperparameter matrix Mpara, wherein the candidate hyperparameter matrix Mpara is expressed as follows:
M
para
=
[
[
K
er
1
,
R
eg
1
]
L
[
K
er
1
,
R
eg
r
]
L
[
K
er
1
,
R
eg
R
]
M
M
M
M
M
[
K
er
k
,
R
eg
1
]
L
[
K
er
k
,
R
eg
r
]
L
[
K
er
k
,
R
eg
R
]
M
M
M
M
M
[
K
er
K
,
R
eg
1
]
L
[
K
er
K
,
R
eg
r
]
L
[
K
er
K
,
R
eg
R
]
]
;
(
26
)
where k=1,L,K K represent a number of candidate kernel parameters; r=1,L,R represents a number of candidate penalty parameters; [Kerk, Regr] represents a kth kernel parameter and rth penalty parameter, a formed hyperparameter pair also is a jth parameter pair in the candidate hyperparameter matrix Mpara, that is, there are Mparaj=[Kerk, Regr]; j=1,L,J, J=K×R represent a number of all hyperparameter pairs in the candidate hyperparameter matrix Mpara; wherein, a hyperparameter {(Kerinitial)i, (Reginitial)i} selected for a first time using the grid search strategy is an element in the candidate hyperparameter matrix Mpara, i.e., there is the {Kerinitial, Reginitial}∈Mpara;
step 2: based on the {(Kerinitial)i, (Reginitial)i} selected by the step 1, a new set of candidate hyperparameters is obtained using following formulas:
(
K
er
vector
)
i
=
(
K
er
initial
)
i
/
k
supara
down
:
k
supara
up
×
(
K
er
initial
)
i
-
(
K
er
initial
)
i
/
k
supara
down
N
ker
:
k
supara
up
×
(
K
er
initial
)
i
;
(
27
)
(
R
er
vector
)
i
=
(
R
eg
initial
)
i
/
k
supara
down
:
k
supara
up
×
(
R
eg
initial
)
i
-
(
R
eg
initial
)
i
/
k
supara
down
N
ker
:
k
supara
up
×
(
R
eg
initial
)
i
;
(
28
)
wherein, (Kervector)i and (Regvector)i represent new candidate hyperparameter sets, corresponding to kernel parameter vector and penalty parameter vector, respectively; Nker and Nreg represent a number of new hyperparameters sets based on experience; ksuparadown and ksuparaup represent hyperparameter shrinkage and expansion factor based on experience, a default value is 10;
by adopting the grid search strategy, adaptively obtaining the hyperparameter pair of an ith sub-model {Keri, Regi};
performing the step 1 and the step 2 for all subsystems, wherein a set of sub-model prediction outputs is expressed as:
Y
^
=
[
y
^
1
,
L
,
y
^
i
,
L
,
y
^
f
]
=
{
y
^
i
}
i
=
1
l
=
[
f
1
(
K
er
1
,
R
eg
1
,
Z
FeSe
⁢
2
⁢
nd
1
)
,
L
,
f
1
(
K
er
i
,
R
eg
i
,
Z
FeSe
⁢
2
⁢
nd
i
)
,
L
,
f
1
(
K
er
l
,
R
eg
l
,
Z
FeSe
⁢
2
⁢
nd
l
)
]
=
{
f
i
(
K
er
i
,
R
eg
i
,
Z
FeSe
⁢
2
⁢
nd
i
)
}
i
=
1
f
(
29
)
wherein, fi(⋅) represents the ith sub-model;
by combining an optimization selection algorithm based on the branch and bound (BB) and the prediction error information entropy weighting algorithm, adaptively optimizing the sub-models and calculating the weight coefficients, wherein given a candidate sub-model and weighting algorithm, an optimal sub-model selection is similar to a weighted optimal feature selection;
for a limited number of candidate sub-models, by running optimization and weighting algorithms multiple times, obtaining SEN models with ensemble sizes of 2 to (I−1); and
ranking the optimized SEN models and selecting a best prediction performance model as a final DXN soft measurement model;
wherein assuming an ensemble size of the final DXN soft measurement model is Isel, a predicted output value ŷ is calculated by formula:
y
ˆ
=
∑
i
se
⁢
1
=
1
I
se
⁢
1
W
i
se
⁢
1
⁢
y
ˆ
i
se
⁢
1
=
∑
i
se
⁢
1
=
1
I
se
⁢
1
w
i
se
⁢
1
·
f
i
sel
(
K
er
i
se
⁢
1
,
R
eg
i
se
⁢
1
,
Z
FeSe
⁢
2
⁢
nd
i
se
⁢
1
)
.
(
30
)