CN108259099B - Electromagnetic radiation prediction method for TD-SCDMA base station - Google Patents

Electromagnetic radiation prediction method for TD-SCDMA base station Download PDF

Info

Publication number
CN108259099B
CN108259099B CN201810071037.1A CN201810071037A CN108259099B CN 108259099 B CN108259099 B CN 108259099B CN 201810071037 A CN201810071037 A CN 201810071037A CN 108259099 B CN108259099 B CN 108259099B
Authority
CN
China
Prior art keywords
frequency sequence
prediction
training
parameter
low
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
Application number
CN201810071037.1A
Other languages
Chinese (zh)
Other versions
CN108259099A (en
Inventor
杨万春
吴涛
张雪
彭艳芬
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Xiangtan University
Original Assignee
Xiangtan University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Xiangtan University filed Critical Xiangtan University
Priority to CN201810071037.1A priority Critical patent/CN108259099B/en
Publication of CN108259099A publication Critical patent/CN108259099A/en
Application granted granted Critical
Publication of CN108259099B publication Critical patent/CN108259099B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B17/00—Monitoring; Testing
    • H04B17/30—Monitoring; Testing of propagation channels
    • H04B17/391—Modelling the propagation channel
    • H04B17/3913—Predictive models, e.g. based on neural network models

Abstract

The invention discloses a TD-SCDMA base station electromagnetic radiation prediction method, which comprises the following steps: the method comprises the steps of taking TD-SCDMA base station electromagnetic radiation historical data as training data, decomposing the training data through DB3 and COIF3 wavelets, obtaining a low-frequency sequence R and a high-frequency sequence T through DB3 wavelet decomposition, obtaining a low-frequency sequence Q and a high-frequency sequence U through COIF3 wavelet decomposition, inputting the low-frequency sequence R, the high-frequency sequence T, the low-frequency sequence Q and the high-frequency sequence U into prediction models respectively for training, and performing combined prediction on TD-SCDMA base station electromagnetic radiation through the trained prediction models. The invention analyzes the electromagnetic radiation prediction mode of the TD-SCDMA base station, and the method can accurately and reliably predict and evaluate the electromagnetic radiation of the TD-SCDMA base station and has certain social benefit.

