CN103245481B - Frequency conversion technology based detection method for air resistance characteristic of large-scale variable-load heat exchanger - Google Patents
Frequency conversion technology based detection method for air resistance characteristic of large-scale variable-load heat exchanger Download PDFInfo
- Publication number
- CN103245481B CN103245481B CN201310164423.2A CN201310164423A CN103245481B CN 103245481 B CN103245481 B CN 103245481B CN 201310164423 A CN201310164423 A CN 201310164423A CN 103245481 B CN103245481 B CN 103245481B
- Authority
- CN
- China
- Prior art keywords
- frequency
- output
- air quantity
- value
- input
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Feedback Control In General (AREA)
Abstract
The invention discloses a frequency conversion technology based detection method for the air resistance characteristic of a large-scale variable-load heat exchanger. The method comprises the specific steps: speculating an initial frequency predicted value f0 of a frequency converter under standard wind volume according to history data in combination with a nerve network; calculating out three parameters K, T and t of an open-loop transfer function K/(1+Tps) *e<-tau s> between frequency converter frequency and the pipeline wind volume through a least squares algorithm through acquiring the wind volume and frequency variations of the frequency converter with frequency changing from 0 to f0; then taking IATE as an optimal index and PID+ K/(1+Tps) *e<-tau s> as the open-loop transfer function, utilizing the closed-loop transfer function as a restraint equation, and performing optimization solution by a nonlinear optimization solver to obtain optimal values Kji, Ti and Td of a PID controller; and obtaining a corrected frequency value f'0 according to approximate linear feature of the wind volume and the frequency at the working point, and regulating the frequency of the frequency converter to f'0. According to the invention, the detection method avoids troubles brought by manually adjusting wind volume, is energy-saving and time-saving, is convenient and quick, and effectively improves the production efficiency.
Description
Technical field
The invention belongs to control technology field, relate to data modeling and the process control of industrial process control field, particularly relate to a kind of detection method of the varying load large heat exchanger vapour lock characteristic based on converter technique.
Background technology
Vapour lock characteristic is the important indicator characterizing large heat exchanger flowing property, is also one of the important indicator that must test before dispatching from the factory of large-scale plate-fin heat exchanger product.Mainly comprise two aspects: one is for certain heat exchanger channel, in standard temperature, normal pressure and normal flow situation, the pressure loss at these heat exchanger channel two ends; Another one is the friction factor of heat exchanger channel in the case.Because test result is different under the condition such as different temperatures, pressure, lack comparability.Therefore required standard situation is normal atmosphere pressure, a standard zero degrees celsius is the status of criterion, and the air quantity now recorded is standard air quantity, and the air quantity recorded under other pressure and temperatures needs to be converted into standard air quantity.Past general orifice flowmeter that adopts in measuring process, by the vapour lock parameter of heat interchanger in the certain air quantity situation of measurement, then converts and obtains the vapour lock characteristic of heat interchanger under normal conditions.Above method is adopted to have obvious shortcoming, one is that the measurement accuracy of orifice flowmeter is poor, the control of air quantity adopts the mode of manual control valve in addition, accurately cannot control air quantity after the conversion, 3rd is that employing orifice plate is measured and manual control is not energy-conservation, inaccurate, subsequent artefacts's calculated amount is large, because each heat interchanger test duration is long, work efficiency is lower.
Adopt the air quantity of Frequency Converter Control blower fan, adopt nozzle measuring system can resolution system power saving, the precision that nozzle measures air quantity be much higher than orifice flowmeter.Owing to cannot ensure in test process that current working is in standard condition, need the temperature and pressure of heat exchanger can convert according to the equation of gas state, conversion method is:
, here
with
represent standard temperature and pressure (STP),
for actual measurement air quantity,
with
for observed pressure and temperature,
it is then the standard air quantity after conversion.For the purpose of accurately, in heat exchanger channel performance test process, air quantity is controlled fast the standard air quantity after conversion, namely to require actual measurement air quantity, by converting standard air quantity to, then to control this standard air quantity and make it equal setting air quantity.Consider efficiency and the power conservation requirement of test, measurement mechanism adopts converter technique to carry out air quantity adjustment.
The different lane testing of heat interchanger also has a feature to be that test has intermittence: after namely testing a passage, before new passage is received measurement mechanism, it is zero that measurement mechanism air channel volume requires.Because the model of each heat interchanger is different with type, in test process, how to determine the original frequency of frequency converter, and how according to the characteristic of system under test (SUT), adopt suitable control method by air quantity stability contorting in the setting value corresponding with standard condition, heat exchanger measuring accuracy and rapidity have material impact.Because the hysteresis quality of measuring system, different passage part throttle characteristics are widely different and the nonlinear characteristic of frequency variation signal change, system is caused to be difficult to realize automatic Quick Measurement.Therefore be necessary to study a kind of method that can realize observing and controlling air quantity fast automatically in the case, make it the industrial requirements meeting the characteristic test of plate type finned heat exchanger vapour lock.
