A method for quickly predicting the cumulative flow of natural gas pipeline networks

Abstract

The present invention relates to a method for quickly predicting the cumulative flow of a natural gas pipeline network. The present invention can quickly predict the cumulative flow of site users according to different natural gas pipeline network distribution conditions. The data selected by the present invention is the data of the local natural gas pipeline network of the prediction site and its upstream and downstream sites. Compared with other algorithms, the amount of selected data is less and the calculation time is shorter; the present invention also segments the selected data again, and the smaller the segmentation interval, the shorter the time required to obtain the optimal empirical parameters and predict the future cumulative flow. The assignment method of the present invention is easy to operate, fast in operation, and suitable for promotion. Different sites only need to obtain upstream and downstream pressure data and pipeline information to achieve their flow measurement, and can be extended to the calculation of the entire pipeline network, with good adaptability.

Description

Technical Field

[0001] The present invention relates to the field of flow prediction using big data, and in particular to a method for quickly predicting the cumulative flow of a natural gas pipeline network. Background Art

[0002] As we all know, natural gas is a high-quality fuel and chemical raw material. In the process of production and life, natural gas is more often used as fuel. It burns completely, has a large unit calorific value, and the products after combustion have little impact on the environment. Therefore, the proportion of natural gas in the world's energy structure continues to rise. With the rapid development of my country's society and economy, the natural gas pipeline network system is becoming increasingly large. However, due to the problem of incomplete or inaccurate data in the process of natural gas pipeline network scheduling, monitoring and maintenance, flow prediction of natural gas pipeline network has become one of the important parts of the operation and management analysis of natural gas pipeline enterprises.

[0003] However, due to the complex distribution of natural gas pipelines, the measurement data has the characteristics of large data volume, strong data time series and diverse data types. At the same time, the data measured by the instrument is inevitably lost and incomplete, which makes the calculation of the predicted value of the cumulative flow of the natural gas pipeline network very cumbersome. At present, the core technology of related prediction software is monopolized by foreign companies, and domestic institutions need to spend a lot of money to use prediction software and maintain the system. Summary of the invention

[0004] In view of the problem that the measurement data is complex or missing due to the huge scale of the natural gas pipeline system, the purpose of the present invention is to propose a method for quickly predicting the cumulative flow of the natural gas pipeline network. The prediction of the cumulative flow of the natural gas pipeline network has a certain reliability and is helpful for the operation and management analysis of the natural gas pipeline enterprises.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is:

[0006] It is known that the measurement time of each day is measured every hour, and each hour has the corresponding measured pressure data at both ends of the pipeline and the cumulative flow of the site users. Therefore, the pressure data at both ends of each pipe section is taken as a known quantity, such as Figure 1 The figure shows the distribution diagram of the straight pipe section stations. The pipelines adjacent to the calculation station are pipeline 1 and pipeline 2. The pressure difference at both ends of pipeline 1 and pipeline 2 is calculated according to formula (1) and formula (2).

[0007] ΔP t1 =|P t1 2 -P t3 2 | (1)

[0008] Where ΔP t1represents the square difference of pressure at both ends of pipeline 1 at the tth hour; P t1 , P t3 Indicates the pressure at both ends of pipeline 1 at the tth hour (unit: MPa).

[0009] ΔP t2 =|P t4 2 -P t2 2 | (2)

[0010] Where ΔP t2 represents the square difference of the pressure at both ends of pipeline 2 at the tth hour; P t2 , P t4 Represents the pressure at both ends of pipeline 2 at the tth hour (unit: MPa).

[0011] According to formula (3), the instantaneous flow Q of the calculation station at the tth hour is established t The relationship between the pressure difference at both ends of the node is used to obtain the instantaneous flow rate Q corresponding to the tth hour. t :

[0012] Q t =α|ΔP t1 -ΔP t2 | (3)

[0013] Where Q t Indicates the instantaneous flow rate of the computing node at the tth hour (unit: Nm 3 / h); α represents the empirical parameter of the calculation site; ΔP t1 , ΔP t2 It represents the square difference of the pressure at both ends of pipeline 1 and pipeline 2 at the tth hour.

[0014] According to formula (4), the instantaneous flow at each time node is accumulated in sequence to obtain the cumulative flow per hour:

[0015]

[0016] Where SQ tP Indicates the cumulative flow rate of the calculation station at the tth hour (unit: Nm 3 / h); SQ0 represents the total cumulative flow rate known in the previous period (unit: Nm 3 / h); Q t It represents the instantaneous flow rate of the calculation station at the tth hour (unit: Nm 3 / h); t represents the tth hour (t=1,2…m); m represents the total number of hours.