Description

Electromagnetic radiation prediction method for TD-SCDMA base station
Technical Field
The invention relates to a TD-SCDMA base station electromagnetic radiation prediction method.
Background
At present, when electromagnetic radiation around a communication base station is predicted, a trained prediction model is usually used for prediction. For example, application No. 2018100095052 discloses a method for predicting electromagnetic radiation of a GSM base station, in the prediction of electromagnetic radiation of a base station, a prediction model is trained first, and then the trained model is used for prediction, but in the prediction, because the model uses a single prediction model, the single prediction model is not stable enough to be influenced by interference in the prediction process, and simultaneously, because the electromagnetic radiation of a TD-SCDMA base station changes violently, the model is not easy to predict.
Aiming at the defects in the prior art, the method combines the regular characteristics of the electromagnetic radiation sequence of the TD-SCDMA base station, takes the electromagnetic radiation historical data of the TD-SCDMA base station as training data to train the model, firstly decomposes the training data through DB3 and COIF3 wavelets, the decomposition layer number is 1, decomposes the DB3 wavelets to obtain a low-frequency sequence R and a high-frequency sequence T, decomposes the COIF3 wavelets to obtain a low-frequency sequence Q and a high-frequency sequence U, then respectively inputs the low-frequency sequence R, the high-frequency sequence T, the low-frequency sequence Q and the high-frequency sequence U into a prediction model to train, and then carries out combined prediction on the electromagnetic radiation of the TD-SCDMA base station by the trained prediction model.
Disclosure of Invention
In order to solve the above technical problem, the present invention provides a method for predicting electromagnetic radiation of a TD-SCDMA base station.
The technical scheme for solving the technical problems comprises the following steps:
1) taking TD-SCDMA base station electromagnetic radiation historical data as training data, decomposing the training data through DB3 and COIF3 wavelets, wherein the number of decomposition layers is 1, obtaining a low-frequency sequence R and a high-frequency sequence T through DB3 wavelet decomposition, and obtaining a low-frequency sequence Q and a high-frequency sequence U through COIF3 wavelet decomposition;
2) inputting the low-frequency sequence R, the high-frequency sequence T, the low-frequency sequence Q and the high-frequency sequence U obtained in the step 1) into a prediction model for training to obtain a parameter omega of the prediction model1、b1、c1，ω2、b2、c2，ω3、b3、c3，ω4、b4、c4；
3) Inputting the input data in the prediction into the prediction model trained in the step 2) respectively to predict the high-frequency sequence and the low-frequency sequence, wherein the expression is as follows:
Figure GDA0002717200330000011
Figure GDA0002717200330000012
Figure GDA0002717200330000021
Figure GDA0002717200330000022
wherein, ω is1、b1、c1，ω2、b2、c2，ω3、b3、c3，ω4、b4、c4For the prediction model parameters obtained by the training of step 2), xRiLow frequency sequence obtained by DB3 wavelet decomposition for historical data before predicted point, and the unit is V/m, xTiHigh frequency sequence obtained by DB3 wavelet decomposition for historical data before predicted point, and the unit is V/m, xQiLow frequency sequence obtained by wavelet decomposition of COIF3 for historical data before predicted point, and the unit is V/m, xUiHigh frequency sequence obtained by wavelet decomposition of COIF3 for historical data before predicted point, and the unit is V/m, yRFor a low frequency sequence xRiThe unit of the predicted value of (a) is V/m, yTIs a high frequency sequence xTiThe unit of the predicted value of (a) is V/m, yQFor a low frequency sequence xQiThe unit of the predicted value of (a) is V/m, yUIs a high frequency sequence xUiThe unit of the predicted value of (2) is V/m;