Summary of the invention
The object of the invention is for the deficiencies in the prior art, a kind of detection method of the varying load large heat exchanger vapour lock characteristic based on converter technique is provided.
The technical solution used in the present invention is as follows:
Step (1) gathers dissimilar plate type finned heat exchanger design parameter and operational factor, sets up the real-time data base comprising design of heat exchanger parameter and operational factor, and the collection of design parameter is obtained by the database in heat exchanger system detection platform.
Described design of heat exchanger parameter comprises tunnel name, design air flow, the design vapour lock of heat interchanger, and operational factor comprises the actual vapour lock of heat interchanger, detected temperatures, air pressure and actual frequency converter frequency, and the preparation method of these parameters is mature technology.
Based on historical test data, adopt the neural network model between neural network design standards air quantity, design vapour lock and actual vapour lock, actual frequency converter frequency, under predicting various heat exchange device channels designs standard air quantity with this, blower fan is in order to reach the original frequency predicted value of this air quantity
, concrete implementation step is as follows:
First, extract the historical data in each heat exchanger channel situation, set up input amendment and output sample collection, input amendment comprises design air flow, design vapour lock, the detected temperatures of heat interchanger, output sample is that actual vapour lock and heat interchanger are reaching the actual frequency converter frequency under design air flow, then neural network is adopted to train these historical datas, neural network structure comprises input layer, hidden layer (middle layer) and output layer, and the neuron number scope of hidden layer and output layer is respectively 4 ~ 10 and 2 ~ 6.In training process using design air flow, design vapour lock and detected temperatures as input, using actual frequency converter frequency and actual vapour lock as output, obtain neural network model by neural network learning, each parameter in neural network model is defined as follows:
Input layer unit input vector is
, object vector
; Hidden layer unit input vector
, output vector
; Output layer unit input vector
, output vector
, k=1,2 ..., m represents sample data number; Input layer is to the connection weight of hidden layer
, i=1,2 ...., n; J=1,2 ... p; Hidden layer is to the connection weight of output layer
, t=1,2 ... q; The output threshold value of each unit of hidden layer
, the output threshold value of each unit of output layer
.
The learning process step of neural network model is as follows:
A) the connection weights and threshold of each layer of initialization, to each connection weight
,
with output threshold value
,
give the random value in interval (-1,1).
B) input amendment and output sample is chosen.
C) with input amendment, connection weight, input threshold value and the output exporting threshold calculations hidden layer and each unit of output layer.
Wherein,
represent the input value of hidden layer unit,
represent the output valve of hidden layer unit,
represent output layer unit input value,
represent the output valve of output layer unit.
D) output layer each unit vague generalization error is calculated
, then utilize hidden layer to the connection weight of output layer
, hidden layer output vector
, output layer each unit vague generalization error
calculate the vague generalization error of each unit of hidden layer
, computing formula is as follows:
;
;
E) the vague generalization error of each unit of output layer is utilized
the connection weight of hidden layer to output layer is revised with the output valve of each unit of hidden layer
, export threshold value
:
;
;
Utilize the vague generalization error of each unit of hidden layer equally
the connection weight of input layer to hidden layer is revised with the input of input layer
, export threshold value
:
,
,
。
represent current connection weight
,
represent revised connection weight;
represent current output threshold value
,
represent revised respective threshold.
represent current connection weight
,
represent revised connection weight,
represent current threshold value,
represent revised threshold value, N=1,2 ..., NN, wherein NN represents the study iterations of setting.
F) choose next input amendment and output sample, turn back to step c), until m training sample training is complete.
G) cumulative errors of all samples are calculated
, cumulative errors account form is
, wherein, q represents output layer unit number, and m represents sample size.If sample cumulative errors
be less than preset value
, or current study iterations is greater than the study iterations of setting, and so learning training terminates.Otherwise again choose sample input and target output, then turn back to step c).
After learning process terminates, by the weights and threshold of neural network each several part obtained, foundation can reflect the neural network model of input and output, by given input information (design air flow, design vapour lock and detected temperatures), obtains the predicted value of frequency converter original frequency.
Step (2): according to the design air flow of real exchanger passage, design vapour lock and detected temperatures, the neural network model utilizing step (1) to set up obtains the original frequency predicted value of this heat interchanger channel condition low-converter
.
