Acoustic technique for monitoring fouling in plate heat exchangers

Recently, many authors have studied changes in acoustic parameters to detect fouling in THEs. Though significant work has been done on ultrasonic/acoustic techniques for pipe fouling detection guided waves by Hay and Rose (2003), guided waves and acoustic impact by Lohr and Rose (2003), Lohr, Rose, and Tavossi (1999), normal beam and pulse-echo by Withers (1996) no application has yet been reported for PHE. Hay and Rose (2003), designed a new piezopolymer guided wave sensor aimed at detecting fouling on the inner surface of piping commonly used in the processing of dairy products. On other hand, Lohr and Rose (2003), investigated guided waves and acoustic impact for pipe fouling detection. In both cases simulated fouling was used by adding a viscous layer on the inner surface of the pipe. Trials were performed under a constant ambient temperature. Wither’s research (Withers, 1996) concentrated on the use of normal-beam ultrasonic through-transmission and pulse-echo velocity measurements to detect and quantify the extent of fouling inside tubular exchangers. Other methods commented upon by Withers include an acoustical-vibrational method in which the change in the nature of the frequency of vibration is related to the extent of fouling, and an optical method in which the reflection from the inner pipe wall is correlated to the extent of fouling. These investigations were carried out using THEs, a context in which acoustics’ sensors can be more easily adapted than in plate heat exchangers. Thus, a real breakthrough in fouling detection inside a PHE has not yet been reached, mainly due to the complex structure of the PHE and a lack of understanding of the fouling mechanism. An optimal monitoring method would accurately indicate the location and extent of the deposit inside the PHE. For this to be done in industry, this information would have to be acquired non-destructively, reproducibly and automatically as well as on-line, in situ, and in real time. In addition, the sensor’s mechanical design would have to take several criteria into account so that any device equipped with these sensors would be robust and easy to use. The objective of this research is to define and implement optimized acoustic instrumentation to monitor and measure the fouling phenomenon inside PHE under critical thermal, hygienic and pressure conditions. The developed technique was studied for a given product under a variety of experimental conditions.


The frequency response analysis (modal analysis) aims to determine the eigenfrequencies and the associated displacements of a given structure. These modes (frequencies) depend on the geometry and mechanical properties of the structure. Their determination is important in the design of structures subjected to excitations. Investigating vibrating plate’s modes helps understanding the influence of a mechanical load on the vibrations of these plates. The governing equations for a flat vibrating plate and the motion of this plate can be found in Rayleigh (1945). The fluid–structure interaction model can be derived from these basic equations and describes the structural motion in the presence of surrounding fluid. However, these equations will not be elaborated here and this study is limited to the determination of the eigenfrequencies for a flat vibrating plate. This is the first step to design the sensor used in this study

During preliminary trials, it was determined that for an acoustic wave to propagate through the PHE with a good signal-to-noise ratio, it must be of very low frequency with a strong amplitude at the emission source. On-site experiments by Lohr et al. (1999) have already shown that acoustic waves generated by acoustic impact can travel through 50 ft of pipe with a very good signal-to-noise ratio. Moreover, it’s possible to maintain a good signal-to-noise ratio while penetrating pipe joints. For this reason a mechanical pulse was used in this study (acoustic impact) to generate propagating acoustic waves on a frequency band ranging from around 0 kHz to 20 kHz. This mechanical pulse was generated by an electromagnet pushrod striking a given surface, and the intensity and periodicity of the shock were controlled by a computer. After choosing the source of the acoustic vibrations, several techniques were compared using acoustic sensors for receiving the transmitted wave. Making use of appropriate sensors to continuously monitor the state of fouling inside the PHE presents a real problem that needed to be solved. The sensors had to be both sensitive to low-frequency acoustic waves and unaffected by temperature fluctuations. These constraints led to the development of a specific sensor using composite materials with an embedded ferroelectric disc (D = 14 mm, e = 1 mm) (Fig. 1).

Fig. 1. Schematic representation of two sensors with a mechanical pulse generator fixed on an exchanger plate.

The chosen composite material (FR4 epoxy laminate) has a low thermal conductivity (0.11 W/m K at 300 K), which renders the entire structure resistant to small temperature fluctuations. This technique is based on the principle of setting in resonance all the mechanical structure of a reduced size sensor which is in contact with the exchanger plate. Some researchers like Nassar and Nongaillard (2004), Degertekin and Khuri-Yakub (1996), Shuyu (1997) and Nikolovski and Royer (1997) used a similar physical principle, but associated a tapered volume with the ceramic pieces, to concentrate the mechanical energy. The eigenfrequencies of the developed sensor are given (Table 2) according to numerical simulations (with FemLab 3.2) and experimental measurements. As it is designed, the sensor has proved to have a good sensitivity to low-frequency (0–20 kHz) acoustic waves. In addition, these adjustable sensors were specifically adapted to the geometry of the heat exchanger plates and were coupled to the PHE (Fig. 1).

An acoustic sensor on the upper section of the V2 exchanger triggers the data acquisition device, while the three sensors in the lower part receive the ultrasonic or acoustic energy and relay the signals to the acquisition-processing device (Fig. 4)

Fig. 4. Schematic representation of the system used to monitor PHE fouling.

When the acoustic impact encounters the trigger sensor, the acquisition device is activated, and the recording window is opened. At the same time, the other sensors are ‘‘listening” for the waveform that travels through the exchanger. When the waveform reaches the receiving sensor, it is recorded and analyzed by the processing device and compared to the response of the trigger sensor. The processing device then provides a real-time value for the power and speed of the waveform caught by the receiving sensor. Up to three different responses (channels) can be obtained by shifting the position of the receiving sensor. The different acoustic channels are numbered from right (Product inlet) to left (Product outlet) of the PHE

For example, Figs. 5 and 6 depict the waveform signals over time and the Power spectral densities for the trigger sensor (St) and central sensor (S1).

The evolution of the acoustic waves that propagated inside the exchanger during the fouling phase and the cleaning phase was studied. Specifically, the changes in the power and propagation time of the acoustic waves received by the acoustic sensors were analyzed. As mentioned above, PHE fouling was also monitored during these experiments by recording the evolution of the overall drop in pressure over time for each exchanger

At the beginning of fouling runs, very slow decreases in acoustic power are recorded for 30 min, with an initial value close to 7.5 and 7 lW, respectively for S0 and S2. This slight decrease can be attributed to a thin layer of irreversibly adsorbed proteins on clean metal surfaces. This indicates that no significant fouling occurred during the first period. At the same time, the pressure drop was constant, confirming the previous observation. After this period, the acoustic power suddenly decreases in function with time until t = 90 min. This decrease in acoustic power may be attributed to the growth of fouled layers. The growth of deposit interferes upon the acoustic waves characteristics

This may indicate that the rate of deposit is more important at the inlet of the PHE. The rise of the pressure drop confirms the growing of deposit. Modifications of rates of acoustic power changes may indicate that the formation of fouled layers is a balance between rates of deposition, rates of re-entrainment of solids from fouled layers into the bulk and particle breakage,

 This study describes a useful acoustic device based on a non-intrusive technique for monitoring the fouling of a plate heat exchanger in real-time. This technique relies on a low frequency acoustic approach that uses a mechanical pulse as a source of acoustic vibrations

In conclusion, the analysis of a low-frequency acoustic wave (0–20 kHz) propagated through the plates demonstrated a good sensitivity that could be exploited to monitor fouling

Read complete article by B. Merheb , G. Nassar , B. Nongaillard a , G. Delaplace b , J.C. Leuliet