4) and performing combined prediction according to the high and low frequency prediction values obtained by training in the step 3), wherein the combined prediction expression is as follows:
yDB3＝yR+yT (5)
yCOIF3＝yQ+yU (6)
y＝(yDB3+yCOIF3)/2 (7)
wherein, yDB3For wavelet decomposition by DB3, the low frequency sequence prediction value y with 1 decomposition levelRAnd the high frequency sequence prediction value yTThe sum of the predicted values is in units of V/m, yCOIF3For wavelet decomposition by COIF3, the low frequency sequence prediction value y with 1 decomposition levelQAnd the high frequency sequence prediction value yUThe unit of the predicted value obtained by summation is V/m, y is the combined predicted value of the proposed model, and the unit is V/m.
In the above method for predicting electromagnetic radiation of TD-SCDMA base station, in step 2), the prediction model is as follows:
Figure GDA0002717200330000023
in the above equation (8), n pieces of history data are decomposed by DB3 wavelet, the number of decomposition layers is 1, the obtained low frequency sequence R is input to the prediction model, and training is performed to obtain the optimal prediction model parameter ω1、b1、c1The parameter ω1、b1Is assigned a value ofRandom number between 1 and 1, and inputting training data into training model to obtain parameter c1To determine the model parameter ω1、b1、c1；
Figure GDA0002717200330000024
In the above formula (9), n pieces of history data are decomposed by DB3 wavelet, the number of decomposition layers is 1, the obtained high frequency sequence T is input to the prediction model, and training is performed to obtain the optimal prediction model parameter ω2、b2、c2The parameter ω2、b2Assigning a random number between-1 and 1, and inputting training data into a training model to obtain a parameter c2To determine the model parameter ω2、b2、c2；
Figure GDA0002717200330000025
In the above equation (10), n pieces of history data are decomposed by COIF3 wavelet, the number of layers is 1, the obtained low-frequency sequence Q is input into the prediction model, and training is performed to obtain the optimal prediction model parameter ω3、b3、c3The parameter ω3、b3Assigning a random number between-1 and 1, and inputting training data into a training model to obtain a parameter c3To determine the model parameter ω3、b3、c3；
Figure GDA0002717200330000031
In the above formula (11), n pieces of history data are decomposed by COIF3 wavelet, the number of decomposition layers is 1, the obtained high frequency sequence U is input into the prediction model, and training is performed to obtain the optimal prediction model parameter ω4、b4、c4The parameter ω4、b4Assigning a random number between-1 and 1, and inputting training data into a training model to obtain parametersNumber c4To determine the model parameter ω4、b4、c4。
The invention has the beneficial effects that: decomposing the electromagnetic radiation historical data of the TD-SCDMA base station into a low-frequency sequence and a high-frequency sequence by decomposing DB3 and COIF3 wavelets with the number of layers being 1 so that the electromagnetic radiation of the base station with violent change becomes gentle and convenient for prediction, training the low-frequency sequence and the high-frequency sequence obtained by decomposition through a prediction model, and performing combined prediction on the electromagnetic radiation of the TD-SCDMA base station through the trained prediction model. The established model can accurately and reliably predict the electromagnetic radiation of the TD-SCDMA base station, and the method has good reference value for base station construction and environmental protection and has certain social benefit.
Drawings
FIG. 1 is a flow chart of the present invention.
Detailed Description
The present embodiment is performed on the premise of the present disclosure, and detailed implementation procedures are given, but the scope of the present disclosure is not limited to the following embodiments.
In the experiment place, in an open and flat area, the implementation object is a teaching building roof base station, the experimental apparatus is a spectrum analyzer AT6024D, and the measurement object is a TD-SCDMA base station.
