Identification of partial blockages in pipelines using genetic algorithms

Abstract

A methodology to identify the partial blockages in a simple pipeline using genetic algorithms for non-harmonic flows is presented in this paper. A sinusoidal flow generated by the periodic on-and-off operation of a valve at the outlet is investigated in the time domain and it is observed that pressure variation at the valve is influenced by the opening size of blockage and its location. In this technique, the unsteady (steady oscillatory) pressure time series at only one location is required to identify two blockages. In the proposed methodology, the solution of the governing hyperbolic PDEs of pipe flow is obtained using the method of characteristics. For any piping system similar to the hypothetical pipe system used in the simulations, generalized best amplitude and best frequency of the valve operation are determined, which give maximum deviation in pressure responses for a specific blockage at different locations for a given constant-head reservoir. The generalized best amplitude and best frequency of the valve operation are also obtained for two blockages. Accuracy of the proposed methodology in identifying blockages in a hypothetical simple pipe system with increased noise in the simulated measurements is studied. A non-dimensional variable is proposed to determine whether the proposed methodology is applicable to isolate partial blockages in a piping system. Finally, the proposed methodology is experimentally validated on a laboratory piping system for a single blockage and two blockages.

Simulation and detection of blockage in a pipe under mean fluid flow using acoustic wave propagation technique

“…Moreover, the transmitted wave continued, and on reaching the pipe outlet, it was reflected backward and arrived back to pressure monitoring locations s 1 at t p = 20.001 ms. Thus, with the data recorded, the wave speed and the location of the blockage from the pipe inlet is calculated as follows: (12) where ℓ is the pipe length (m), l s is the distance from the pipe inlet to pressure monitoring location (m), c is the wave speed, t p is the time it takes the wave to travel from the inlet to the pressure monitoring location then to the pipe exit and back to pressure monitoring location (s n ), and t s is the time it takes the wave to travel from the inlet to pressure monitoring location s n ,and the location of the blockage from s 1 is given by…”

“…The relevant data for location s 1 to s 6 and the estimated location of the leakage from the data are shown in Table 2. The absolute distance of the acoustic pulse from the pipe inlet to the discontinuity in the pipe is calculated using Equation (12) with data generated by pressure monitoring location upstream of the defect and thus tabulated in the eighth column of the table. Furthermore, comparing the value of the predicted with the actual distance from the inlet to the defect location, the percentage error (E) is computed and tabulated in the ninth column of the table.…”

“…Figure 9b shows a magnified plot of the dynamic pressure versus time waveform predicted for pressure monitoring location s 1 . Considering the predicted time of flight data, the wave speed can be deduced for all the times predicted for pressure monitoring locations s 1 to s 6 using Equation (12). The wave speed c according to the time of flight for other pressure monitoring locations (s 1 ,s 2 ,s 3 ,s 4 ,s 5 ,and s 6 ) are computed and tabulated in Table 3.…”

Estimation and examination of linepack pressures in long liquid pipelines

“…Rapidly varying flow variables (water-hammer) in the conduit flows are functions of space x ð Þ and time t ð Þ coordinates. The spatial and temporal variations in pressure and velocity were calculated by solving the following continuity and momentum equations [11].…”

Corrosion Prediction for Naphtha and Gas System of Atmospheric Distillation Tower Based on Artificial Neural Network and Genetic Algorithm

“…In this paper, the tansig function is selected as the training function, in which the output function H of the hidden layer is determined by formula 2, where x is the input variable, w ij is the connection weight number between the input layer and the hidden layer, a is the threshold value of the hidden layer, initialization assignment of the weight and threshold is randomly selected within the interval (0, 1) [10].…”