Step (3): the original frequency predicted value obtained according to step (2)
, blower fan frequency is adjusted to from 0
80%, be designated as
, when blower fan frequency reaches
and after stable operation 10 ~ 20s, record and now survey standard air quantity and be designated as
; By blower fan frequency adjustment be
100%, be designated as
, when blower fan frequency reaches
and after stablizing 10 ~ 20s, record and now survey standard air quantity and be designated as
; And blower fan frequency to
and stablize in 10 ~ 20s process, record actual measurement standard air quantity, the actual measurement vapour lock and blower fan frequency of heat exchanger channel with cycle 0.5 ~ 1s, then, the actual measurement standard air quantity that obtains will be recorded in this process and blower fan frequency deducts respectively
with
, obtain air quantity changing value
, frequency change
, wherein
with
represent of record respectively
individual air quantity change and frequency change.
Step (4): according to pipeline gas flowing Controlling model feature, set up the transport function between frequency and air quantity.Air quantity control system transport function is that one order inertia adds delay link, therefore the transport function between frequency and standard air quantity can be set to
, wherein
represent Open loop gain cofficient, time constant and time delay respectively,
for plural number.According to the frequency change that step (3) obtains
, will
give positive initial value respectively, pass through transport function
calculate at sample point frequency change
output under effect
, then with
for target, employing least-squares algorithm simulates three parameters in transport function
, obtain the concrete transport function of this heat interchanger passage.
Step (5): establish the concrete transport function of heat interchanger passage in step (4) after, is adjusted to preset standard air quantity by air quantity, the parameter of the PID that adjusts in order to ensure fast and stable.With IATE integral performance for optimum index, with parameter in PID controller
,
,
for variable, with PID+
as open-loop transfer function, take closed loop transfer function, as equation of constraint, with
,
,
being on the occasion of being variable bound, adopting nonlinear optimization solution technique to be optimized and solving, obtaining optimum PID controller parameter
,
,
value, wherein
,
,
represent ratio, integration and differentiation parameter respectively.
Step (6): obtain original frequency predicted value according to step (2)
with the frequency that step (3) obtains
, calculate linear relationship parameter approximate between air quantity and frequency, and determine the modified value of original frequency
, by blower fan frequency adjustment to modified value
, and stable
second.According to the principle of similitude, frequency
with air quantity
there is approximation relation is
, wherein
with
for needing the linear relationship parameter obtained.By original frequency predicted value
with
and the air quantity of correspondence, bring above relational expression into and obtain parameter
with
.Obtaining
with
after, according to
relation, order
for established standards air quantity, obtain the modified value of original frequency under established standards air quantity
.
Step (7): blower fan frequency setting is existed
frequency point, equifrequent reaches setting value and stablizes
after second.Then employing incremental timestamp device is by the actual measurement Boiler pressure control of heat exchanger channel at established standards air quantity place, obtains the vapour lock characteristic of heat exchanger channel under established standards air quantity.The output form of PID is:
;
here three parameters of incremental timestamp device
,
,
for obtaining obtaining those three parameters in step (5).
represent the sampling period,
represent the error between the setting value of corresponding step number and value of feedback,
,
,
represent ongoing frequency value, frequency change and next step frequency values respectively.By incremental timestamp, actual measurement airflow value can be controlled automatically in established standards airflow value, control accuracy is within 0.5%.The vapour lock characteristic now obtained by measurement is exactly the vapour lock characteristic under established standards air quantity, by the comparison between actual measurement vapour lock and design vapour lock, can obtain heat interchanger vapour lock characteristic performance index.
Beneficial effect of the present invention:
The present invention not only replaces manual inspection heat interchanger vapour lock Characterization method in the past completely, has the advantages that automatic detection, automatically calculating and measuring accuracy are high.The robotization of this method and rapidity are very good, can adapt to different loads and air quantity requirement, while raising measuring accuracy, reduction labor workload, greatly can accelerate heat interchanger test lot number.
Embodiment
Based on the detection method of the varying load large heat exchanger vapour lock characteristic of converter technique, to the different channel characteristic of various heat exchange device, carry out automatic Quick Measurement vapour lock characteristic.First the present invention carries out data mining according to historical test data, adopt the relation between neural network design of heat exchanger parameter and original frequency, then the control characteristic of heat exchanger channel is obtained by the control characteristic parameter automatic Identification of various heat exchange device passage, optimal performance index is adopted to adjust the pid parameter of self-actuated controller, then the standard value after adopting incremental timestamp device to be changed by air channel volume controls to established standards air quantity place, obtains its vapour lock characteristic by the vapour lock value detected now.
