Cerro Matoso SA (CMSA) is located in Montelibano, Colombia. It is one of the biggest producers of ferronickel in the world. The structural health monitoring process performed in the electric arc furnaces at CMSA, is of great importance in the maintenance and control of ferronickel production. The control of thermal and dimensional conditions of the electric furnace aims to detect and prevent failures that may affect its physical integrity. A network of thermocouples distributed radially and at different heights from the furnace wall, is responsible for monitoring the temperatures in the electric furnace lining. In order to optimize the operation of the electric furnace, it is important to predict the temperature at some points. However, this can be difficult due the number of variables which it depends on. To predict the temperature behavior in the electric furnace lining, a deep learning model for time series prediction has been developed. ARIMA, LSTM, GRU and other combinations were tested. Gated Recurrent Unit (GRU) characterized by its multivariate and multi output type had the lowest square error. A study of the best input variables for the model that influence the temperature behavior is also carried out. Some of the input variables are the power, current, impedance, calcine chemistry, temperature history, among others. The methodology to tune the parameters of the GRU deep learning model is described. Results show an excellent behavior for predicting the temperatures 6 hours into the future with root mean square errors of 3%. This model will be integrated to a software that obtains data for a time window from the distributed control system (DCS) to feed the model. In addition, this software will have a graphical user interface used by the operators furnace in the control room. Results of this work will improve the process of structural control and health monitoring at Cerro Matoso S.A.

The development of portable ECG has found growing markets, from wearable ECG sensor to ambulatory ECG recorder, encountering challenges of moderately complex to tightly regulated device. This study investigates how the different modalities of ECG are affected by motion artifact and their reduction techniques, when using analog front-end (AFE) integrated circuits such as the AD823X family. It is known that the typical range of current mobile health ECG devices ADC is 10-12 bits, and sometimes 16-bits, which might be insufficient to show up the small potential amplitudes useful in diagnosis; but the interest now is on the interplay of how a digital resolution choice and variable gain can cope with motion artifacts. With our methodology for a rapid prototyping of ECG device, and using the AFE AD8232 and Bluetooth communication, multiple ECG configurations are evaluated under two microcontroller systems of different resolution: a generic Arduino Nano board which features a 10-bit ADC and the 24-bit ADC converter of Silicon Labs C8051F350 board. From the ECG configuration applications (such as: cardiac monitor, fitness, HMR in chest, diagnostic, and Holter), analog filters characteristics and gain are explored under these two different microcontroller systems, in their capacity to solve the problems of motion artifacts alongside performance of the special parameters of each application.

The main component of the ion selective surfactant sensor is the sensing membrane. Ion selective sensing membranes are usualy fabricated as a liquid type membrane with matrix made of a PVC, ionophore and a plasticizer. Plasticizers soften the PVC but due to their liphopilicity they influence the ionophore, the ion exchange across the membrane, membrane resistance and consequently the analytical signal. The aim of the research was to investigate the influence of four different plasticizers on the analytical properties of the surfactant sensor membrane towards cationic surfactants. 1,3-didecyl-2-methyl-imidazolium-tetraphenylborate (DMI-TPB) ion exchange complex was used as an ionophore. Two cationic surfactants; cetylpyridinium chloride (CPC) and hexadecyltrimethylammonium (CTAB), were used to measure the membrane response characteristics for all plasticizer formulations. Next, the same membranes were used to perform the titrations of the CPC and CTAB with dodecyl sulfate (DS) as a titrant. The plasticizer with the best analytical properties was selected for testing on real samples of six pharmaceutical disinfectants. The results were compared with standard surfactant sensor.

Commercially-available activity trackers have recently exploded in popularity. Use of an activity tracker has been shown to increase physical activity for some individuals, but abandonment of these devices is high. Many of these activity trackers have been found to be accurate at tracking exercise, though concerns remain about how well these devices track non-traditional movement. This study compares several models of activity trackers for accuracy in step counts while riding a motorcycle a moderate distance.

A motorcyclist wore an Apple Watch (Generation 1) (AW), Fitbit Charge HR (FCH), and a Fitbit Zip (FZ) while completing ten, 14-mile out-and-back trip trials. Step counts of each device were recorded after each trial, and the difference between the number of steps measured by the devices and the actual step counts (zero in all cases) were calculated. All trials contained miscounted steps by all the devices. The average miscounted steps for the AW were 12.9 steps (P<0.05), the FCH miscounted 211.0 steps (P<0.05), and the FZ miscounted 305.3 steps (P<0.05). The range of miscounted steps varied from 2 miscounted steps for the AW to 811 miscounted steps for the FZ. There were also noted differences in step counts between the “out” and “back” portions of the trips. In this study, commercially-available activity trackers were shown to misrepresent riding a motorcycle as step activity. Reported step counts, and therefore distance and calorie information, from these activity trackers should be interpreted with caution among motorcyclists.

Rectangular glass micro-capillaries are very interesting devices that can be inserted in a micro-fluidic path and exploited for label-free optical sensing of ultra-low volumes of fluids. In the past, such devices have been used for the detection of the refractive index of fluids. In this work, we developed a smart micro-opto-fluidic platform that can distinguish water and alcohol samples flowing in the micro-channel thanks to the profile of their absorption spectrum in the near infrared (NIR) region from 1.15 to 1.65 µm. The readout technique is non-contact, remote and non-invasive. The micro-capillary, with wall thickness of 280 µm and channel depth equal to 400 µm, is laid flat onto an Aluminum bulk mirror and the light from a Tungsten lamp is shone on its upper flat side with an angle of incidence of 14°. The readout beam crosses the glass walls and the channel depth twice, since it is reflected by the mirror, and it is then coupled to the monochromator input of an optical spectrum analyzer. The theoretical transmission spectra T(λ) of the capillary filled just with air as well as with distilled water, isopropanol, ethylene glycol, and 95% ethanol (with 5% water content) are obtained using analytical equations including the wavelength-dependent attenuation due to fluid absorption. Then experimental measurements are carried out and the experimental spectral response, defined as SR(λ) = T_sample(λ)/T_air(λ), is compared with the theoretical one, revealing a very good level of agreement.

The present work proposes an electronic tongue arrangement for the detection of biogenic amines (BAs) in ternary complex mixtures.

Since the formation of BAs is directly proportional to the increase of temperature and the presence of bacteria, the elevated concentration of that kind of compounds could be related easily with the quality of the food industry products. The most regulated field is the fish industry, which has set limits for histamine as the marker compound: 100 mg·Kg^-1 in the European Union and 50 mg·Kg^-1 in the United States of America.

Herein it is proposed a voltammetric sensor array for the quick and efficient detection of histamine (Hys), cadaverine (Cad) and tyramine (Tyr) which, together with advanced chemometric tools such as artificial neural networks (ANN) and partial least squares (PCA), leads to models able to predict the individual concentration of each BA in the analyzed samples.

The final ANN structure had 51 input neurons, 5 neurons in the hidden layer, and 3 neurons in the output layer. The functions used for the hidden and output layers were Tansig and Purelin, respectively. The results show that this is a valid model with slopes near to 1 and intercepts close to 0. Moreover, it is important to remark that the worst correlation has a value of almost 0.900.

Biosensors and wearable sensor systems with transmitting capabilities are currently developed and used for the monitoring of health data, exercise activities, and other performance data. Unlike conventional approaches, these devices enable convenient, continuous, and/or unobtrusive monitoring of a user’s behavioral signals in real time. Examples include signals relative to hand movements and individual grip force data, which directly translate into spatiotemporal grip force profiles for the different measurement loci on the fingers and/or palm of the hand. Wearable sensor systems combine innovation in sensor design, electronics, data transmission, power management, and signal processing for statistical analysis, as will be shown in this presentation. The first part briefly summarizes current state of the art in grip force profiling to highlight important functional aspects. Then, wearable sensor technology in the form of sensor glove systems for the real-time monitoring of task skill evolution during training in a simulator task will be described on the basis of the spatiotemporal evolution of individual grip force profiles and their statistical and functional analysis. Although a lot of research is currently devoted to this area, many technological aspects still remain to be optimized, and new methods for data analysis and knowledge representation are urgently needed, as will be clarified in the discussion. Wearable sensor technology represents an open challenge for the scientific community and its further development partly relies on contributions from women researchers from multiple disciplines, as pointed out in the final conclusions of this presentation.

In this paper, a dual band planar inverted F antenna (PIFA) is designed for wireless communication intended to be used in wireless body sensor networks. The designed PIFA operates at two different frequency bands, 2.45 GHz (ISM band) and 5.2 GHz (HiperLAN band). In body-centric wireless networks antennas need to be integrated with wireless wearable sensors. An antenna is an essential part of wearable body sensor networks. For on-body communications antennas need to be less sensitive from human body effects. For body-centric communications wearable devices need to communicate with the devices located over the surface and also there is a need of communication from on-body devices to off-body units. Based on this need a dual band planar inverted F antenna is designed which works at two different frequency bands, i.e., 2.45 GHz and 5.2 GHz. The 2.45 GHz is proposed for establishing communication among the wireless sensor devices attached on the human body while 5.2 GHz is proposed for the communications form on-body to off-body devices. The proposed antenna is very compact and due to having ground plane at the backside it shows less sensitive to the effects of the human body tissues. Computer Simulation Technology (CST) microwave studio™ was used for antenna design and simulation purposes. Performance parameters such as return loss, bandwidth, radiation pattern and efficiency of this antenna are shown and investigated. These performance parameters of the proposed antenna have been investigated at free space and close proximity to the human body. Simulation results and analysis show that the performance parameters of the proposed very good results at both frequency bands. Due to its compact size, less sensitive to the human body tissues and dual band functionality, it will be a good candidate for wireless wearable body sensor networks.

With the growth of population, the demand of milk is increasing at very fast rate. Due to this increasing demand, adulteration of milk by various substances has been very common in throughout the world, which not only reduces the nutritional value but also causes various diseases to human being. The common adulterants used in adulteration are detergents, ammonium sulphate [(NH₄)SO₄], sodium hydroxide (NaOH), sodium-bi-carbonate (NaHCO₃) ,common salt (NaCl) and fat.

The aim of this paper is to identify / detect the minimum detectable limit for the above six adulterants in the milk. A capacitor sensor is used, which is sensitive to an electrical property ( Relative Permittivity ) of the measuring medium and gives different dielectric loss angle of the milk sample under measurement, when sensor is immersed in different adulterated/unadulterated milk. The tangent of dielectric loss angle (tan Delta) is measured by Schering Bridge. By this sensor measurement system, detection of adulterants in milk with different % (from 5% to 20%) is studied. Packet milk and raw-milk (directly collected from the milk collection center) are used for the experiment. Experimental results of various adulteration levels are plotted for verification of the result and to check the data consistency.

Current progress in sensor technology is setting the stage to move closer to satisfactory solutions to challenging engineering problems, like e.g. system identification and structural health monitoring (SHM). In civil engineering, SHM is often based on the analysis of vibrational recordings, represented by time histories of displacements and/or accelerations collected through pervasive sensor networks and shaped as Multivariate Time Series (MTS). Despite the great advances in soft computing techniques like neural networks, inverse problems featuring regression tasks on the raw vibrational measurements are still challenging. Developing dimensionality reduction tools, able to infer complex correlations within and across the recorded time series, stands as a must. In this work, we have designed an AutoEncoder (AE) capable of condensing MTS-shaped data in a vector featuring a few latent variables only. The obtained reduced data representation allows the solution of inverse problems, like e.g. the identification of the parameters governing the dynamic load applied to a structural system. Inception modules and residual learning are respectively exploited for the encoding and the decoding parts of the AE, enhancing the informative content of the latent variables. Numerical examples, aimed at the identification of the loading conditions on a shear-type building, are reported to assess the effectiveness of the proposed procedure.