The invention is further described below with reference to the figures and examples.
Fig. 1 is a block diagram of the technical scheme of the invention, and the specific steps are as follows:
step 1): the method comprises the steps of taking TD-SCDMA base station electromagnetic radiation historical data as training data, decomposing the training data through DB3 and COIF3 wavelets, wherein the number of decomposition layers is 1, obtaining a low-frequency sequence R and a high-frequency sequence T through DB3 wavelet decomposition, and obtaining a low-frequency sequence Q and a high-frequency sequence U through COIF3 wavelet decomposition.
In the embodiment, the electromagnetic radiation history data of the TD-SCDMA base station is taken as training data, 12 history data are taken as training data, decomposition is performed by decomposing DB3 and COIF3 wavelets with the number of layers of 1, 12 low-frequency sequences R and 12 high-frequency sequences T are obtained by decomposing DB3 wavelets with the number of layers of 1, and 12 low-frequency sequences Q and 12 high-frequency sequences U are obtained by decomposing COIF3 wavelets with the number of layers of 1.
Step 2): inputting the low-frequency sequence R, the high-frequency sequence T, the low-frequency sequence Q and the high-frequency sequence U obtained in the step 1) into a prediction model for training to obtain a parameter omega of the prediction model1、b1、c1，ω2、b2、c2，ω3、b3、c3，ω4、b4、c4。
In this embodiment, the low-frequency sequence R, the high-frequency sequence T, the low-frequency sequence Q, and the high-frequency sequence U obtained in step 1) are taken as training data, and are respectively input to a model for training, and when the model is trained by the low-frequency sequence R, R is used for training the model1,i＝[0.197623,0.202044,...,0.191881]，i＝1,2,...,12,yR＝[0.188432]The following model is input:
Figure GDA0002717200330000041
in the above equation (12), the history data is wavelet-decomposed by DB3 with the number of decomposition layers being 1, and the obtained low frequency sequence R is input to the prediction model while the parameter ω is input1And b1Assigned as a random number between-1 and 1, ω1Assigned value of 0.8129, b1The assignment is 0.2119 training to obtain the prediction model parameter c1＝-0.62911；
When the high-frequency sequence T trains the model, the T is used1,i＝[-0.00093,-0.00041,...,0.002599]，i＝1,2,...,12，yT＝[-0.00058]The following model is input:
Figure GDA0002717200330000042
in the above equation (13), the history data is wavelet-decomposed by DB3 with the number of decomposition layers being 1, and the obtained high-frequency sequence T is input to the prediction model, while the parameter ω is input2And b2Assigned as a random number between-1 and 1, ω2Assigned value of 0.6213, b2Assignment of valueObtaining prediction model parameters c for 0.3175 training2＝-0.3248；
When the low-frequency sequence Q trains the model, Q is used1,i＝[0.196319,0.201446,...,0.192642]，i＝1,2,...,12，yQ＝[0.190339]The following model is input:
Figure GDA0002717200330000043
in the above equation (14), the history data is decomposed into 1 number of layers by the COIF3 wavelet, the obtained low frequency sequence Q is input to the prediction model, and the parameter ω is input at the same time3And b3Assigned as a random number between-1 and 1, ω3The value is-0.2315, b3The assignment is 0.2614 training to obtain the prediction model parameter c3＝0.717792；
When the model is trained by the high-frequency sequence U, the U is trained1,i＝[0.000371,-0.00012,...,0.001838]，i＝1,2,...,12，yu＝[-0.00249]The following model is input:
Figure GDA0002717200330000044
in the above equation (15), the history data is decomposed by the COIF3 wavelet, the number of layers is 1, the obtained high frequency sequence U is input to the prediction model, and the parameter ω is input4And b4Assigned as a random number between-1 and 1, ω4Assigned value of 0.2755, b4The assignment is 0.3145 training to obtain the prediction model parameter c4＝-0.3125。
Step 3), inputting the input data in prediction into the prediction model trained in step 2) respectively to predict high and low frequency sequences, wherein the expression is as follows:
Figure GDA0002717200330000051
Figure GDA0002717200330000052
Figure GDA0002717200330000053
Figure GDA0002717200330000054
wherein, ω is1、b1、c1，ω2、b2、c2，ω3、b3、c3，ω4、b4、c4For the prediction model parameters obtained by the training of step 2), xRiLow frequency sequence obtained by DB3 wavelet decomposition for historical data before predicted point, and the unit is V/m, xTiHigh frequency sequence obtained by DB3 wavelet decomposition for historical data before predicted point, and the unit is V/m, xQiLow frequency sequence obtained by wavelet decomposition of COIF3 for historical data before predicted point, and the unit is V/m, xUiHigh frequency sequence obtained by wavelet decomposition of COIF3 for historical data before predicted point, and the unit is V/m, yRFor a low frequency sequence xRiThe unit of the predicted value of (a) is V/m, yTIs a high frequency sequence xTiThe unit of the predicted value of (a) is V/m, yQFor a low frequency sequence xQiThe unit of the predicted value of (a) is V/m, yUIs a high frequency sequence xUiThe unit of the predicted value of (2) is V/m.
In this embodiment, input data during prediction is input into the prediction model trained in step 2), and high and low frequency sequences are predicted for 14 times, and predicted values are obtained as shown in the following table:
TABLE 1 prediction of high and low frequency sequences
y R 0.2 123 8 0.2 149 61 0.2 186 3 0.2 245 95 0.2 205 02 0.2 133 21 0.2 137 27 0.2 148 49 0.2 135 5 0.2 196 35 0.2 134 73 0.2 113 62 0.2 128 88 0.212662
y T 0.0 006 4 - 0.0 001 2 - 0.0 002 2 0.0 006 45 - 0.0 007 9 8.8 9E- 05 0.0 003 33 - 0.0 013 1 0.0 018 0.0 004 05 - 0.0 014 6 0.0 006 18 0.0 001 32 -0.00056
y Q 0.1 835 91 0.1 884 21 0.1 951 96 0.1 961 14 0.1 942 89 0.1 897 28 0.1 858 98 0.1 890 25 0.1 931 45 0.1 919 14 0.1 892 13 0.1 876 63 0.1 868 47 0.18827
y U 0.0 036 82 0.0 005 74 - 0.0 027 6 0.0 031 21 - 0.0 005 4 - 0.0 021 9 0.0 022 68 - 0.0 013 4 - 0.0 007 7 0.0 022 09 - 0.0 003 6 - 0.0 012 1 0.0 003 6 0.000988
Step 4): performing combined prediction according to the high and low frequency prediction values obtained by training in the step 3), wherein the combined prediction expression is as follows:
yDB3＝yR+yT (20)
yCOIF3＝yQ+yU (21)
y＝(yDB3+yCOIF3)/2 (22)
wherein, yDB3For wavelet decomposition by DB3, the low frequency sequence prediction value y with 1 decomposition levelRAnd the high frequency sequence prediction value yTThe sum of the predicted values is in units of V/m, yCOIF3For wavelet decomposition by COIF3, the low frequency sequence prediction value y with 1 decomposition levelQAnd the high frequency sequence prediction value yUThe unit of the predicted value obtained by summation is V/m, y is the combined predicted value of the proposed model, and the unit is V/m.
Substituting the high and low frequency predicted values obtained in the step 3) into the formulas (20) and (21) respectively to obtain yDB3、yCOIF3A value of (a) and then yDB3、yCOIF3The value of (a) is substituted for the formula (22) to obtain the value of y, wherein yDB3、yCOIF3The values of y are all set forth in the following table:
TABLE 2yDB3、yCOIF3Value of y
yDB3 0.21 242 0.21 484 0.21 8412 0.22 524 0.21 9712 0.21 341 0.21 406 0.21 354 0.21 835 0.22 004 0.21 471 0.21 198 0.21 242 0.2 121
yCOI F 0.18 7273 0.18 8995 0.19 2436 0.19 9238 0.19 3749 0.18 7538 0.18 8266 0.18 7685 0.19 2405 0.19 4123 0.18 8853 0.18 6123 0.18 7207 0.1 892 5
y 0.20 0147 0.20 1914 0.20 5426 0.21 2239 0.20 6732 0.20 0474 0.20 1133 0.20 0613 0.20 5377 0.20 7082 0.20 1481 0.19 9067 0.20 0113 0.2 021 4
Measuring Measurement of Value of 0.20 20 0.20 38 0.20 73 0.21 42 0.20 86 0.20 24 0.20 24 0.20 25 0.20 73 0.20 90 0.20 37 0.20 09 0.20 20 0.2 041
The experimental results show that each predicted value is close to the measured value, and the electromagnetic radiation of the TD-SCDMA base station can be stably predicted at different times, which shows that the method can realize accurate and reliable prediction of the electromagnetic radiation of the TD-SCDMA base station, and simultaneously the experimental results verify the effectiveness of the method used by the invention.