[0017] Then the cumulative flow SQ of the calculation site at the tth hour is calculated tpThe measured cumulative flow SQ at the known t hour t0 The relative error is calculated according to formula (5):

[0018]

[0019] Where δ t Represents the relative error at the tth hour; SQ tP Indicates the cumulative flow rate of the calculation station at the tth hour (unit: Nm 3 / h); SQ t0 Indicates the measured flow rate at the calculation site in the tth hour (unit: Nm 3 / h).

[0020] The known actual flow test pressure data is divided into n time periods according to the day unit, where the first time period is from the first day to the n-1 day; the second time period is from the second day to the n-1 day; the third time period is from the third day to the n-1 day. Similarly, the n-2 time period is from the n-2 day and the n-1 day; the n-1 time period is from the n-1 day; the n-n time period is from the n day, and the first n-1 time periods are used to fit the most suitable empirical parameters to predict the cumulative flow of the n-time period.

[0021] Calculate the corresponding optimal empirical parameter α for the first n-1 periods of time n-1 , by continuously iterating and calculating the empirical parameters, the iterative calculation process for each period of time is as follows:

[0022] (1) Adjust the empirical parameter from 0.01 to iterate, with a step size of 0.01, and obtain the first empirical parameter α1;

[0023] (2) According to formulas (1) and (2), the square difference of the pressure at both ends of each pipeline at the tth hour (t = 1, 2…m) is calculated in sequence to obtain ΔP at the corresponding time. t1 , ΔP t2 ;

[0024] (3) Calculate the instantaneous flow Q of the calculation site corresponding to each hour according to formula (3): t ;

[0025] (4) According to formula (4), the instantaneous flow rate of each hour is accumulated to obtain the cumulative flow rate per hour, and the cumulative flow rate SQ of the tth hour is calculated in turn. tP ;

[0026] (5) According to formula (5), the cumulative flow SQ at the tth hour is tP The measured cumulative flow SQ at the corresponding time t0 The relative error calculation is performed to obtain the relative error δ at this moment. t, there are m in total;

[0027] (6) Compare the sizes of the m relative errors and record the maximum relative error δ under this adjustment parameter α1 max ;

[0028] (7) Taking 0.01 as the step size of the empirical parameter, the i-th empirical parameter α is obtained in turn. i , repeat steps (1) to (6) and record the maximum relative error under i different empirical parameters;

[0029] (8) Compare the i maximum relative errors, and take the empirical parameter corresponding to the minimum value among the i values ​​as the optimal empirical parameter α for this time period. n-1 .

[0030] The first n-1 periods of time each have a corresponding optimal adjustment parameter α n-1 And under this adjustment parameter, there is a maximum relative error δ in each period of time n-1 . Compare the error δ of the previous n-1 periods n-1 , and the optimal adjustment parameter result corresponding to its minimum value is used as the prediction experience parameter α n , to predict the cumulative flow in the nth time period.

[0031] Furthermore, the time period grouping unit of this method is not limited to days, and can be grouped as needed according to actual conditions. It only needs to ensure that the length of the time period for solving the empirical parameters is much longer than the length of the predicted time period, and the optimal empirical parameters within the error range are used to predict the cumulative flow in the future time period.

[0032] Furthermore, this method is not only applicable to straight pipe sections. For the calculation of sites with multiple pipelines, it is only necessary to calculate the square difference of the pressures of the redundant pipelines according to formulas (1) and (2), add the pressure difference of the redundant pipelines to formula (3), and finally calculate it in the same way as the straight pipe section.

[0033] Compared with the existing method, the present invention has the following beneficial effects:

[0034] (1) The present invention does not require high computing hardware conditions, and can quickly predict the cumulative flow of site users according to different natural gas pipeline network distribution conditions. The data selected by the present invention is the data of the local natural gas pipeline network of the prediction site and its upstream and downstream sites. Compared with other algorithms, the amount of selected data is smaller and the calculation time is shorter; the present invention also divides the selected data into segments again. The smaller the segmentation interval, the shorter the time required to obtain the optimal empirical parameters and predict the future cumulative flow.

[0035] (2) The method for quickly predicting the cumulative flow of a natural gas pipeline network described in the present invention is more cost-effective than a horizontally arranged online gas composition measuring instrument.

