# American Institute of Mathematical Sciences

doi: 10.3934/jimo.2019091

## Numerical solution to an inverse problem on a determination of places and capacities of sources in the hyperbolic systems

 1 Baku State University, Z. Khalilov, 23, Baku, Azerbaijan 2 Institute of Control Systems of Azerbaijan, National Academy of Sciences, B. Vahabzade, 9, AZ1141 Baku, Azerbaijan

* Corresponding author

Received  September 2018 Revised  April 2019 Published  July 2019

In the work we investigate the numerical solution to a class of inverse problems with respect to the system of differential equations of hyperbolic type. The specialties of considered problems are: 1) the impulse impacts are present in the system and it is necessary to determine the capacities and the place of their location; 2) the differential equations of the system are only related to boundary values, and arbitrarily; 3) because of the long duration of the object functioning, the exact values of the initial conditions are not known, but a set of possible values is given. The inverse problem under consideration is reduced to the problem of parametric optimal control without initial conditions with non-separated boundary conditions. For the solution it is proposed to use first-order optimization methods. The results of numerical experiments are given on the example of the inverse problem of fluid transportation in the pipeline networks of complex structure. The problem is to determine the locations and the volume of leakage of raw materials based on the results of additional observations of the state of the transportation process at internal points or at the ends of sections of the pipeline network.

Citation: Yegana Ashrafova, Kamil Aida-Zade. Numerical solution to an inverse problem on a determination of places and capacities of sources in the hyperbolic systems. Journal of Industrial & Management Optimization, doi: 10.3934/jimo.2019091
##### References:

show all references

##### References:
The scheme of pipeline network in the problem 1
Plots of the exact ($q_{loss} = q^{(1,2)^\ast }(t) = 50 - 10e^{- 0.0003 t} {m^3} /{hour}$ and obtained leak volume functions $q^{(1,2)^\ast }(t)$ for five various initial approximations $(\xi ^{(1,2)},\,\,q^{(1,2)}(t))_0^i ,i = 1,..,5.$ for the iteration process (10) in problem 1
Plots of exact ($q^{loss*}$) and computed leak volume functions ($q^*(t)$)for various suspected segments
Results of the solution to the problem 1 regarding leakage in the segment (1, 2) in the presence of noise for various initial states $(\xi ^{(1,2)},\,\,q^{(1,2)}(t))_0^i ,i = 1,..,5.$
 $\xi _0^{(1,2)}$ 60 20 90 10 45,685 $\eta$ $q_0^{(1,2)} (t)$ $90 - 10e^{3\alpha t}$ $20 -10e^{3\alpha t}$ $30-10e^{3\alpha t}$ $66 + 20e^{3\alpha t}$ $66 + 20e^{3\alpha t}$ $\tilde {\Phi _0}$ 76.104 16.704 11.664 42.48 57.636 $\tilde {\Phi }$ 5.73$\cdot 10^{ - 7}$ 1.26$\cdot 10^{ - 7}$ 3.19$\cdot 10^{ - 6}$ 1.85$\cdot 10^{ - 6}$ 7.43$\cdot 10^{ - 7}$ 0% $\tilde {\xi }^{(1,2)}$ 30.003 29.998 30.008 29.994 29.998 $N_{iter}$ 6 5 16 14 8 $\delta \xi ^{(1,2)}$ 0.00009 0.00006 0.0003 0.0002 0.00006 $\delta q^{(1,2)}$ 0.0003 0. 0006 0.0003 0.0002 0.0003 $\tilde {\Phi _0}$ 76.352 16.837 11.683 42.148 57.732 $\tilde {\Phi }$ 0.023 0.014 0.024 0.017 0.020 0.5% $\tilde {\xi }^{(1,2)}$ 29.841 30.332 30.068 29.654 29.796 $N_{iter}$ 6 5 14 12 7 $\delta \xi ^{(1,2)}$ –0.005 0.011 0.002 –0.011 –0.006 $\delta q^{(1,2)}$ 0.043 0.030 0.049 0.041 0.042 $\tilde{\Phi _0}$ 77.119 16.924 11.832 43.744 57.413 $\tilde {\Phi }$ 0.067 0.071 0.073 0.062 0.065 1% $\tilde {\xi }^{(1,2)}$ 28.527 29.392 29.923 30.597 29.839 $N_{iter}$ 6 5 14 12 7 $\delta \xi ^{(1,2)}$ –0.049 –0.020 –0.002 0.020 –0.005 $\delta q^{(1,2)}$ 0.109 0.092 0.107 0.094 0.093
 $\xi _0^{(1,2)}$ 60 20 90 10 45,685 $\eta$ $q_0^{(1,2)} (t)$ $90 - 10e^{3\alpha t}$ $20 -10e^{3\alpha t}$ $30-10e^{3\alpha t}$ $66 + 20e^{3\alpha t}$ $66 + 20e^{3\alpha t}$ $\tilde {\Phi _0}$ 76.104 16.704 11.664 42.48 57.636 $\tilde {\Phi }$ 5.73$\cdot 10^{ - 7}$ 1.26$\cdot 10^{ - 7}$ 3.19$\cdot 10^{ - 6}$ 1.85$\cdot 10^{ - 6}$ 7.43$\cdot 10^{ - 7}$ 0% $\tilde {\xi }^{(1,2)}$ 30.003 29.998 30.008 29.994 29.998 $N_{iter}$ 6 5 16 14 8 $\delta \xi ^{(1,2)}$ 0.00009 0.00006 0.0003 0.0002 0.00006 $\delta q^{(1,2)}$ 0.0003 0. 0006 0.0003 0.0002 0.0003 $\tilde {\Phi _0}$ 76.352 16.837 11.683 42.148 57.732 $\tilde {\Phi }$ 0.023 0.014 0.024 0.017 0.020 0.5% $\tilde {\xi }^{(1,2)}$ 29.841 30.332 30.068 29.654 29.796 $N_{iter}$ 6 5 14 12 7 $\delta \xi ^{(1,2)}$ –0.005 0.011 0.002 –0.011 –0.006 $\delta q^{(1,2)}$ 0.043 0.030 0.049 0.041 0.042 $\tilde{\Phi _0}$ 77.119 16.924 11.832 43.744 57.413 $\tilde {\Phi }$ 0.067 0.071 0.073 0.062 0.065 1% $\tilde {\xi }^{(1,2)}$ 28.527 29.392 29.923 30.597 29.839 $N_{iter}$ 6 5 14 12 7 $\delta \xi ^{(1,2)}$ –0.049 –0.020 –0.002 0.020 –0.005 $\delta q^{(1,2)}$ 0.109 0.092 0.107 0.094 0.093
The results of the solution to the problem 1 regarding leakage in the section (1, 2) for different $\nu$ and for different initial states $(\xi ^{(1,2)},\,\,q^{(1,2)}(t))_0^i ,i = 1,..,5.$
 $\xi _0^{(1,2)}$ 60 20 90 10 45.685 $\nu$ $q_0^{(1,2)} (t)$ $90-10e^{3\alpha t}$ $20-10e^{3\alpha t}$ $30-10e^{3\alpha t}$ $66 + 20e^{3\alpha t}$ $66 + 20e^{3\alpha t}$ $\tilde{\Phi _0}$ 155.248 22.323 22.323 84.303 118.809 $\tilde {\Phi}$ 8.2$\cdot 10^{ - 4}$ 2.7$\cdot 10^{ - 4}$ 1.0$\cdot 10^{ - 3}$ 2.2$\cdot 10^{ - 4}$ 7.4$\cdot 10^{ - 5}$ 3 $\tilde {\xi }^{(1,2)}$ 30.145 29.996 30.190 30.022 30.029 $\delta \xi ^{(1,2)}$ 0.004 0.0001 0.006 0.0007 0.0005 $\delta q^{(1,2)}$ 0.0005 0.0004 0.0001 0.0009 0.0005 $\tilde{\Phi _0}$ 155.390 33.608 22.443 84.086 118.854 $\tilde {\Phi }$ 1.6$\cdot 10^{ - 4}$ 3.9$\cdot 10^{ - 4}$ 4.3$\cdot 10^{ - 4}$ 2.8$\cdot 10^{ - 4}$ 6.4$\cdot 10^{ - 5}$ 5 $\tilde {\xi }^{(1,2)}$ 29.963 30.093 30.100 29.985 29.995 $\delta \xi ^{(1,2)}$ 0.001 0.003 0.003 0.0005 0.0001 $\delta q^{(1,2)}$ 0.0006 0.0003 0.0005 0.0004 0.0002
 $\xi _0^{(1,2)}$ 60 20 90 10 45.685 $\nu$ $q_0^{(1,2)} (t)$ $90-10e^{3\alpha t}$ $20-10e^{3\alpha t}$ $30-10e^{3\alpha t}$ $66 + 20e^{3\alpha t}$ $66 + 20e^{3\alpha t}$ $\tilde{\Phi _0}$ 155.248 22.323 22.323 84.303 118.809 $\tilde {\Phi}$ 8.2$\cdot 10^{ - 4}$ 2.7$\cdot 10^{ - 4}$ 1.0$\cdot 10^{ - 3}$ 2.2$\cdot 10^{ - 4}$ 7.4$\cdot 10^{ - 5}$ 3 $\tilde {\xi }^{(1,2)}$ 30.145 29.996 30.190 30.022 30.029 $\delta \xi ^{(1,2)}$ 0.004 0.0001 0.006 0.0007 0.0005 $\delta q^{(1,2)}$ 0.0005 0.0004 0.0001 0.0009 0.0005 $\tilde{\Phi _0}$ 155.390 33.608 22.443 84.086 118.854 $\tilde {\Phi }$ 1.6$\cdot 10^{ - 4}$ 3.9$\cdot 10^{ - 4}$ 4.3$\cdot 10^{ - 4}$ 2.8$\cdot 10^{ - 4}$ 6.4$\cdot 10^{ - 5}$ 5 $\tilde {\xi }^{(1,2)}$ 29.963 30.093 30.100 29.985 29.995 $\delta \xi ^{(1,2)}$ 0.001 0.003 0.003 0.0005 0.0001 $\delta q^{(1,2)}$ 0.0006 0.0003 0.0005 0.0004 0.0002
The results of the solution to the problem 1 regarding leakage in the section (5, 6) for different initial states
 $\xi _0^{(5,6)}$ $q_0^{(5,6)} (t)$ $\tilde {\xi }^{(5,6)}$ $\tilde{\Phi_0}$ $\tilde {\Phi}$ $N_{iter}$ 25 $5+10e^{3\alpha t}$ 15.219 2.556 3.55 $\cdot 10^{ - 5}$ 165 10 $30 +10e^{3\alpha t}$ 14.781 44.615 3.55 $\cdot 10^{ - 5}$ 204 22.56 $15+10e^{3\alpha t}$ 15.222 3.467 3.59 $\cdot 10^{ - 5}$ 197
 $\xi _0^{(5,6)}$ $q_0^{(5,6)} (t)$ $\tilde {\xi }^{(5,6)}$ $\tilde{\Phi_0}$ $\tilde {\Phi}$ $N_{iter}$ 25 $5+10e^{3\alpha t}$ 15.219 2.556 3.55 $\cdot 10^{ - 5}$ 165 10 $30 +10e^{3\alpha t}$ 14.781 44.615 3.55 $\cdot 10^{ - 5}$ 204 22.56 $15+10e^{3\alpha t}$ 15.222 3.467 3.59 $\cdot 10^{ - 5}$ 197
Results obtained by solving Problem 2
 $\xi ^{0}$ $q^{0} (t)$ Segment $\tilde{\xi }^{ks}$ $\Im _{0}$ $\tilde{\Im }$ Number of indices $(k,s)$ iterations (1, 2) 5.991 43.272 39.614 6 (5, 2) 2.279 139.536 60.386 58 25 $15+10e^{-0.0003t}$ (3, 2) 29.974 10.188 1.22$\cdot 10^{-5}$ 12 (5, 4) 2.244 54.185 37.958 79 (5, 6) 2.995 72.396 20.621 185 (1, 2) 5.986 46.836 39.614 5 (5, 2) 2.305 84.132 60.408 5 5 $5+10e^{-0.0003t}$ (3, 2) 29.935 32.148 2.44$\cdot 10^{-5}$ 12 (5, 4) 1.898 38.628 37.728 150 (5, 6) 2.947 21.786 20.621 72
 $\xi ^{0}$ $q^{0} (t)$ Segment $\tilde{\xi }^{ks}$ $\Im _{0}$ $\tilde{\Im }$ Number of indices $(k,s)$ iterations (1, 2) 5.991 43.272 39.614 6 (5, 2) 2.279 139.536 60.386 58 25 $15+10e^{-0.0003t}$ (3, 2) 29.974 10.188 1.22$\cdot 10^{-5}$ 12 (5, 4) 2.244 54.185 37.958 79 (5, 6) 2.995 72.396 20.621 185 (1, 2) 5.986 46.836 39.614 5 (5, 2) 2.305 84.132 60.408 5 5 $5+10e^{-0.0003t}$ (3, 2) 29.935 32.148 2.44$\cdot 10^{-5}$ 12 (5, 4) 1.898 38.628 37.728 150 (5, 6) 2.947 21.786 20.621 72
 [1] Hi Jun Choe, Do Wan Kim, Yongsik Kim. Meshfree method for the non-stationary incompressible Navier-Stokes equations. Discrete & Continuous Dynamical Systems - B, 2006, 6 (1) : 17-39. doi: 10.3934/dcdsb.2006.6.17 [2] Yinnian He, Yanping Lin, Weiwei Sun. Stabilized finite element method for the non-stationary Navier-Stokes problem. Discrete & Continuous Dynamical Systems - B, 2006, 6 (1) : 41-68. doi: 10.3934/dcdsb.2006.6.41 [3] Volodymyr O. Kapustyan, Ivan O. Pyshnograiev, Olena A. Kapustian. Quasi-optimal control with a general quadratic criterion in a special norm for systems described by parabolic-hyperbolic equations with non-local boundary conditions. Discrete & Continuous Dynamical Systems - B, 2019, 24 (3) : 1243-1258. doi: 10.3934/dcdsb.2019014 [4] Fredi Tröltzsch, Alberto Valli. Optimal voltage control of non-stationary eddy current problems. Mathematical Control & Related Fields, 2018, 8 (1) : 35-56. doi: 10.3934/mcrf.2018002 [5] Kaïs Ammari, Thomas Duyckaerts, Armen Shirikyan. Local feedback stabilisation to a non-stationary solution for a damped non-linear wave equation. Mathematical Control & Related Fields, 2016, 6 (1) : 1-25. doi: 10.3934/mcrf.2016.6.1 [6] Zhendong Luo. A reduced-order SMFVE extrapolation algorithm based on POD technique and CN method for the non-stationary Navier-Stokes equations. Discrete & Continuous Dynamical Systems - B, 2015, 20 (4) : 1189-1212. doi: 10.3934/dcdsb.2015.20.1189 [7] Ciprian G. Gal, Alain Miranville. Robust exponential attractors and convergence to equilibria for non-isothermal Cahn-Hilliard equations with dynamic boundary conditions. Discrete & Continuous Dynamical Systems - S, 2009, 2 (1) : 113-147. doi: 10.3934/dcdss.2009.2.113 [8] Franck Boyer, Pierre Fabrie. Outflow boundary conditions for the incompressible non-homogeneous Navier-Stokes equations. Discrete & Continuous Dynamical Systems - B, 2007, 7 (2) : 219-250. doi: 10.3934/dcdsb.2007.7.219 [9] Ciprian G. Gal, Mahamadi Warma. Reaction-diffusion equations with fractional diffusion on non-smooth domains with various boundary conditions. Discrete & Continuous Dynamical Systems - A, 2016, 36 (3) : 1279-1319. doi: 10.3934/dcds.2016.36.1279 [10] Lu Yang, Meihua Yang, Peter E. Kloeden. Pullback attractors for non-autonomous quasi-linear parabolic equations with dynamical boundary conditions. Discrete & Continuous Dynamical Systems - B, 2012, 17 (7) : 2635-2651. doi: 10.3934/dcdsb.2012.17.2635 [11] José M. Arrieta, Simone M. Bruschi. Very rapidly varying boundaries in equations with nonlinear boundary conditions. The case of a non uniformly Lipschitz deformation. Discrete & Continuous Dynamical Systems - B, 2010, 14 (2) : 327-351. doi: 10.3934/dcdsb.2010.14.327 [12] Ciprian G. Gal, Maurizio Grasselli. The non-isothermal Allen-Cahn equation with dynamic boundary conditions. Discrete & Continuous Dynamical Systems - A, 2008, 22 (4) : 1009-1040. doi: 10.3934/dcds.2008.22.1009 [13] Rafael Abreu, Cristian Morales-Rodrigo, Antonio Suárez. Some eigenvalue problems with non-local boundary conditions and applications. Communications on Pure & Applied Analysis, 2014, 13 (6) : 2465-2474. doi: 10.3934/cpaa.2014.13.2465 [14] Nikolai Dokuchaev. On forward and backward SPDEs with non-local boundary conditions. Discrete & Continuous Dynamical Systems - A, 2015, 35 (11) : 5335-5351. doi: 10.3934/dcds.2015.35.5335 [15] Mohamed Boulanouar. On a Mathematical model with non-compact boundary conditions describing bacterial population (Ⅱ). Evolution Equations & Control Theory, 2019, 8 (2) : 247-271. doi: 10.3934/eect.2019014 [16] Shitao Liu, Roberto Triggiani. Recovering damping and potential coefficients for an inverse non-homogeneous second-order hyperbolic problem via a localized Neumann boundary trace. Discrete & Continuous Dynamical Systems - A, 2013, 33 (11&12) : 5217-5252. doi: 10.3934/dcds.2013.33.5217 [17] Alessio Fiscella, Enzo Vitillaro. Local Hadamard well--posedness and blow--up for reaction--diffusion equations with non--linear dynamical boundary conditions. Discrete & Continuous Dynamical Systems - A, 2013, 33 (11&12) : 5015-5047. doi: 10.3934/dcds.2013.33.5015 [18] François Bolley, Arnaud Guillin, Xinyu Wang. Non ultracontractive heat kernel bounds by Lyapunov conditions. Discrete & Continuous Dynamical Systems - A, 2015, 35 (3) : 857-870. doi: 10.3934/dcds.2015.35.857 [19] Xiaoyu Fu. Stabilization of hyperbolic equations with mixed boundary conditions. Mathematical Control & Related Fields, 2015, 5 (4) : 761-780. doi: 10.3934/mcrf.2015.5.761 [20] Chao Xu, Yimeng Dong, Zhigang Ren, Huachen Jiang, Xin Yu. Sensor deployment for pipeline leakage detection via optimal boundary control strategies. Journal of Industrial & Management Optimization, 2015, 11 (1) : 199-216. doi: 10.3934/jimo.2015.11.199

2018 Impact Factor: 1.025