Concrete steps of the present invention are as follows:
Step (1) gathers dissimilar plate type finned heat exchanger design parameter and operational factor, sets up the real-time data base comprising design of heat exchanger parameter and operational factor, and the collection of design parameter is obtained by the database in heat exchanger system detection platform.
Described design of heat exchanger parameter comprises tunnel name, design air flow, the design vapour lock of heat interchanger, and operational factor comprises the actual vapour lock of heat interchanger, detected temperatures, air pressure and actual frequency converter frequency, and the preparation method of these parameters is mature technology.
Based on historical test data, adopt the neural network model between neural network design standards air quantity, design vapour lock and actual vapour lock, actual frequency converter frequency, under predicting various heat exchange device channels designs standard air quantity with this, blower fan is in order to reach the original frequency predicted value of this air quantity
, concrete implementation step is as follows:
First, extract the historical data in each heat exchanger channel situation, set up input amendment and output sample collection, input amendment comprises design air flow, design vapour lock, the detected temperatures of heat interchanger, output sample is that actual vapour lock and heat interchanger are reaching the actual frequency converter frequency under design air flow, then neural network is adopted to train these historical datas, neural network structure comprises input layer, hidden layer (middle layer) and output layer, and the neuron number scope of hidden layer and output layer is respectively 4 ~ 10 and 2 ~ 6.In training process using design air flow, design vapour lock and detected temperatures as input, using actual frequency converter frequency and actual vapour lock as output, obtain neural network model by neural network learning, each parameter in neural network model is defined as follows:
Input layer unit input vector is
, object vector
; Hidden layer unit input vector
, output vector
; Output layer unit input vector
, output vector
, k=1,2 ..., m represents sample data number; Input layer is to the connection weight of hidden layer
, i=1,2 ...., n; J=1,2 ... p; Hidden layer is to the connection weight of output layer
, t=1,2 ... q; The output threshold value of each unit of hidden layer
, the output threshold value of each unit of output layer
.
The learning process step of neural network model is as follows:
H) the connection weights and threshold of each layer of initialization, to each connection weight
,
with output threshold value
,
give the random value in interval (-1,1).
I) input amendment and output sample is chosen.
J) with input amendment, connection weight, input threshold value and the output exporting threshold calculations hidden layer and each unit of output layer.
Wherein,
represent the input value of hidden layer unit,
represent the output valve of hidden layer unit,
represent output layer unit input value,
represent the output valve of output layer unit.
K) output layer each unit vague generalization error is calculated
, then utilize hidden layer to the connection weight of output layer
, hidden layer output vector
, output layer each unit vague generalization error
calculate the vague generalization error of each unit of hidden layer
, computing formula is as follows:
;
;
L) the vague generalization error of each unit of output layer is utilized
the connection weight of hidden layer to output layer is revised with the output valve of each unit of hidden layer
, export threshold value
:
;
;
Utilize the vague generalization error of each unit of hidden layer equally
the connection weight of input layer to hidden layer is revised with the input of input layer
, export threshold value
:
,
,
。
represent current connection weight
,
represent revised connection weight;
represent current output threshold value
,
represent revised respective threshold.
represent current connection weight
,
represent revised connection weight,
represent current threshold value,
represent revised threshold value, N=1,2 ..., NN, wherein NN represents the study iterations of setting.
M) choose next input amendment and output sample, turn back to step c), until m training sample training is complete.
N) cumulative errors of all samples are calculated
, cumulative errors account form is
, wherein, q represents output layer unit number, and m represents sample size.If sample cumulative errors
be less than preset value
, or current study iterations is greater than the study iterations of setting, and so learning training terminates.Otherwise again choose sample input and target output, then turn back to step c).
After learning process terminates, by the weights and threshold of neural network each several part obtained, foundation can reflect the neural network model of input and output, by given input information (design air flow, design vapour lock and detected temperatures), obtains the predicted value of frequency converter original frequency.
Step (2): according to the design air flow of real exchanger passage, design vapour lock and detected temperatures, the neural network model utilizing step (1) to set up obtains the original frequency predicted value of this heat interchanger channel condition low-converter
.
Step (3): the original frequency predicted value obtained according to step (2)
, blower fan frequency is adjusted to from 0
80%, be designated as
, when blower fan frequency reaches
and after stable operation 10 ~ 20s, record and now survey standard air quantity and be designated as
; By blower fan frequency adjustment be
100%, be designated as
, when blower fan frequency reaches
and after stablizing 10 ~ 20s, record and now survey standard air quantity and be designated as
; And blower fan frequency from
arrive
and stablize in 10 ~ 20s process, record actual measurement standard air quantity, the actual measurement vapour lock and blower fan frequency of heat exchanger channel with cycle 0.5 ~ 1s, then, the actual measurement standard air quantity that obtains will be recorded in this process and blower fan frequency deducts respectively
with
, obtain air quantity changing value
, frequency change
, wherein
with
represent of record respectively
individual air quantity change and frequency change.
