Diagnosis of the Technical Condition of Standing Wave Support for Pipeline

Abstract

Field experiments were conducted to record acoustic noise on the surface of an above-ground pipeline (existing heating main). Research was carried out on sections of the pipeline with various types of fastening—rigid (pipes welded to the support) and non-rigid (heat-insulating tube freely lies on the support rack). Experiments have shown that noise waves can create certain frequencies and shapes of bending waves that occur in spans of a pipeline. Depending on how rigid and reliable the fasteners are, they can be used to diagnose sections of the pipeline by acoustic noise. In computer simulation by the finite element method, the frequencies of bending waves were obtained, which are close to some experimental ones. The interaction between the nodes and beams of bending waves passing along the spans of pipes with different types of fastening to quality is consistent with laboratory experiments conducted earlier.

Periodically occurring accidents on pipelines cause damage to the environment, increase land pollution and sometimes human losses.

A significant decrease in the rigidity of the pipe fastening to the support leads to an actual increase in the length of the span of the pipeline, which can lead to damage or even destruction of this pipeline section. On the other hand, such an increase in the span length should lower its natural frequencies, which can serve as an indicator of a decrease in the stability of a pipeline section at an early stage of this process, when its consequences are not yet visible by naked eye.

In this paper (Kolesnikov et al. 2012), the results of physical modeling showed that the acoustic noise data recorded on the surface of both empty and fluid-filled pipes can determine the frequencies and modes of oscillations of bending standing waves generated between their supports. To do this, it is sufficient to accumulate a large number of amplitude spectra of acoustic noise records. This information makes it possible to unambiguously diagnose a violation of a rigid pipe fastening to a support, since in places of rigid fastenings there should be no oscillations of standing waves (these are their nodal points), and if there is a violation of hard contact, vibrations in such places can be observed at least at some natural frequencies of the pipe.

To determine the natural frequencies of the pipeline elements, various techniques are used. For example, this can be done by exciting vibrations using artificial sources such as shock (Al-Sahib et al. 2010) or mechanical vibrations (Li and Guo 1990; Al-Sahib et al. 2010). This paper presents the results of a full-scale experiment demonstrating the possibility of diagnosing the technical condition of pipeline supports by standing waves generated in pipes by acoustic noise.

Two spans of the existing above-ground pipeline (heat pipeline) were selected for experimental work. The pipeline consists of two parallel steel pipes with a diameter of 46 cm, partially covered with thermal insulation materials. Every 10 m pipes are rigidly welded to massive steel supports (Fig. 1a), between which they are raised by an average of 25–30 cm above the surface of the earth, except for the intersection of sharp local landslides. with the exception of places where sharp local relief depressions intersect. In such places, pipes are laid without hard mount on higher steel stands (Fig. 1b).

Fig. 1

A a Pipeline section with rigid pipe fastening to the support b pipes laid on the rack without rigid fastening recorder and c geophone mounted on the pipe

A pipe filled with hot water flowing under pressure was selected for measurements. In one span, the pipe was rigidly fastened to the supports on both sides, as shown in Fig. 1a. In the second span, on one side there was a rigid fastening and on the other, the pipe was loosely laid on the stand shown in Fig. 1b. Measurements were carried out using vertical geophones GS-20DX and single-channel digital recorders RefTek-125A (Fig. 1c) with a sampling rate of 1 kHz. Geophones were attached to the pipe using magnetic disks. On each span, continuous noise recording was carried out for 10 min at points along the upper part of the pipe with a step of 20 cm along the entire span length.

During processing, the noise recordings recorded at each observation point were broken down into fragments with a duration of approximately 8.2 s (8192 counts), the amplitude spectra of these fragments were calculated and averaged. As a result, sharp quasi-regular peaks appear on the averaged amplitude spectra, which, as it will be shown below, correspond to the bending standing waves. For more accurate determination of the frequencies of these peaks, generalized spectra averaged over all points of the profile were also constructed. A generalized spectrum for the passage of a pipe between two rigid supports is shown in Fig. 2.

Fig. 2

Generalized amplitude spectrum of noise recordings recorded on the pipe surface between two rigid supports

The joint visualization of the averaged spectra obtained for all observation points in each pipe span (Fig. 3) allows us to verify that the peaks observed in the amplitude spectra correspond to standing waves. It can be seen that at the frequencies of the spectral peaks (natural frequencies of the pipes), alternation of the maxima and minima of the spectral amplitudes is observed along the spans.

Fig. 3

(a) Distribution of averaged amplitude spectra of noise recordings on rigidly fixed supports on both sides of the pipe section and (b) on the pipe section rigidly fixed on one side and freely lying on the rack on the other side

Since, in the general case, standing waves of different types can form in the pipes, we compared the experimental results for a span with a pipe rigidly fixed on both sides (Figs. 2 and 3a) with the results of numerical simulation in the MSC Nastran finite element system. The comparison showed that the experimentally observed alternations of the maxima and minima of the spectral amplitudes correspond to the nodes and antinodes of the five modes of bending standing waves. The differences in the frequencies of standing waves obtained in the field experiment and as a result of numerical simulation do not exceed 5.5%.

Returning to the experimental results, we note that the nature of the distribution of nodes and antinodes along the span for the two cases under consideration is significantly different. The sharp quasiregular peaks in the spectrum for a pipe with two rigid fixtures (Fig. 3a) are located on the frequency axis approximately two times less often than for a pipe rigidly fixed on one side only (Fig. 3b).

This pattern means that in the second case, noise generated by bending standing waves are formed in a double span. This is also confirmed by the distribution features of the antinodes and nodes of standing waves along the spans. Thus, from Fig. 3a it follows that at the extreme points of the observation profile corresponding to the places of rigid fastening of the pipe to the supports, there are practically no vibrations, that is, these are the nodal points of all the bending standing waves formed in the span.

Figure 3b differs significantly from Fig. 3a. In addition to doubling the number of natural frequencies, the distribution of amplitude maxima and minima at these frequencies indicates the absence of hard contact at a point at 10 m, corresponding to the place where the pipe rests without rigid fastening on the rack shown in Fig. 1b. This is evidenced by the presence of antinodes in this place for every second mode of standing waves, which could not have been in the case of a rigid fastening in this place.

The results showed that acoustic noise recordings can be successfully used to diagnose the stability loss of sections of above-ground pipelines caused by a decrease in the rigidity of fixing pipe spans.