Long-term acoustic and optical turbidity monitoring in a sewer

Abstract

In order to monitor these phenomena, the Suspended Solids Concentration is usually measured either by ad hoc analyzes on samples or continuously by optical turbidity. We compared these conventional techniques to explain, in terms of particle presence, the variations of real time acoustic turbidity. The Ultrasonic Doppler Velocity Profiler (UVP) also provides velocity and water height data. Major differences have been observed between dry weather and storm event conditions. .

Full text

Long-term acoustic and optical turbidity monitoring in a
sewer
A. Pallarès*, S. Fischer**, X. France***, M.N. Pons****, P. Schmitt*
* Laboratoire ICube – UMR 7357, Université de Strasbourg, 2, rue Boussingault, 67000 Strasbourg,
France (Email: anne.pallares@unistra.fr, pschmitt@unistra.fr)
** UBERTONE, 11, rue de l'Académie, 67000 Strasbourg (Email: stephane.fischer@ubertone.fr)
***GEMCEA, 149, rue Gabriel Péri, 54500, Vandoeuvre-les-Nancy (E-mail:
xavier.france@gemcea.org)
**** Laboratoire Réactions et Génie des Procédés – UMR CNRS 7274, Université de Lorraine, 1, rue
Grandville, BP 20451, F-54001 Nancy cedex, France (Email: marie-noelle.pons@univ-lorraine.fr)
In order to monitor these phenomena, the Suspended Solids Concentration is usually measured either by ad hoc
analyzes on samples or continuously by optical turbidity. We compared these conventional techniques to explain, in
terms of particle presence, the variations of real time acoustic turbidity. The Ultrasonic Doppler Velocity Profiler (UVP)
also provides velocity and water height data.
Major differences have been observed between dry weather and storm event conditions.
.
Keywords: Acoustic, Backscattering, Doppler, Velocity, Turbidity, Sediment, Suspended Solids,
Wastewater.
1 INTRODUCTION
The knowledge of sediment transport characteristics is an important issue in terms of sewer and surface
water management. Indeed, the Suspended Solids (SS) transported by (waste)water are a vector of pollution
and they may also be physically damaging [Lazar et al., 2010, Bilotta et al. 2008]. A significant sedimentation
in structures can lead to progressive silting thereof. Appropriate flow management, in order to limit these
phenomena, with high temporal frequency suspended solids data, is needed.
In order to monitor these phenomena in sewage, the Suspended Solids Concentration (SSC) is usually
measured either by ad hoc analyzes on samples or continuously by optical turbidity.
Optical turbidity is the most commonly used continuous measurement technology for SSC estimation as well
in natural water flows like rivers [Anant et al. 2015, Haimann et al. 2014] or in sewer systems [Hannouche
2012]. Optical turbidity depends on the colour, size and shape of the SS. As widely discussed in [Downing
2006, Hannouche 2012], optical turbidity can, after adequate calibration, be linearly linked to the SSC.
However, it should be noted that it is a point measurement which might not be representative of the whole
flow. Furthermore the calibration is site specific. Its major drawback is its sensitivity to biofouling which leads
to a signal degradation [Flemming 2002].
Acoustic backscattering (or acoustic turbidity) is widely used in marine environment [Moate et al. 2012] and
rivers [Moore et al. 2012]. The use of multi-frequency instruments allows to monitor the particle size and
concentration. As shown in [Thorne et al. 2014, Wilson et al. 2015] and the references therein, inversion
techniques exist and are satisfying in flows with limited particle size and nature. This is unfortunately not the
case in wastewater for which some attempts have been made [Abda et al. 2009] but no systematic inversion
technique exists. Acoustic devices are quite insensitive to biofilm and have a main asset: the possibility to
measure the sediment characteristics over a cross-section.
Without regular calibration, both optical and acoustic turbidity give a rough estimation of the SSC. As in the
complex medium that is wastewater, inversion techniques are still discussed, we focused on the general
optical and acoustic turbidity trends during long time observation trying to highlight their similitudes or
differences.
2 BACKGROUND
The interesting fact with acoustic pulsed Doppler instruments is that the flow velocity, the backscattered
signal amplitude and the turbidity profiles can be simultaneously recorded on the insonified water column.

