Anomaly Detection and Inter-Sensor Transfer Learning on Smart Manufacturing Datasets.

Smart manufacturing systems are considered the next generation of manufacturing applications. One important goal of the smart manufacturing system is to rapidly detect and anticipate failures to reduce maintenance cost and minimize machine downtime. This often boils down to detecting anomalies within the sensor data acquired from the system which has different characteristics with respect to the operating point of the environment or machines, such as, the RPM of the motor. In this paper, we analyze four datasets from sensors deployed in manufacturing testbeds. We detect the level of defect for each sensor data leveraging deep learning techniques. We also evaluate the performance of several traditional and ML-based forecasting models for predicting the time series of sensor data. We show that careful selection of training data by aggregating multiple predictive RPM values is beneficial. Then, considering the sparse data from one kind of sensor, we perform transfer learning from a high data rate sensor to perform defect type classification. We release our manufacturing database corpus (4 datasets) and codes for anomaly detection and defect type classification for the community to build on it. Taken together, we show that predictive failure classification can be achieved, paving the way for predictive maintenance.

Temperature Self-Calibration of Always-On, Field-Deployed Ion-Selective Electrodes Based on Differential Voltage Measurement.

Originally developed for use in controlled laboratory settings, potentiometric ion-selective electrode (ISE) sensors have recently been deployed for continuous, in situ measurement of analyte concentration in agricultural (e.g., nitrate), environmental (e.g., ocean acidification), industrial (e.g., wastewater), and health-care sectors (e.g., sweat sensors). However, due to uncontrolled temperature and lack of frequent calibration in these field applications, it has been difficult to achieve accuracy comparable to the laboratory setting. In this paper, we propose a novel temperature self-calibration method where the ISE sensors can serve as their own thermometer and therefore precisely measure the analyte concentration in the field condition by compensating for the temperature variations. We validate the method with controlled experiments using pH and nitrate ISEs, which use the Nernst principle for electrochemical sensing. We show that, using temperature self-calibration, pH and nitrate can be measured within 0.3% and 5% of the true concentration, respectively, under varying concentrations and temperature conditions. Moreover, we perform a field study to continuously monitor the nitrate concentration of an agricultural field over a period of 6 days. Our temperature self-calibration approach determines the nitrate concentration within 4% of the ground truth measured by laboratory-based high-precision nitrate sensors. Our approach is general and would allow battery-free temperature-corrected analyte measurement for all Nernst principle-based sensors being deployed as wearable or implantable sensors.

A non-invasive multipoint product temperature measurement for pharmaceutical lyophilization.

Monitoring product temperature during lyophilization is critical, especially during the process development stage, as the final product may be jeopardized if its process temperature exceeds a threshold value. Also, in-situ temperature monitoring of the product gives the capability of creating an optimized closed-loop lyophilization process. While conventional thermocouples can track product temperature, they are invasive, limited to a single-point measurement, and can significantly alter the freezing and drying behavior of the product in the monitored vial. This work has developed a new methodology that combines non-invasive temperature monitoring and comprehensive modeling. It allows the accurate reconstruction of the complete temperature profile of the product inside the vial during the lyophilization process. The proposed methodology is experimentally validated by combining the sensors' wirelessly collected data with the advanced multiphysics simulations. The flexible wireless multi-point temperature sensing probe is produced using micro-manufacturing techniques and attached outside the vial, allowing for accurate extraction of the product temperature.

A biodegradable chipless sensor for wireless subsoil health monitoring.

Precision Agriculture (PA) is an integral component of the contemporary agricultural revolution that focuses on enhancing food productivity in proportion to the increasing global population while minimizing resource waste. While the recent advancements in PA, such as the integration of IoT (Internet of Things) sensors, have significantly improved the surveillance of field conditions to achieve high yields, the presence of batteries and electronic chips makes them expensive and non-biodegradable. To address these limitations, for the first time, we have developed a fully Degradable Intelligent Radio Transmitting Sensor (DIRTS) that allows remote sensing of subsoil volumetric water using drone-assisted wireless monitoring. The device consists of a simple miniaturized resonating antenna encapsulated in a biodegradable polymer material such that the resonant frequency of the device is dependent on the dielectric properties of the soil surrounding the encapsulated structure. The simple structure of DIRTS enables scalable additive manufacturing processes using cost-effective, biodegradable materials to fabricate them in a miniaturized size, thereby facilitating their automated distribution in the soil. As a proof-of-concept, we present the use of DIRTS in lab and field conditions where the sensors demonstrate the capability to detect volumetric water content within the range of 3.7-23.5% with a minimum sensitivity of 9.07 MHz/%. Remote sensing of DIRTS can be achieved from an elevation of 40 cm using drones to provide comparable performance to lab measurements. A systematic biodegradation study reveals that DIRTS can provide stable readings within the expected duration of 1 year with less than 4% change in sensitivity before signs of degradation. DIRTS provides a new steppingstone toward advancing precision agriculture while minimizing the environmental footprint.

Antidelaminating, Thermally Stable, and Cost-Effective Flexible Kapton Platforms for Nitrate Sensors, Mercury Aptasensors, Protein Sensors, and p-Type Organic Thin-Film Transistors.

The inkjet printing of metal electrodes on polymer films is a desirable manufacturing process due to its simplicity but is limited by the lack of thermal stability and serious delaminating flaws in various aqueous and organic solutions. Kapton, adopted worldwide due to its superior thermal durability, allows the high-temperature sintering of nanoparticle-based metal inks. By carefully selecting inks (Ag and Au) and Kapton substrates (Kapton HN films with a thickness of 135 µm and a thermal resistance of up to 400 °C) with optimal printing parameters and simplified post-treatments (sintering), outstanding film integrity, thermal stability, and antidelaminating features were obtained in both aqueous and organic solutions without any pretreatment strategy (surface modification). These films were applied in four novel devices: a solid-state ion-selective (IS) nitrate (NO3-) sensor, a single-stranded DNA (ssDNA)-based mercury (Hg2+) aptasensor, a low-cost protein printed circuit board (PCB) sensor, and a long-lasting organic thin-film transistor (OTFT). The IS NO3- sensor displayed a linear sensitivity range between 10-4.5 and 10-1 M (r2 = 0.9912), with a limit of detection of 2 ppm for NO3-. The Hg2+ sensor exhibited a linear correlation (r2 = 0.8806) between the change in the transfer resistance (RCT) and the increasing concentration of Hg2+. The protein PCB sensor provided a label-free method for protein detection. Finally, the OTFT demonstrated stable performance, with mobility values in the linear (µlin) and saturation (µsat) regimes of 0.0083 ± 0.0026 and 0.0237 ± 0.0079 cm2 V-1 S-1, respectively, and a threshold voltage (Vth) of -6.75 ± 3.89 V.

Multi-Point Wireless Temperature Sensing System for Monitoring Pharmaceutical Lyophilization.

This work presents the design and evaluation of a fully wireless, multi-point temperature sensor system as a Process Analytical Technology (PAT) for lyophilization. Each sensor contains seven sensing elements which measure the product temperature at various positions of the contents of a glass vial. The sensor performance was studied by freeze drying experiments with sensor placement in both center and edge of full shelf of 6R glass vials with 4 ml fill volume. Product temperature profile and primary drying time measured at the bottom center position in the glass vial by the wireless sensor as well as the primary drying time are in close comparison with the thermocouple data. The drying times during primary drying were determined at the top, higher middle, lower middle and bottom positions which are 3.26 mm apart vertically in the vial by the wireless sensor based on the temperature profile measured at different positions. For a center vial, the drying time from the start of primary drying to each layer was measured at 3.9, 9.3, 14.2, and 21 h respectively, allowing to track the sublimation interface during primary drying phase. In addition, sublimation rate at each layer was calculated based on the drying time and theoretical weight loss of ice in the product. The sublimation rate at the beginning of the primary drying was similar to the sublimation rate by gravimetric method. Furthermore, the vial heat transfer coefficient (Kv ) was also calculated based on the sublimation rate. Thus, allowing the use of the multi-point wireless sensor to rapidly monitor the sublimation rate and Kv for every batch as continuous process verification. Similar tests were also conducted with 3% w/v mannitol solutions and the results were consistent demonstrating potential for real-time monitoring, process verification and cycle optimization for pharmaceutical lyophilization.

A nonlithographic top-down electrochemical approach for creating hierarchical (micro-nano) superhydrophobic silicon surfaces.

Superhydrophobic surfaces are biomimetic structures with potential applications in several key technological areas. In the past decade, several top-down and bottom-up fabrication methods have been developed to create such surfaces. These typically combine a hierarchical structure and low surface energy coatings to increase the contact angle and decrease the rolling angles. Silicon-based superhydrophobic surfaces are particularly attractive since they can be integrated with active electronics in order to protect them from the detrimental effects of environmental water and moisture. In this work, we introduce a simple and inexpensive process incorporating electrochemical surface modification (to create a fractal shape micro-nano topography) in combination with a final wet etching step to fabricate a superhydrophobic silicon surface with a contact angle of 160 degrees and a sliding angle of less than 1 degree.