[0036] (3) The value assignment method of the present invention is easy to operate, fast in operation, and suitable for promotion. Different sites only need to obtain upstream and downstream pressure data and pipeline information to achieve their flow measurement, and can be extended to the calculation of the entire pipeline network, with good adaptability.

[0037] (4) The value assignment algorithm constructed by the present invention can obtain high-precision prediction results that are close to the actual measured values ​​for different time periods, and has wide applicability and a certain degree of reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 Calculate site distribution maps for straight pipe sections;

[0039] Figure 2 The site distribution map of Sanmen Station for calculation example;

[0040] Figure 3 The data graph and relative error graph of the predicted and measured values ​​of the Sanmen station in the calculation example. DETAILED DESCRIPTION

[0041] The present invention will be further described below with reference to the accompanying drawings and examples.

[0042] The present invention uses Sanmen Station in Zhejiang Province as a calculation example, and its station distribution is as follows: Figure 2 As shown. It is known that the lengths of pipeline 1 and pipeline 2 are 41.44km and 31.25km respectively, and the pipe diameters are both 813mm. The pressure differences at the inlet and outlet of pipeline 1 are P1 and P3 respectively, and the pressure differences at the inlet and outlet of pipeline 2 are P2 and P4 respectively. The time period used in the calculation example is: July 27, 2020 to August 2, 2020. The measurement data is based on each hour as a node, and there is a corresponding pipeline pressure difference P and site user instantaneous flow Q every hour. t And the site user cumulative traffic SQ t0 .

[0043] Taking the seven-day period of a week at Sanmen Station as an example, the actual flow test pressure data are divided into seven sections according to the calculation time: the first period is from July 27 to August 1; the second period is from July 28 to August 1; the third period is from July 29 to August 1; the fourth period is from July 30 to August 1; the fifth period is from July 31 to August 1; the sixth period is August 1; and the seventh period is August 2. The empirical parameter α corresponding to the minimum relative error obtained by fitting the first six periods is used to predict the cumulative flow on August 2.

[0044] Taking the calculation of the optimal empirical parameters for the sixth period as an example, the iterative calculation process is as follows:

[0045] (1) Adjust the empirical parameter and start iterating from 0.01, with a step size of 0.01, to obtain the first empirical parameter α1

[0046] (2) According to formulas (1) and (2), the square difference of the pressure at both ends of each pipeline at the tth hour (t = 1, 2…24) is calculated in sequence to obtain ΔP at the corresponding time. t1 , ΔP t2 , the calculation results are as follows:

[0047] t=1,ΔP 11 =2.3609,ΔP 12 =0.44361

[0048] t=2,ΔP 21 =2.1474,ΔP 22 =0.27742

[0049] t=3,ΔP 31 =2.2495,ΔP 32 =0.39239

[0050] t=4,ΔP 41 =1.6432, ΔP 42 =0.97601

[0051] t=5,ΔP 51 =1.6432, ΔP 52 =0.97601

[0052] t=6,ΔP 61 =2.1174,ΔP 62 =0.61029

[0053] t=7,ΔP 71 =2.3412, ΔP 72 =0.35956

[0054] t=8,ΔP 81 =2.4281,ΔP 82 =0.56352

[0055] t=9,ΔP 91 =2.2875,ΔP 92 =0.74203

[0056] t=10,ΔP 101 =2.3514, ΔP 102 =0.60511

[0057] t=11,ΔP 111 =2.3207, ΔP 112 =0.46842

[0058] t=12,ΔP 121 =0.56078,ΔP 122 =0.96551

[0059] t=13,ΔP 131 =3.6347,ΔP 132 =1.2236

[0060] t=14,ΔP 141 =3.8910,ΔP 142 =0.29396

[0061] t=15,ΔP 151 =3.0685,ΔP 152 =2.0228

[0062] t=16,ΔP 161 =5.0408,ΔP 162 =2.4304

[0063] t=17,ΔP 171 =5.6217,ΔP 172 =2.2415

[0064] t=18,ΔP 181 =5.4590,ΔP 182 =2.6656

[0065] t=19,ΔP 191 =7.0416,ΔP 192 =2.4205

[0066] t=20,ΔP 201 =6.6681,ΔP 202 =2.1054

[0067] t=21,ΔP 211 =7.8441,ΔP 212 =2.8859

[0068] t=22,ΔP 221 =8.0909,ΔP 222 =3.0854

[0069] t=23,ΔP 231 =8.7960,ΔP 232 =3.1901

[0070] t=24,ΔP 241 =8.6471,ΔP 242 =3.5166

[0071] (3) Calculate the instantaneous flow Q of the calculation site corresponding to each hour according to formula (3): t ,calculate

[0072] The result is shown below

[0073] t=1,Q1=1.6827

[0074] t=2,Q2=1.4549

[0075] t=3,Q3=1.5851

[0076] t=4,Q4=1.5715

[0077] t=5,Q5=1.6366

[0078] t=6,Q6=1.6204

[0079] t=7,Q7=1.7949

[0080] t=8,Q8=1.8177

[0081] t=9,Q9=1.7739

[0082] t=10,Q 10 =1.6734

[0083] t=11,Q 11 =0.91577

[0084] t=12,Q 12 =1.4466

[0085] t=13,Q 13 =2.1582

[0086] t=14,Q 14 =0.62742

[0087] t=15,Q 15 =1.5662

[0088] t=16,Q 16 =2.0281

[0089] t=17,Q 17 =1.6760

[0090] t=18,Q 18 =2.7726

[0091] t=19,Q 19 =2.7376

[0092] t=20,Q20 =2.9749

[0093] t=21,Q 21 =3.0033

[0094] t=22,Q 22 =3.3635

[0095] t=23,Q 23 =3.0782

[0096] t=24,Q 24 =0.72048

[0097] (4) According to formula (4), the instantaneous flow rate of each hour is accumulated to obtain the cumulative flow rate per hour, and the cumulative flow rate SQ of the tth hour is calculated in turn. tP ;

[0098] t=1,SQ 1P =431107.6827

[0099] t=2,SQ 2P =431109.1376

[0100] t=3,SQ 3P =431110.7228

[0101] t=4,SQ 4P =431112.2943

[0102] t=5,SQ 5P =431113.9310

[0103] t=6,SQ 6P =431115.5514

[0104] t=7,SQ 7P =431117.3464

[0105] t=8,SQ 8P =431119.1641

[0106] t=9,SQ 9P =431120.9380

[0107] t=10,SQ 10P =431122.6115

[0108] t=11,SQ 11P =431123.5273

[0109] t=12,SQ 12P=431124.9740

[0110] t=13,SQ 13P =431127.1322

[0111] t=14,SQ 14P =431127.7596

[0112] t=15,SQ 15P =431129.3259

[0113] t=16,SQ 16P =431131.3540

[0114] t=17,SQ 17P =431133.0301

[0115] t=18,SQ 18P =431135.8027

[0116] t=19,SQ 19P =431138.5404

[0117] t=20,SQ 20P =431141.51533

[0118] t=21,SQ 21P =431144.5186

[0119] t=22,SQ 22P =431147.8822

[0120] t=23,SQ 23P =431150.9604

[0121] t=24,SQ 24P =431151.6809

[0122] (5) According to formula (5), the cumulative flow SQ at the tth hour is tP The measured cumulative flow SQ at the corresponding time t0 The relative error calculation is performed to obtain the relative error δ at this moment. t ,

[0123] There are 24 in total, and the calculation results are as follows:

[0124] t=1,δ1=0.00039%

[0125] t=2,δ2=0.45%

[0126] t=3,δ3=0.91%

[0127] t=4,δ4=1.39%

[0128] t=5,δ5=1.91%

[0129] t=6,δ6=2.55%

[0130] t=7,δ7=3.34%

[0131] t=8,δ8=4.31%

[0132] t=9,δ9=5.35%

[0133] t=10,δ 10 =5.35%

[0134] t=11,δ 11 =5.35%

[0135] t=12,δ 12 =5.35%

[0136] t=13,δ 13 =5.35%

[0137] t=14,δ 14 =5.35%

[0138] t=15,δ 15 =10.60%

[0139] t=16,δ 16 =11.31%

[0140] t=17,δ 17 =11.98%

[0141] t=18,δ 18 =12.62%

[0142] t=19,δ 19 =13.24%

[0143] t=20,δ 20 =13.81%

[0144] t=21,δ 21 =14.27%

[0145] t=22,δ 22 =14.71%

[0146] t=23,δ 23 =15.09%

[0147] t=24,δ 24 =15.45%

[0148] (6) Compare the sizes of the 24 relative errors and record the maximum relative error δ under this adjustment parameter α1 max ;

[0149] δ max =15.45%

[0150] (7) Taking 0.01 as the step size of the empirical parameter, the i-th empirical parameter α is obtained in turn. i , repeat steps (1) to (7) and record the maximum relative error under i different empirical parameters. The calculation results are as follows:

[0151] α1=0.01,δ1=15.45%

[0152] α2=0.02,δ2=15.44%

[0153] α3=0.03,δ3=15.43%

[0154] α4=0.04,δ4=15.42%

[0155] …

[0156] α 1680 =16.8,δ 1680 =2.71%

[0157] α 1681 =16.81,δ 1681 =2.70%

[0158] α 1682 =16.82,δ 1682 =2.71%

[0159] α 1683 =16.83,δ 1683 =2.70%

[0160] α 1684 =16.84,δ 1684 =2.69%

[0161] …

[0162] α 3628 =36.28,δ 3628 =17.04%

[0163] α 3629 =36.29,δ 3629 =17.05%

[0164] α 3630 =36.3,δ 3630 =17.07%

[0165] α 3631 =36.31,δ 3631 =17.06%

[0166] α 3632 =36.32,δ 3632 =17.07%

[0167] …

[0168] (8) Compare the i maximum relative errors, and take the empirical parameter corresponding to the minimum value among the i as the optimal empirical parameter for this time period. The calculation results are as follows:

[0169] Minimum value of the maximum relative error: δ1 = 2.69%

[0170] Optimal empirical parameter: α = 16.84

[0171] According to the calculation method in the sixth paragraph, the optimal empirical parameters and relative errors for other time periods are calculated. The calculation results are shown in the table below.

[0172]

[0173] The relative error δ1 of the first period is very large, up to 57.34%, and the corresponding adjustment parameter α1=10.15, which means that there is a large deviation between the predicted cumulative flow value obtained after tracing the data of six days and the actually measured cumulative flow value. It is not appropriate to use this empirical parameter to predict the cumulative flow in the seventh period; the relative error δ2=8.86% of the second period, and the empirical parameter α2=21.03; the relative error δ3=5.02% of the third period, and the empirical parameter α3=22.00; the relative error δ4=3.27% of the fourth period, and the empirical parameter α4=21.66; the relative error δ5=3.42% of the fifth period, and the empirical parameter α5=20.38; the relative error δ6=2.69% of the sixth period, and the empirical parameter α6=16.84. In order to more accurately predict the cumulative flow of the seventh period, the empirical parameters corresponding to the time period with the smallest relative error are selected for prediction, that is, the sixth period is selected: the relative error δ6 is only 2.69%, and the empirical parameter α6 is adjusted to 16.84.

[0174] The cumulative flow on August 2 is predicted using the adjustment parameter α6 of the sixth period, and the predicted cumulative flow SQ P , and then the cumulative flow SQ0 measured by the instrument is compared with the predicted cumulative flow SQ P Comparative verification is performed, and the results are as follows Figure 3 The square dotted line in the figure represents the measured cumulative flow SQ0, and the circular dotted line represents the predicted cumulative flow SQ P, the solid line represents the relative error curve, and the left ordinate represents the cumulative flow rate (unit: Nm 3 / h), the ordinate on the right represents the relative error value, and the abscissa represents time. Figure 3 It shows that the predicted cumulative flow and the measured cumulative flow have the same change trend. Through the analysis of relative error, it is known that the maximum relative error between the two is δ 7max Only 1.59%, the minimum value δ 7min This indicates that the optimal empirical parameters obtained by fitting the cumulative flow in the previous period and the measured cumulative flow can quickly and accurately predict the cumulative flow in the future period, indicating that the cumulative flow prediction and assignment algorithm of the established natural gas local pipeline network has a certain accuracy in predicting the natural gas flow measurement.

[0175] In summary, the present invention can effectively predict the cumulative flow of upstream and downstream users. The established assignment algorithm can be extended to the entire natural gas pipeline network, and can effectively predict the local cumulative flow of the natural gas pipeline network over a period of time. It has a certain reliability, which is beneficial to the operation and management analysis of natural gas pipeline enterprises and lays the foundation for flow measurement in natural gas transactions.

[0176] The present invention is not limited to the above-mentioned embodiments. All equivalent changes and modifications made within the scope of application of the present invention should fall within the scope of the present invention.

Claims

1. A method for quickly predicting the cumulative flow of a natural gas pipeline network, characterized in that The method comprises the following steps: It is known that the measurement time of each day is every hour, and each hour has the corresponding measured pressure data at both ends of the pipeline and the accumulated flow of the site users; The pressure data at both ends of each pipe section are taken as known quantities. The pipes adjacent to the calculation site are pipe 1 and pipe 2. The pressure difference at both ends of pipe 1 and pipe 2 is calculated according to formula (1) and formula (2) respectively. ΔP t1 =|P t1 2 -P t3 2 | (1) Where ΔP t1 represents the square difference of pressure at both ends of pipeline 1 at the tth hour; P t1 , P t3 represents the pressure at both ends of pipeline 1 at the tth hour; ΔP t2 =|P t4 2 -P t2 2 | (2) Where ΔP t2 represents the square difference of the pressure at both ends of pipeline 2 at the tth hour; P t2 , P t4 represents the pressure at both ends of pipeline 2 at the tth hour; According to formula (3), the instantaneous flow Q of the calculation station at the tth hour is established t The relationship between the pressure difference at both ends of the node is used to obtain the instantaneous flow rate Q corresponding to the tth hour. t : Q t =α|ΔP t1 -ΔP t2 | (3) Where Q t represents the instantaneous flow of the computing node at the tth hour; α represents the empirical parameter of the computing site; ΔP t1 , ΔP t2 It represents the square difference of the pressure at both ends of pipeline 1 and pipeline 2 at the tth hour; According to formula (4), the instantaneous flow at each time node is accumulated in sequence to obtain the cumulative flow per hour: Where SQ tP It represents the cumulative flow of the calculation site in the tth hour; SQ0 represents the total cumulative flow of the previous period of time; m represents the total number of hours; Then the cumulative flow SQ of the calculation site at the tth hour is calculated tp The measured cumulative flow SQ at the known t hour t0 Calculate the relative error: Where δ t represents the relative error at the tth hour; The known actual flow test pressure data is divided into n time periods according to the day unit, wherein the first time period is selected from the first day to the n-1 day; the second time period is selected from the second day to the n-1 day; the third time period is selected from the third day to the n-1 day; and so on, the n-2 time period is selected from the n-2 day and the n-1 day; the n-1 time period is selected from the n-1 day; the n time period is selected from the n day, and the most suitable empirical parameters are fitted with the first n-1 time periods to predict the cumulative flow of the n time period; Calculate the corresponding optimal empirical parameter α for the first n-1 periods of time n-1 , by continuously iterating and calculating the empirical parameters, the iterative calculation process for each period of time is as follows: (1) Adjust the empirical parameter from 0.01 to iterate, with a step size of 0.01, and obtain the first empirical parameter α1; (2) According to formulas (1) and (2), the square difference of the pressure at both ends of each pipeline at the tth hour is calculated in turn, and the corresponding ΔP is obtained respectively. t1 , ΔP t2 ; (3) Calculate the instantaneous flow Q of the calculation site corresponding to each hour according to formula (3): t ; (4) According to formula (4), the instantaneous flow rate of each hour is accumulated to obtain the cumulative flow rate per hour, and the cumulative flow rate SQ of the tth hour is calculated in turn. tP ; (5) According to formula (5), the cumulative flow SQ at the tth hour is tP The measured cumulative flow SQ at the corresponding time t0 The relative error calculation is performed to obtain the relative error δ at this moment. t , there are m in total; (6) Compare the sizes of the m relative errors and record the maximum relative error δ under this adjustment parameter α1 max ; (7) Taking 0.01 as the step size of the empirical parameter, the i-th empirical parameter α is obtained in turn. i , repeat steps (1) to (6) and record the maximum relative error under i different empirical parameters; (8) Compare the i maximum relative errors, and take the empirical parameter corresponding to the minimum value among the i as the optimal empirical parameter α for this time period. n-1 ; The first n-1 periods of time each have a corresponding optimal adjustment parameter α n-1 And under this adjustment parameter, there is a maximum relative error δ in each period of time n-1 ; Compare the error δ of the previous n-1 periods n-1 , and the optimal adjustment parameter result corresponding to its minimum value is used as the prediction experience parameter α n , to predict the cumulative flow in the nth time period.

2. A method for rapidly predicting the cumulative flow of a natural gas pipeline network according to claim 1, characterized in that: For multi-pipeline site calculations, calculate the square difference of the pressures of the redundant pipes according to formulas (1) and (2), add the pressure difference of the redundant pipes in formula (3), and finally calculate using the straight pipe section calculation method.

Images