r
ts
rms eM
r
kk
V
a
2
2
1-
=
On the theoretical point of view [Shen et al. 1988, Thorne et al. 2012, Thorne et al. 2014], for an acoustic
flowmeter, the recorded root-mean-square voltage of the backscattered signal, or amplitude, can be written
at range r, distance from the transducer, as follows:
(1)
This expression stands for data taken outside the transducer near-field. a and ks are medium depending
variables. M is the particle concentration. a can be divided into two contributions: αw which is the attenuation
due to the water absorption tabulated through a semi-empirical formula [Fisher et al. 1977], and
a
s which is
the particle attenuation depending on the particle density, their mean radius and their acoustic normalized
total scattering cross-section. ks represents the particle backscattering properties and depends on the
backscattering characteristics of the particles, thus their nature, their density, and their radius.
In addition to amplitude, the UVP generates directly the corresponding acoustic turbidity. Given equation (1),
its expression would be:
r
s
r
t
rms sw
eMke
k
rV
T
aa
2
2
1
2-
==
(2)
In equation (2), the right part only depends on the particle characteristics. Thus, the acoustic turbidity directly
includes information about the particles encountered in the explored medium. An estimation of the water
height can easily be computed through the flow bottom echo in the acoustic turbidity.
When using multi-frequency backscattering measurements, it should be noted that when s
a
pl
2>
mostly
backscattering is observed and scattering losses rise rapidly with increasing sediment size. Therefore, if the
smallest particle to be detected has a radius amin, with c the speed of sound, the optimal choice of the
observation frequency F would be:
min
2a
c
F
p
=
(3)
3 EXPERIMENTAL SET-UP
3.1 Site description
The measurement campaign was done in the entry chamber of the wastewater treatment plant of Greater
Nancy (250 000 p.e.). Its reference flow is 120 000 m3/day. 65% of the water comes from a combined sewer
system. As seen on figure 1, all instruments or sampling devices are fixed on a floating arm located close to
one of the inflows.
Figure 1: Site set-up.
3.2 In-situ monitoring
An UVP, an UB-Flow 315 (Ubertone, Strasbourg, France), was used for the acoustic measurements. It was
floating on the water surface and looking down at the chamber bottom.

In parallel to the acoustic measurements, continuous optical turbidity was recorded every 5 minutes by a
Solitax-ts-line sc device (Hach-Lange, Lognes, France) also fixed on the floating arm as illustrated by the
figure 2. Its optical turbidity range is 0.001- 4000 FNU.
The reader should be aware that the buoy will move with the water level. This explains later on the variable
depth on the acoustic turbidity profiles.
Figure 2: Global instrumental set-up view.
The UVP and the optical turbidimeter provided continuous data from May to November 2014. Additional data
were available for pluviometry (rainfall over 5 minutes in mm/h) and water height in the chamber (MSP900
acoustic level sensor from Mobrey, Bron, France).
3.3 Sampling
According to the weather conditions, dry or rainy, punctual SSC measurements were also performed and
repeated. Samples were taken every hour during dry weather and every 15 minutes during a rain event.
Figure 3 shows the automatized wastewater sampling process. Once being sampled, the water analysis was
done according to the usual procedure NF EN872- 2005.
Figure 3: SSC sampling cycle.

4 EXPERIMENTAL RESULTS
The following table shows the different configurations used for the acoustic measurements. These
configurations were successively and continuously screened, the whole screening process taking around 6
minutes. Thus, for a given configuration, one profile is available every 6 minutes.
Each configuration has a different emission frequency, pulse repetition frequency, number of recorded
profiles, position of the first water cell with its depth and inter- cell distance, as well as number of cells and
number of samples used in each cell to evaluate one profile.
In configurations 1 to 4, the emission frequencies are different but the other characteristics
are similar in
order to match different suspended particles sizes. The optimal sensitivities given in Table 1 correspond to
the theoretically largest radius detected at that frequency. Configuration 5, with less statistics, is only used
for post-processed water height measurement.
Table 1. Settings of the different configurations
.
4.1 Long term overview
Figure 4 shows characteristic recordings of optical turbidity and acoustic turbidity at the two extreme
frequencies.
Figure 4.a) shows the most usual observation, here over more than six days of recording in dry weather
conditions. In this case, all turbidities reflect the daily cycle of human activity [Enfinger 2009]. Optical
turbidity has an average value of 150 NTU. The acoustic turbidity at 4.167 MHz is dominant, with a factor up
to 5 compared to the one at 0.815 MHz. This suggests that mainly small particles under 120 µm diameter are
present in the flow.
Figure 4.b) shows the same measurements over quite three days. Rainfall was observed at the beginning of
this period. The daily cycle of human activity can still be seen of the optical turbidity which has a slighter
lower average value than in Figure 4.a) (70 NTU). Compared to Figure 4.a), both acoustic turbidities
considerably increased, by a factor 100 compared to previously. Their growth completely masks the daily
cycle which can be seen after the rain event with a zoom on smaller turbidity values. In this situation, the
acoustic turbidity at 0.815 MHz is dominant suggesting the presence of bigger particles around 600 µm
diameter. The increase of the acoustical turbidities values could be explained by an increase in SSC or a
modification of the particle composition.
1 2 3 4 5
Frequency (MHz)
0.815 0.987 1.974 4.167 1.5
Position of the first
cell (mm)
21.3 21.3 13.7 13.8 13.8
Cell
depth (mm)
5.4 5.2 3.7 5.1 8.3
Inter
-
cell distance
(mm)
5.6 5.6 3.5 4.8 8.1
PRF (Hz)
699 799 1199 548 300
Number of
cell
s
182 160 168 200 200
Number of samples
per profile
128 128 128 128 8
Number of profiles
40 40 40 30 1
Optimal
particle
diameter sensitivity
(µm)
580 480 240 120 320

