Dataset of near infrared spectroscopy measurements to predict rheological parameters of sludge

In the dataset presented in this article, 36 sludge samples were characterized. Rheological parameters were determined and near infrared spectroscopy measurements were realized. In order to assess the potential of near infrared spectroscopy to predict rheological parameters of sludge, Partial Least Square algorithm was used to build calibration models.


a b s t r a c t
In the dataset presented in this article, 36 sludge samples were characterized. Rheological parameters were determined and near infrared spectroscopy measurements were realized. In order to assess the potential of near infrared spectroscopy to predict rheological parameters of sludge, Partial Least Square algorithm was used to build calibration models.
& 2016 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

Subject area
Physics, Spectroscopy More specific subject area

Wastewater treatment
Type of data

Experimental features
Near Infrared Spectroscopy coupled with chemometric analysis was used to test the feasibility to predict rheological parameters of sludge samples.

Data accessibility
The data is available with this article

Value of the data
The data can be used as supplements on the physical properties of sludge and can be compared to other studies.
Those data establish a link between physical properties and reflectance spectra on various sludge samples.
Near infrared spectroscopy and multivariate analysis are able to predict rheological parameters of sludge.

Data
Several measurements on 36 sludge samples of different types (primary, secondary, digested, and dehydrated) were made. Rheological parameters (elastic and viscous moduli, yield stress, and viscosity) were determined (Table 1). In parallel, reflectance spectra were measured using an integrating sphere (Fig. 3). With a Partial Least Square (PLS) algorithm, predicting models were obtained for the dry matter (Fig. 4) and four rheological parameters (Figs. 5-8).

Sludge sample
36 sludge samples were collected in different wastewater treatment plants in France (Table 1). Consequently, a various panel of samples (primary, secondary, digested, and dehydrated) is available to construct the database. Moreover, knowing that sludges evolve over a large period of time, some samples were measured at different times over a period of 3 months. Additionally, two samples were mixed to create a new sludge. The database is so formed of 36 measurements. Finally, once collected, the samples were stored in sealed cans in the fridge before being characterized.
The dry matter of each sample was determined at 105°C for 24 h (Table 1).

Rheological measurements
A controlled stress rheometer (Mars II Thermofisher) was used with a coaxial cylinders geometry (R in ¼ 19 mm, H in ¼55 mm and R out ¼21.5 mm). In addition, both surfaces were rough, which avoids wall slip. The temperature was kept constant (at 20°C) through a thermostatic bath (C25P Haake).
The procedure consisted in mixing the samples at 300 rpm for 10 min with a blending (RW20 Ika) in order to homogenize them. Then, they were left at rest for 30 min in the measurement geometry in order for the sludge to be restructured. After this rest, viscoelastic properties (Fig. 1) were measured by applying oscillations at a frequency of 1 Hz for a strain range from 0.01% to 200%. Fifty measurement points were recorded according to a logarithmic distribution between those two limits. For each sample, a value of the elastic (G') and the viscous (G'') moduli in the linear viscoelastic region can be extracted (Table 1).
Finally, flow properties were obtained by applying a ramp of decreasing shear rates from 1000 s À 1 to 0.01 s À 1 (Fig. 2). Thirty measurement points, each for a time of 40 s, were used according to a   logarithmic distribution between the two limits. In order to determine the yield stress (τ 0 ) and the plastic viscosity (α 0 ) of each sample (Table 1), a modified Herschel-Bulkley model proposed by Baudez et al. [1] was used.

Spectral measurements
All the spectra measurements were realized simultaneously (but separately) with the rheological measurements. The samples had the same history: a mixing at 300 rpm for 10 min and a rest of 30 min. Data were acquired, exported and converted to Matlab readable files.  Acquisitions were taken with a pre-dispersive spectrometer double beam (JASCO V-670) equipped with an integrating sphere. Samples were analyzed in a quartz cell with optical path of 1 cm (Hellma). Spectral data (Fig. 3) were collected in the wavelength region of 1200-1800 nm at 5 nm intervals and a spectral bandwidth of 12 nm. The baseline was measured with a diffuse reflectance standard (Spectralon@). The manipulation of the experiments was undertaken at controlled room temperature (22 70.5°C).

PLS algorithm
A Partial Least Square (PLS) [2] algorithm was used to model the physical properties of the sludge. A general PLS model was built using the whole calibration set. The number of latent variables was determined by comparing performances by leave-one-out cross-validation [3]. Model results