Claims (2)

1. A TD-SCDMA base station electromagnetic radiation prediction method is characterized in that the method comprises the following steps:
1) taking TD-SCDMA base station electromagnetic radiation historical data as training data, decomposing the training data through DB3 and COIF3 wavelets, wherein the number of decomposition layers is 1, obtaining a low-frequency sequence R and a high-frequency sequence T through DB3 wavelet decomposition, and obtaining a low-frequency sequence Q and a high-frequency sequence U through COIF3 wavelet decomposition;
2) inputting the low-frequency sequence R, the high-frequency sequence T, the low-frequency sequence Q and the high-frequency sequence U obtained in the step 1) into a prediction model for training to obtain a parameter omega of the prediction model1、b1、c1，ω2、b2、c2，ω3、b3、c3，ω4、b4、c4；
3) Inputting the input data in the prediction into the prediction model trained in the step 2) respectively to predict the high-frequency sequence and the low-frequency sequence, wherein the expression is as follows:
Figure FDA0002717200320000011
Figure FDA0002717200320000012
Figure FDA0002717200320000013
Figure FDA0002717200320000014
wherein, ω is1、b1、c1，ω2、b2、c2，ω3、b3、c3，ω4、b4、c4For the prediction model parameters obtained by the training of step 2), xRiLow frequency sequence obtained by DB3 wavelet decomposition for historical data before predicted point, and the unit is V/m, xTiObtained by DB3 wavelet decomposition for historical data before predicted pointHigh frequency sequence with unit of V/m, xQiLow frequency sequence obtained by wavelet decomposition of COIF3 for historical data before predicted point, and the unit is V/m, xUiHigh frequency sequence obtained by wavelet decomposition of COIF3 for historical data before predicted point, and the unit is V/m, yRFor a low frequency sequence xRiThe unit of the predicted value of (a) is V/m, yTIs a high frequency sequence xTiThe unit of the predicted value of (a) is V/m, yQFor a low frequency sequence xQiThe unit of the predicted value of (a) is V/m, yUIs a high frequency sequence xUiThe unit of the predicted value of (2) is V/m;
4) and performing combined prediction according to the high and low frequency prediction values obtained by training in the step 3), wherein the combined prediction expression is as follows:
yDB3＝yR+yT (5)
yCOIF3＝yQ+yU (6)
y＝(yDB3+yCOIF3)/2 (7)
wherein, yDB3For wavelet decomposition by DB3, the low frequency sequence prediction value y with 1 decomposition levelRAnd the high frequency sequence prediction value yTThe sum of the predicted values is in units of V/m, yCOIF3For wavelet decomposition by COIF3, the low frequency sequence prediction value y with 1 decomposition levelQAnd the high frequency sequence prediction value yUThe unit of the predicted value obtained by summation is V/m, y is the combined predicted value of the proposed model, and the unit is V/m.
2. The method of claim 1, wherein in step 2), the prediction model is as follows:
Figure FDA0002717200320000021
in the above equation (8), n pieces of history data are decomposed by DB3 wavelet, the number of decomposition layers is 1, the obtained low frequency sequence R is input to the prediction model, and training is performed to obtain the optimal prediction model parameter ω1、b1、c1The parameter ω1、b1Assigning a random number between-1 and 1, and inputting training data into a training model to obtain a parameter c1To determine the model parameter ω1、b1、c1；
Figure FDA0002717200320000022
In the above formula (9), n pieces of history data are decomposed by DB3 wavelet, the number of decomposition layers is 1, the obtained high frequency sequence T is input to the prediction model, and training is performed to obtain the optimal prediction model parameter ω2、b2、c2The parameter ω2、b2Assigning a random number between-1 and 1, and inputting training data into a training model to obtain a parameter c2To determine the model parameter ω2、b2、c2；
Figure FDA0002717200320000023
In the above equation (10), n pieces of history data are decomposed by COIF3 wavelet, the number of layers is 1, the obtained low-frequency sequence Q is input into the prediction model, and training is performed to obtain the optimal prediction model parameter ω3、b3、c3The parameter ω3、b3Assigning a random number between-1 and 1, and inputting training data into a training model to obtain a parameter c3To determine the model parameter ω3、b3、c3；
Figure FDA0002717200320000024
In the above formula (11), n pieces of history data are decomposed by COIF3 wavelet, the number of decomposition layers is 1, the obtained high frequency sequence U is input into the prediction model, and training is performed to obtain the optimal prediction model parameter ω4、b4、c4Will beParameter omega4、b4Assigning a random number between-1 and 1, and inputting training data into a training model to obtain a parameter c4To determine the model parameter ω4、b4、c4。
CN201810071037.1A 2018-01-25 2018-01-25 Electromagnetic radiation prediction method for TD-SCDMA base station Active CN108259099B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810071037.1A CN108259099B (en) 2018-01-25 2018-01-25 Electromagnetic radiation prediction method for TD-SCDMA base station