a)
b)
Figure 4: Acoustic and optical turbidity variation with time.
Figure 5 shows the long term evolution of the water height which was measured by a standard acoustic
water level sensor and by configuration 5 as a function of time. The time period covers roughly a month. The
peaks reflect rain events as correlated to the pluviometry data. As already observed, in regard to the
conventional technique, a systematic bias of about 20 cm is observed compared to the UVP data. This is
due to a sediment layer of about 20-30 cm at the bottom of the chamber [Pallarès et al. 2016].
Figure 5: Water height as a function of time.
As a clear weather dependency is observed, we focused on a more detailed weather associated data
analysis.

4.2 Dry weather
During dry weather conditions, urban sewage is a mix of grey-water (washing, food preparation, baths,
toilets, etc…) and industrial wastewater. On the later case pollutants will be site dependent and specifically
related to the industry type. The observed flowrate will be time dependent, reflecting the human activity cycle
[Enfinger 2009]. The daily flow rates are known to be regular. Minimal values are observed around 4 am and
maximal values will occur between 8 am and 1 pm [Dégrémont 1991, Henze et al. 2002].
Figure 6 shows the variations of acoustic turbidity with time over one dry weather day for the different
frequencies. The highest turbidity value is observed for the highest frequency, sensitive to particle under 120
µm diameter. A similar global variation is observed at all frequencies. As can be seen, all turbidities are
linked to human activities: a first increase is observed between 7 am and 1 pm followed by a second one
lasting from 8 to 11 pm. The peaks at 5 am suggest an industrial run-off.
Figure 6: Acoustic turbidity, at different frequencies, as a function of time under dry weather conditions.
Over the same period of time, we compared variations of the acoustic turbidity with those of sampling and
optical turbidity. The values of the acoustical turbidity, for the two extreme frequencies, 0.815 and 4.167 MHz
are displayed. On figure 7, one can see that all measurement techniques show the same global variation
trends with time.
Figure 7: Comparison of measurement techniques during dry weather conditions.

a)
b)
c)
Figure 8: Acoustic turbidity profile as a function of time during dry weather conditions.a- for 0.815 MHz, b- for 4.167
MHz,c-for 4.167 MHz over 24 hours

Figure 8 shows different aspects of the time variations of the acoustic turbidity. The upper part of each figure
shows the turbidity profile over depth as a function of time. The figures are restricted to the exploitable zone
going from a few centimeters after the transducer to close to the flow bottom. Thus the variation of acoustic
turbidity profile is here plotted between 9 and 14 cm for the lowest frequency and 5 and 12 cm for the
highest. Figure 8 a) shows the evolution for 0.815 MHz acoustic turbidity over several days. The daily
wastewater cycle can clearly be identified with quite clear water during the nights. One can see that the
acoustic turbidity clearly increases during day time. This is confirmed by b) showing the same evolution for
4.167 MHz. A zoom of b) over a 24 hour period is given in c). Between 4 and 8 am, the turbidity is quite null
showing a clear water flow. As soon as human activity begins, the turbidity sharply increases reaching a
maximum between 8:30 and 11 am. A second, smoother increase is observed between 8 and 11 pm. This is
coherent with other observations as for example [Abda et al. 2009]. At this frequency, one can also see that
the turbidity values are higher near the flow bottom. This is suggesting a concentration gradient during dry
weather periods.
4.3 Storm weather
Because of the presence of multiple impervious urban surfaces which do not allow rain to infiltrate into the
ground, runoff is generated. A first flush is the initial runoff of a rainstorm. During this phase, polluted water
entering storm drains in areas with high proportions of impervious surfaces is typically more concentrated
compared to the remainder of the storm. Consequently, these high concentrations of urban runoff result in
high levels of pollutants discharged from storm sewers to surface waters.
Figure 9 shows again the variations of the acoustic turbidity with time for the different frequencies over
roughly 24 hours. Compared to the dry weather, acoustic turbidity values are multiplied by 100 during a rain
event. The highest turbidity values are observed at 0.815 and 0.937 MHz. Mainly particles under 600 µm
diameter are observed in the flow. Approximatively half of these acoustic turbidity values are observed at
1.974 and 4.167 MHz corresponding to particle radius under 240 µm.
During storm weather, the acoustic turbidity strongly increases. To explain these variations, a huge increase
in SSC or/and a brutal change in particle characteristics is expected. The daily cycle is completely masked
by the radical increase of turbidity during storm weather.
Figure 9: Acoustic turbidity, at different frequencies, as a function of time during storm weather.
Figure 10 shows again the comparison of the different measurement techniques and the pluviometry data
during the two rain events. Both events show different behaviours regarding to the measurement technique:
no real correlation can be seen between SSC, acoustic and optical turbidities.
In both cases, the acoustic turbidity values are multiplied by 100 compared to the dry weather values.

Concerning the SSC seen by optical turbidity and sampling, a huge difference is observed for the rain event
of the 21th. For both rain events, final SSC are low compared to dry weather conditions.
The associated pluviometry data show that the measurements, even if both qualified as rain event, were
taken at different moments during the rainfall. Thus the storm event of the 8th July is linked to moderate
rainfall lasting a few hours. On the contrary, for the 21th, the sampling begins with a heavy rain peak soon
followed by a second one. Thus, the observations of the 21th can be associated to the first flush which also
explains the decreasing SSC.
As SSC decreases and observed acoustic turbidities increase, the nature of the particles present in the flow
during a storm event is completely different compared to the ones present during dry weather.
Figure 10: Comparison of measurement techniques during storm weather.

a)
b)
Figure 11: Acoustic turbidity profile as a function of time during storm weather conditions. a- for 0.815 MHz, b- for 4.167
MHz.
Figure 11 shows the variation of the acoustic turbidity profile as a function of time, centered on a storm
event. a) shows the evolution at 0.815 MHz and b) the one ay 4.167 MHz. The change of the acoustic
turbidity profile is here plotted between 8 and 23 cm for the lowest frequency and 4 and 20 cm for the
highest. One can see that in both cases, the acoustic turbidity is very high at the flow surface. It also varies
quickly from measurement to measurement showing a huge evolution in a couple of minutes. After the rain
event, the turbidity decreases at the water surface and increases at the flow bottom thus transcribing
sedimentation processes. A clue confirming this hypothesis is this phenomenum is more intense at the
lowest frequency, thus bigger particles.
5. Discussion
During dry weather conditions, the acoustic turbidity at 4.167 MHz is predominant suggesting the dominant
presence of fine particles with diameter under 120 µm. As has been observed previously, all particle survey
techniques show similar trends during dry weather conditions [Pallarès et al. 2015].
On the acoustic turbidity profile, a concentration gradient (higher turbidity values near the flow bottom) can
be seen especially during the morning hours. Thus a vertical concentration gradient during dry weather

conditions exists, phenomena which is in contradiction with other observations [Larrarte et al. 2011,
Randrianarimanana et al. 2015].
During storm weather, no common behaviour between particle survey techniques is observed. As can be
seen on the pluviometry data, one storm event can differ greatly from another making comparison difficult
even impossible. Anyhow, comparing the intensity of the signal for a same technique, it seems that sampling
and optical turbidity decrease by a factor 4 during an established rain event; on the contrary, the acoustic
turbidities increase by a factor 100 whatever the frequency is. Assuming the exactitude of the SSC value
obtained on the samples, their concentration goes from 200 to 40 mg/l. In the same time, the acoustic
turbidity is multiplied by 100.
As mentioned previously, only a drastic change in the nature of particle can explain the increase of acoustic
turbidities. The acoustic scattering varies with the composition of the particles [Moate et al. 2012]. Thus, we
suppose a high proportion of mineral particles during a storm event compared to the rather organic fraction
present during dry weather [Pallarès et al. 2011]. Urban storm water is known to be source of heavy metal
pollution [Davis et al. 2001, Dujcik et al. 2016, Beck et al. 2012, Herngen et al. 2005, Gromaire et al. 1998].
In combination with the presence of mineral particles, the presence of metallic compounds could also explain
the rise of acoustic turbidity.
In comparison, it seems that sampling and optical turbidity are mainly sensitive to the particle sizes. It is well
known that the larger the particle size, the less intense will be the scattering of light thus the optical turbidity
[Merten et al. 2014]. The decrease in optical turbidity is here confirmed by the frequency sensitivity of the
acoustic turbidity which shows the presence of coarser particles as during dry weather.
During storm weather, the different instruments clearly don’t see the same things. The acoustic signal is very
sensitive to the particle nature where optical turbidity mainly depends on particle size. This again questions
the representativeness of measurements and also the comparability of storm events.
5 CONCLUSION
This paper is an attempt to prove the usefulness of acoustic SS sewage survey without data inversion. As
demonstrated, it can be considered as a robust qualitative description of the flow. When using multi-
frequency devices as in this study, the behaviour of the signal at the different frequencies give indications
about the size and the nature of the particles contained in the flow.
The acoustic turbidity signal can clearly be used as storm weather indicator during which it increases, at all
frequencies, of a factor 100. This is not the case for optical turbidity which not shows such significative
differences.
It also brings out the flow evolution time and spatial scale. One can see that particle concentration, through
the acoustic turbidity, seems rather uniform during dry weather conditions and very variable during storm
weather. As a turbidity profile of the illuminated flow section is available, indications about the SSC profile is
available. During dry weather, acoustic data suggest the existence of an increasing concentration gradient
with depth. During storm weather, the mineral particles seem concentrated at the surface, in the first 20 cm
of the flow.
The UVP water height measurements are coherent with other external means. By comparison of both
values, information about an eventual sediment layer at the bottom of the flow can be obtained.

REFERENCES
[Lazar et a. 2010] Lazar A.N., Butterfield D, Futter M.N., Rankinen K, Thouvenot-Korppoo M, Jarritt N, Lawrence D.S.,
Wade A.J., Whitehead P.G. (2010) An assessment of the fine sediments dynamics in an upland river system: INCA-Sed
modifications and implications for fisheries. Science of the Total Environment, 408, 2555-2566.
[Anant et al. 2015] Anant Kr Rai, Arun Kumar (2015) Continuous measurement of suspended sediment
concentration: Technological advancement and future outlook, Measurement, 76, 209–227.
[Haimann et al. 2014] M. Haimann, M. Liedermann, P. Lalk, H. Habersack (2014) An integrated suspended sediment
transport monitoring and analysis concept, Int. J. Sedim. Res. 29,135–148.
[Hannouche 2012] Ali Hannouche, Analyse du transport solide en réseau d'assainissement unitaire par temps de pluie :
exploitation de données acquises par les observatoires français en hydrologie urbaine, Université Paris-Est, 2012.
[Bilotta et al. 2008] Bilotta G.S., Brazier R.E. (2008) Understanding the influence of suspended solids on waterquality and
aquatic biota, Water Research, 42, 2841-2861.
[Downing 2006] Downing J. (2006) Twenty-five years with OBS sensors: The good, the bad,and the ugly, Continental
Shelf Research, 26, 2299–2318.
[Flemming 2002] Flemming H.C., Biofouling in water systems – cases, causes and countermeasures, Appl. Microbiol.
Biotechnol. (2002) 59:629–640
[Moate et al. 2012] Benjamin D.Moate, Peter D.Thorne (2012) Interpreting acoustic backscatter from suspended
sediments of different and mixed mineralogical composition, Continental Shelf Research 46, 67–82.
[Moore et al. 2012] Moore S.A., LeCoz J., Hurther D., Paquier A., (2012) On the application of horizontal ADCPs to
suspended sediment transport surveys in rivers, Continental Shelf Research, 46, 50-63.
[Thorne et al. 2014] Thorne, P.D. and Hurter, D. (2014), An overview on the use of backscattered sound for measuring
suspended particle size and concentration profiles in non-cohesive inorganic sediment transport studies, Continental
Shelf Research, 22, 603-632.
[Wilson et al. 2015] G.W.Wilson, A.E.Hay (2015) Acoustic backscatter inversion for suspended sediment concentration
and size: A new approach using statistical inverse theory, Continental Shelf Research 106, 130–139.
[Abda et al. 2009] Abda F., Azbaid A, Ensminger D, Fischer S, François P, Schmitt P, Pallarès A. (2009) Ultrasonic
device for real-time sewage velocity and suspended particles concentration measurements, Water Sci Technol., 60, 117-
25.
[Davis et al. 2001] Davis A. P., Shokouhian M., Ni S. (2001) Loading estimates of lead, copper, cadmium, and zinc in
urban runoff from specific sources, Chemosphere, 44(5), 997-1009.
[Djukić et al. 2016] Djukić, A., Lekić, B., Rajaković-Ognjanović, V., Veljović, D., Vulić, T., Djolić, M., Prodanović, D. (2016)
Further insight into the mechanism of heavy metals partitioning in stormwater runoff, Journal of environmental
management, 168, 104-110.
[Beck et al. 2012] Beck, H. J., & Birch, G. F. (2012) Spatial and temporal variance of metal and suspended solids
relationships in urban stormwater—implications for monitoring, Water, Air, & Soil Pollution, 223(3), 1005-1015.
[Herngren et al. 2005] Herngren, L., Goonetilleke, A., & Ayoko, G. A. (2005) Understanding heavy metal and suspended
solids relationships in urban stormwater using simulated rainfall, Journal of environmental management, 76(2), 149-158.
[Gromaire et al. 1998] Gromaire-Mertz, M. C., Garnaud, S., Gonzalez, A., & Chebbo, G. (1999) Characterisation of urban
runoff pollution in Paris, Water Science and Technology, 39(2), 1-8.
[Sheng et al. 1988] Sheng, J. and Hay, A. E. (1988), An examination of the spherical scatter approximation in aqueous
suspensions of sand, J. Acoust. Soc. Am., 83(2), 598-610.
[Thorne et al. 2002] Thorne, P.D. and Hanes, D.M. (2002), A review of acoustic measurement of small-scale sediment
processes, Continental Shelf Research, 22, 603-632.
[Fisher et al. 1977] Fisher, F.H., Simmons, V.P. (1977), Sound, absorption in sea water. Journal of the Acoustical
Society of America, 62, 558–564.
[Enfinger et al. 2009] Enfinger, K.L. and Mitchell, P.S. (2009) Sewer Sociology – San Diego Metropolitan Area,
Proceedings of the Pipeline Division Specialty Conference; San Diego, CA. American Society of Civil Engineers: Reston,
VA.
[Pallarès et al. 2014] A. Pallarès, S. Fischer, X. France, M.N. Pons, P. Schmitt (2014) Acoustic turbidity as online
monitoring tool for rivers and sewer networks, Flow Measurement and Instrumentation, Available online.
[Dégrémont 1991] Dégremont (1991) Mémento technique de l’eau, Lavoisier, Paris.
[Henze et al. 2002] Henze M, Harremoes P, Cour Jansen J, Arvin E.(2002) Wastewater treatment: Biological
and chemical processes. Berlin: Springer.
[Pallarès et al. 2011] A. Pallarès, P. François, M.N. Pons and P. Schmitt (2011) Suspended particles in wastewater: their
optical, sedimentation and acoustical characterization and modelling, Water Science & Technology, 63.2, 240-247.
[Merten et al. 2014] Gustavo Henrique Merten & Paul D. Capel & Jean P. G. Minella (2014) Effects of suspended
sediment concentration and grain size on three optical turbidity sensors, J Soils Sediment, 14, 1235–1241.
[Larrarte et al. 2011] F. Larrarte, M.-N. Pons (2011) Suspended solids concentration in wastewater: Influence of sampling
conditions, Urban Water Journal, 8, 397-404.
[Randrianarimanana et al. 2015] J.J. Randrianarimanana, L. Lebouc, E. Szturyz, F. Larrarte, Experimental study of the
velocity and suspended solids distribution in wastewater system, E-proceedings of the 36th IAHR World Congress
28 June – 3 July, 2015, The Hague, the Netherlands.