Step (4): according to pipeline gas flowing Controlling model feature, set up the transport function between frequency and air quantity.Air quantity control system transport function is that one order inertia adds delay link, therefore the transport function between frequency and standard air quantity can be set to
, wherein
represent Open loop gain cofficient, time constant and time delay respectively,
for plural number.According to the frequency change that step (3) obtains
, will
give positive initial value respectively, pass through transport function
calculate at sample point frequency change
output under effect
, then with
for target, employing least-squares algorithm simulates three parameters in transport function
, obtain the concrete transport function of this heat interchanger passage.
Step (5): establish the concrete transport function of heat interchanger passage in step (4) after, is adjusted to preset standard air quantity by air quantity, the parameter of the PID that adjusts in order to ensure fast and stable.With IATE integral performance for optimum index, with parameter in PID controller
,
,
for variable, with PID+
as open-loop transfer function, take closed loop transfer function, as equation of constraint, with
,
,
being on the occasion of being variable bound, adopting nonlinear optimization solution technique to be optimized and solving, obtaining optimum PID controller parameter
,
,
value, wherein
,
,
represent ratio, integration and differentiation parameter respectively.
Step (6): obtain original frequency predicted value according to step (2)
with the frequency that step (3) obtains
, calculate linear relationship parameter approximate between air quantity and frequency, and determine the modified value of original frequency
, by blower fan frequency adjustment to modified value
, and stable
second.According to the principle of similitude, frequency
with air quantity
there is approximation relation is
, wherein
with
for needing the linear relationship parameter obtained.By original frequency predicted value
with
and the air quantity of correspondence, bring above relational expression into and obtain parameter
with
.Obtaining
with
after, according to
relation, order
for established standards air quantity, obtain the modified value of original frequency under established standards air quantity
.
Step (7): blower fan frequency setting is existed
frequency point, equifrequent reaches setting value and stablizes
after second.Then employing incremental timestamp device is by the actual measurement Boiler pressure control of heat exchanger channel at established standards air quantity place, obtains the vapour lock characteristic of heat exchanger channel under established standards air quantity.The output form of PID is:
;
here three parameters of incremental timestamp device
,
,
for obtaining obtaining those three parameters in step (5).
represent the sampling period,
represent the error between the setting value of corresponding step number and value of feedback,
,
,
represent ongoing frequency value, frequency change and next step frequency values respectively.By incremental timestamp, actual measurement airflow value can be controlled automatically in established standards airflow value, control accuracy is within 0.5%.The vapour lock characteristic now obtained by measurement is exactly the vapour lock characteristic under established standards air quantity, by the comparison between actual measurement vapour lock and design vapour lock, can obtain heat interchanger vapour lock characteristic performance index.
Claims (1)
1., based on the detection method of the varying load large heat exchanger vapour lock characteristic of converter technique, it is characterized in that the method comprises the steps:
Step (1) gathers dissimilar plate type finned heat exchanger design parameter and operational factor, sets up the real-time data base comprising design of heat exchanger parameter and operational factor, and the collection of design parameter is obtained by the database in heat exchanger system detection platform;
Described design of heat exchanger parameter comprises tunnel name, design air flow, the design vapour lock of heat interchanger, and operational factor comprises the actual vapour lock of heat interchanger, detected temperatures, air pressure and actual frequency converter frequency;
Based on historical test data, adopt the neural network model between neural network design standards air quantity, design vapour lock and actual vapour lock, actual frequency converter frequency, under predicting various heat exchange device channels designs standard air quantity with this, blower fan is in order to reach the original frequency predicted value f of this air quantity
0, concrete implementation step is as follows:
First, extract the historical data in each heat exchanger channel situation, set up input amendment and output sample collection, input amendment comprises design air flow, design vapour lock, the detected temperatures of heat interchanger, output sample is that actual vapour lock and heat interchanger are reaching the actual frequency converter frequency under design air flow, then adopts neural network to train these historical datas; Neural network structure comprises input layer, hidden layer and output layer, and the neuron number scope of hidden layer and output layer is respectively 4 ~ 10 and 2 ~ 6; In training process using design air flow, design vapour lock and detected temperatures as input, using actual frequency converter frequency and actual vapour lock as output, obtain neural network model by neural network learning, each parameter in neural network model is defined as follows:
Input layer unit input vector is P
k=(a
1, a
2..., a
n), object vector T
k=(d
1, d
2..., d
n); Hidden layer unit input vector S
k=(s
1, s
2..., s
p), output vector B
k=(b
1, b
2..., b
p); Output layer unit input vector L
k=(l
1, l
2..., l
q), output vector C
k=(c
1, c
2..., c
q), k=1,2 ..., m represents sample data number; Input layer is to the connection weight w of hidden layer
ij, i=1,2 ..., n; J=1,2 ... p; Hidden layer is to the connection weight v of output layer
jt, t=1,2 ... q; The output threshold value θ of each unit of hidden layer
j, the output threshold value y of each unit of output layer
t;
The learning process step of neural network model is as follows:
A) the connection weights and threshold of each layer of initialization, to each connection weight w
ij, v
jtwith output threshold value θ
j, y
tgive the random value in interval (-1,1);
B) input amendment and output sample is chosen;
C) with input amendment, connection weight, input threshold value and the output exporting threshold calculations hidden layer and each unit of output layer;
Wherein, s
jrepresent the input value of hidden layer unit, b
jrepresent the output valve of hidden layer unit, l
trepresent the input value of output layer unit, c
trepresent the output valve of output layer unit;
D) output layer each unit vague generalization error is calculated
then utilize hidden layer to the connection weight v of output layer
jt, hidden layer output vector B
k=(b
1, b
2..., b
p), output layer each unit vague generalization error
calculate the vague generalization error of each unit of hidden layer
computing formula is as follows:
E) the vague generalization error of each unit of output layer is utilized
the connection weight v of hidden layer to output layer is revised with the output valve of each unit of hidden layer
jt, export threshold value y
t:
Equally, the vague generalization error of each unit of hidden layer is utilized
the connection weight w of input layer to hidden layer is revised with the input of input layer
ij, export threshold value θ
j:
Wherein, v
jt(N) current connection weight v is represented
jt, v
jt(N+1) revised connection weight is represented; y
t(N) current output threshold value y is represented
t, y
t(N+1) revised respective threshold is represented; w
ij(N) current connection weight w is represented
ij, w
ij(N+1) revised connection weight is represented, θ
j(N) current threshold value is represented, θ
j(N+1) revised threshold value is represented, N=1,2 ..., NN, wherein NN represents the study iterations of setting;
F) choose next input amendment and output sample, turn back to c), until m training sample training is complete;
G) calculate the cumulative errors E of all samples, cumulative errors account form is
wherein, q represents output layer unit number, and m represents sample size; If sample cumulative errors E is less than preset value ε, or current study iterations is greater than the study iterations of setting, and so learning training terminates; Otherwise again choose sample input and target output, then turn back to c);
After learning process terminates, by the weights and threshold of neural network each several part obtained, set up the neural network model that can reflect input and output, by given input amendment information, obtain the predicted value of frequency converter original frequency;
Step (2) is according to the design air flow of real exchanger passage, design vapour lock and detected temperatures, and the neural network model utilizing step (1) to set up obtains the original frequency predicted value f of this heat interchanger channel condition low-converter
0;
The original frequency predicted value f that step (3) obtains according to step (2)
0, blower fan frequency is adjusted to f from 0
080%, be designated as f
1, when blower fan frequency reaches f
1and after stable operation 10 ~ 20s, record and now survey standard air quantity and be designated as Q
w1; Be f by blower fan frequency adjustment
0100%, be designated as f
2, when blower fan frequency reaches f
2and after stablizing 10 ~ 20s, record and now survey standard air quantity and be designated as Q
w; And in blower fan frequency from f
1to f
2and stablize in 10 ~ 20s process, record actual measurement standard air quantity, the actual measurement vapour lock and blower fan frequency of heat exchanger channel with cycle 0.5 ~ 1s, then, the actual measurement standard air quantity that obtains will be recorded in this process and blower fan frequency deducts Q respectively
w1and f
1, obtain air quantity changing value
frequency change u=[u
1, u
2..., u
cc], wherein
and u
ccrepresent cc air quantity change and the frequency change of record respectively;
Step (4), according to pipeline gas flowing Controlling model feature, sets up the transport function between frequency and air quantity; Air quantity control system transport function is that one order inertia adds delay link, therefore the transport function between frequency and standard air quantity can be set to K