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810071037.1A CN108259099B (en) 2018-01-25 2018-01-25 Electromagnetic radiation prediction method for TD-SCDMA base station

Publications (2)

Publication Number Publication Date
CN108259099A CN108259099A (en) 2018-07-06
CN108259099B true CN108259099B (en) 2021-04-23

Family

ID=62742692

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810071037.1A Active CN108259099B (en) 2018-01-25 2018-01-25 Electromagnetic radiation prediction method for TD-SCDMA base station

Country Status (1)

Country Link
CN (1) CN108259099B (en)

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107294623A (en) * 2017-06-20 2017-10-24 湘潭大学 A kind of Novel Communication base station electromagnetic radiation Forecasting Methodology

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101819253B (en) * 2010-04-20 2013-10-23 湖南大学 Probabilistic neural network-based tolerance-circuit fault diagnosis method
CN102056183B (en) * 2010-12-10 2013-08-21 北京交通大学 Network flow prediction method and device based on cognitive network
CN104408529A (en) * 2014-11-21 2015-03-11 广东工业大学 Short-term load predicting method of power grid
CN104992248A (en) * 2015-07-07 2015-10-21 中山大学 Microgrid photovoltaic power station generating capacity combined forecasting method
CN107545321A (en) * 2017-07-25 2018-01-05 东南大学 A kind of ARMA RBF by-product gas generating capacity combination forecasting methods based on wavelet transformation
CN107292462A (en) * 2017-08-25 2017-10-24 广东工业大学 A kind of short-term load forecasting method, apparatus and system

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107294623A (en) * 2017-06-20 2017-10-24 湘潭大学 A kind of Novel Communication base station electromagnetic radiation Forecasting Methodology

Also Published As

Publication number Publication date
CN108259099A (en) 2018-07-06

Similar Documents

Publication Publication Date Title
CN111193256A (en) Power load prediction method based on variational modal decomposition and gated cyclic unit
CN112308008B (en) Radar radiation source individual identification method based on working mode open set of transfer learning
CN102083119B (en) Method and device for evaluating network coverage interference
CN107862114A (en) Small echo PSO SVM Diagnosis Method of Transformer Faults based on three ratio feature amounts
CN110113075B (en) Hybrid network station frequency hopping parameter blind estimation method based on STFT-SPWVD
CN109375253B (en) Earthquake motion parameter evaluation method based on maximum credible earthquake of all earthquake-generating structures
CN105760347A (en) HHT end effect restraining method based on data/extreme value joint symmetric prolongation
Akpinar et al. Naive forecasting of household natural gas consumption with sliding window approach
CN108734264A (en) Deep neural network model compression method and device, storage medium, terminal
CN108183754B (en) Electromagnetic radiation prediction method for GSM base station
CN106483563A (en) seismic energy compensation method based on complementary set empirical mode decomposition
CN109918776A (en) The engineering prediction on fatigue crack growth method of engineering goods based on two-step least square method
CN108259099B (en) Electromagnetic radiation prediction method for TD-SCDMA base station
CN115099420A (en) Model aggregation weight dynamic distribution method for wireless federal learning
CN106295005B (en) The Multi-factor estimation method and apparatus of contact fatigue life of spray coating layer
CN105915299A (en) Time-frequency two-dimensional LMBP neural network based frequency spectrum prediction method in ISM frequency range
CN103970129A (en) Control valve adhesion detecting method
CN112241800B (en) Method for calculating VOCs pollutant emission amount of coke oven
CN104008292B (en) Broad-band antenna super-broadband electromagnetic impulse response prediction method
CN114441111B (en) Pipeline leakage infrasonic wave signal analysis method and system
Mallik et al. EME-Net: A U-net-based indoor EMF exposure map reconstruction method
CN109257128A (en) A kind of spectrum signal recognition methods and system based on Fourier space fitting denoising
CN113675854A (en) Distribution network cable transformer verification method and device considering transformer area voltage loss
CN112330046A (en) Power demand prediction method based on multi-dimensional gray-neural network hybrid coordination
CN111428932A (en) Medium-and-long-term air traffic flow prediction method based on wavelet transformation and gray prediction

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant