Research Article

Mathematical Model Based on BP Neural Network Algorithm for the Deflection Identification of Storage Tank and Calibration of Tank Capacity Chart

Table 2

The oil reserve of oil-out and analysis of percentage error with BP neural network statistical table.

Site 1Site 2
OH/mmAOR/LTOR/LFENN/LPECOH/mmAOR/LTOR/LFENN/LPEC

1020.653464.743405.10 59.610.00%715.322214.742222.10 10.08 0.12%
1007.733414.743361.80 53.150.01%705.432164.742180.70 11.76 0.19%
994.323364.743315.90 48.640.01%693.522114.742130.60 13.97 0.09%
980.963314.743269.40 45.200.00%682.52064.742084.30 16.09 0.17%
967.103264.743220.10 42.050.08%671.022014.742036.00 18.31 0.15%
956.013214.743180.10 39.590.15%658.681964.741984.00 20.60 0.07%
941.543164.743127.20 36.340.04%647.741914.741938.00 22.50 0.04%
929.693114.743083.20 33.600.07%635.761864.741887.60 24.34 0.08%
916.443064.743033.40 30.450.03%624.611814.741840.70 25.80 0.01%
904.143014.742986.60 27.460.02%612.531764.741790.00 27.08 0.10%
891.92964.742939.50 24.470.03%600.691714.741740.30 27.99 0.14%
879.232914.742890.30 21.390.10%589.401664.741693.10 28.55 0.01%
868.992864.742850.20 18.960.15%577.001614.741641.30 28.82 0.14%
855.132814.742795.50 15.810.12%564.581564.741589.60 28.76 0.25%
844.022764.742751.20 13.450.00%554.331514.741547.10 28.48 0.26%
831.642714.742701.60 11.070.08%540.761464.741491.00 27.83 0.11%
820.472664.742656.50 9.200.04%528.651414.741441.20 27.02 0.04%
808.162614.742606.50 7.500.03%517.191364.741394.20 26.09 0.25%
796.002564.742556.90 6.240.06%504.871314.741344.10 24.97 0.33%
785.042514.742511.90 5.510.11%490.781264.741287.00 23.55 0.10%
773.072464.742462.60 5.180.12%478.061214.741235.90 22.20 0.09%
762.092414.742417.20 5.300.12%465.971164.741187.70 20.88 0.18%
750.812364.742370.30 5.860.01%452.401114.741134.00 19.39 0.01%
739.422314.742322.90 6.840.06%439.981064.741085.30 18.04 0.24%
727.092264.742271.408.320.07%425.831014.741030.3016.550.10%

Where OH, AOR, TOR, FENN, and PEC are, respectively, oil height, actual oil reserve, theoretical oil reserve, fitting error of neural network, and proportional error with correction.