ze
-τ s/ (1+T
ts), wherein K
z, T
t, τ represents Open loop gain cofficient, time constant and time delay respectively, s is plural number; According to the frequency change u=[u that step (3) obtains
1, u
2..., u
cc], by K
z, T
t, τ gives positive initial value, respectively by transport function K
ze
-τ s/ (1+T
ts) the output Q=(Q under sample point frequency change u effect is calculated
1, Q
2..., Q
cc), then with
for target, employing least-squares algorithm simulates three parameter K in transport function
z, T
t, τ, obtain the concrete transport function of this heat interchanger passage;
Air quantity, after step (4) establishes the concrete transport function of heat interchanger passage, is adjusted to preset standard air quantity, the parameter of the PID that adjusts in order to ensure fast and stable by step (5); With IATE integral performance for optimum index, with parameter K in PID controller
p, T
i, T
dfor variable, with PID+K
ze
-τ s/ (1+T
ts) as open-loop transfer function, be equation of constraint with closed loop transfer function, with K
p, T
i, T
dbeing on the occasion of being variable bound, adopting nonlinear optimization solution technique to be optimized and solving, obtaining optimum PID controller parameter K
p, T
i, T
dvalue, wherein K
p, T
i, T
drepresent ratio, integration and differentiation parameter respectively;
The original frequency predicted value f that step (6) obtains according to step (2)
0with the frequency f that step (3) obtains
1, calculate linear relationship parameter approximate between air quantity and frequency, and determine the modified value f ' of original frequency
0, by blower fan frequency adjustment to modified value f '
0, and stablize τ second; According to the principle of similitude, it is Q ≈ dd that frequency f and air quantity Q exist approximation relation
1f+dd
2, wherein dd
1and dd
2for needing the linear relationship parameter obtained; By original frequency predicted value f
0and f
1and the air quantity of correspondence, substitute into above relational expression and obtain parameter d d
1and dd
2; Obtaining dd
1and dd
2after, according to f=(Q-dd
2)/dd
1relation, make Q be established standards air quantity, obtain the modified value f ' of original frequency under established standards air quantity
0;
Step (7) by blower fan frequency setting at f '
0frequency point, equifrequent reaches setting value and after stablizing τ second; Then employing incremental timestamp device is by the actual measurement Boiler pressure control of heat exchanger channel at established standards air quantity place, obtains the vapour lock characteristic of heat exchanger channel under established standards air quantity; The output form of PID is:
Uu (kk+1)=uu (kk)+Δ u
kkhere three parameter K of incremental timestamp device
p, T
i, T
dfor obtaining obtaining those three parameters in step (5); T represents the sampling period, e (kk), e (kk-1), e (kk-2) represent the error between the setting value of corresponding step number and value of feedback, and uu (kk), Δ uu (kk), uu (kk+1) represent ongoing frequency value, frequency change and next step frequency values respectively; By incremental timestamp, actual measurement airflow value can be controlled automatically in established standards airflow value, control accuracy is within 0.5%; The vapour lock characteristic now obtained by measurement is exactly the vapour lock characteristic under established standards air quantity, by the comparison between actual measurement vapour lock and design vapour lock, can obtain heat interchanger vapour lock characteristic performance index.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310164423.2A CN103245481B (en) | 2013-05-07 | 2013-05-07 | Frequency conversion technology based detection method for air resistance characteristic of large-scale variable-load heat exchanger |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310164423.2A CN103245481B (en) | 2013-05-07 | 2013-05-07 | Frequency conversion technology based detection method for air resistance characteristic of large-scale variable-load heat exchanger |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103245481A CN103245481A (en) | 2013-08-14 |
CN103245481B true CN103245481B (en) | 2015-07-15 |
Family
ID=48925149
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201310164423.2A Active CN103245481B (en) | 2013-05-07 | 2013-05-07 | Frequency conversion technology based detection method for air resistance characteristic of large-scale variable-load heat exchanger |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103245481B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104569625B (en) * | 2015-01-20 | 2015-11-04 | 中国人民解放军国防科学技术大学 | A kind of large-scale antenna directional diagram measuring method based on rotatable auxiliary antenna |
CN105929683B (en) * | 2016-06-23 | 2019-01-18 | 东南大学 | A kind of differential is adjustable PID controller parameter engineering turning model and method |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6460359B1 (en) * | 1998-05-26 | 2002-10-08 | Atlas Copco Airpower, Nv | Method and device for cool-drying |
CN101392939A (en) * | 2008-11-18 | 2009-03-25 | 天津大学 | Nonlinear prediction and control method for independence energy supply temperature of buildings |
CN102096373A (en) * | 2010-12-07 | 2011-06-15 | 昆明理工大学 | Microwave drying PID (proportion integration differentiation) control method based on increment improved BP (back propagation) neural network |
CN102645315A (en) * | 2012-04-28 | 2012-08-22 | 杭州电子科技大学 | Automatic, fast and accurate detection method for air resistance characteristics of large heat exchanger |
CN102680037A (en) * | 2012-06-06 | 2012-09-19 | 上海华东电脑***工程有限公司 | Air quantity differential pressure calibration method applied to liquid cooling type frame |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPS59205523A (en) * | 1983-05-09 | 1984-11-21 | Omron Tateisi Electronics Co | Combustion controlling device |
-
2013
- 2013-05-07 CN CN201310164423.2A patent/CN103245481B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6460359B1 (en) * | 1998-05-26 | 2002-10-08 | Atlas Copco Airpower, Nv | Method and device for cool-drying |
CN101392939A (en) * | 2008-11-18 | 2009-03-25 | 天津大学 | Nonlinear prediction and control method for independence energy supply temperature of buildings |
CN102096373A (en) * | 2010-12-07 | 2011-06-15 | 昆明理工大学 | Microwave drying PID (proportion integration differentiation) control method based on increment improved BP (back propagation) neural network |
CN102645315A (en) * | 2012-04-28 | 2012-08-22 | 杭州电子科技大学 | Automatic, fast and accurate detection method for air resistance characteristics of large heat exchanger |
CN102680037A (en) * | 2012-06-06 | 2012-09-19 | 上海华东电脑***工程有限公司 | Air quantity differential pressure calibration method applied to liquid cooling type frame |
Non-Patent Citations (2)
Title |
---|
大型板翅式换热器气阻特性测控***的开发;李相发等;《第30届中国控制会议》;20111231;第136-141页 * |
板翅式换热器翅片表面流动特性测试***;徐丁等;《杭州电子科技大学学报》;20100831;第30卷(第4期);第5769-5774页 * |
Also Published As
Publication number | Publication date |
---|---|
CN103245481A (en) | 2013-08-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112987566B (en) | Aerodynamic-thermal supercoiled nonlinear fractional order sliding-mode model-free control method | |
CN109583585B (en) | Construction method of power station boiler wall temperature prediction neural network model | |
CN104533701B (en) | A kind of automatic setting method of Turbine Governor System control parameter | |
CN104750131B (en) | A kind of fluidized-bed temperature control method based on parameter identification | |
CN111829003B (en) | Power plant combustion control system and control method | |
CN101216715B (en) | PID control temperature instrument using nerve cell network adjustment parameter and its control method | |
CN107272412A (en) | A kind of identifying approach of intermittent wind tunnel flow field control | |
CN104748807A (en) | Online power station main steam flow calculation method based on flow correction | |
CN101887267A (en) | Mach number controller in wind tunnel | |
CN201476905U (en) | Neural network PID temperature controlled thermocouple automatic verification system | |
CN204807295U (en) | A wind tunnel experiment device for different plate -fin heat exchanger plain fin | |
CN108106679B (en) | Method and system for measuring inlet air volume of coal mill of power station | |
CN103245481B (en) | Frequency conversion technology based detection method for air resistance characteristic of large-scale variable-load heat exchanger | |
CN109636677A (en) | Building thermal technique performance estimating method based on model calibration | |
CN105528000A (en) | Intelligent temperature control meter for aircraft | |
CN101813920B (en) | Virtual redundancy method for temperature sensor of power station turboset | |
CN102645315B (en) | Automatic, fast and accurate detection method for air resistance characteristics of large heat exchanger | |
CN202210005U (en) | Heat energy meter flow full-automatic detection device | |
CN110779132A (en) | Water pump equipment operation control system of air conditioning system based on reinforcement learning | |
CN105400951B (en) | A kind of humidification hybrid control system and its control method for annealing process of silicon steel | |
CN103279030B (en) | Dynamic soft measuring modeling method and device based on Bayesian frame | |
CN104034378A (en) | Constant-current thermal gas mass flow meter and measuring method implemented by same | |
CN111998423A (en) | Electric heat storage circulating air quantity-water temperature control system and prediction control method thereof | |
CN103413029A (en) | Moving horizon estimation method for multispeed sampling continuous agitating pot type reactor | |
CN204064363U (en) | A kind of sensor circuit for thermal type gas quality flow meter |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
EE01 | Entry into force of recordation of patent licensing contract |
Application publication date: 20130814 Assignee: ZHEJIANG YINBO INTELLIGENT TECHNOLOGY Co.,Ltd. Assignor: HANGZHOU DIANZI University Contract record no.: X2020330000114 Denomination of invention: Detection method of gas resistance characteristics of large heat exchanger with variable load based on frequency conversion technology Granted publication date: 20150715 License type: Common License Record date: 20201216 |
|
EE01 | Entry into force of recordation of patent licensing contract |