On-line Sensor Research Articles

CUSTOMIZABLE IoT SOLUTION BASED ON ESP32 MCU

Sensing and perception are one of the key aspects in robotics. It is necessary for robots to be able to measure some physical parameters in the world and making sense of such data for performing different tasks and act according to surrounding environment conditions. Right now, robotics is seeing a revolution in use of sensors. Traditionally robots have been designed to have maximum stiffness and applications have been designed to be predictable in their operation. As robots emerge from the fences areas and are designed for a wider range of applications: from collaborative robotics to autonomously driving cars. It is essential to have perception capabilities that allow estimation of the state of the robot but also the state of the surrounding environment. Due to these new requirements the importance of sensing and perception has increased significantly over the last decade and will without doubt to continue to grow in the future [18]. The objective of this paper is to develop customizable IoT solution for remote monitoring of selected physical parameters. The created prototype is based on ESP32 MCU and consists of two parts: hardware and software & cloud. Also, mobile application for wireless device configuration and online sensors data monitoring was developed. Several filtering methods were implemented for proper data preprocessing of near-surface water level, outside temperature, solar power and self-diagnostics sensors’ data. Also, possibility of remote software upgrade is implemented. The prototype was setup with solar power supply and some experimental results are demonstrated in order to approve that this approach is fully functional, autonomous and low-cost. Keywords : IoT, digital filters, wireless device configuration, mqtt, cloud, sensing and perception.

On-Line Monitoring of Biofilm Accumulation on Graphite-Polypropylene Electrode Material Using a Heat Transfer Sensor.

Biofilms growing on electrodes are the heart piece of bioelectrochemical systems (BES). Moreover, the biofilm morphology is key for the efficient performance of BES and must be monitored and controlled for a stable operation. For the industrial use of BES (i.e., microbial fuel cells for energy production), monitoring of the biofilm accumulation directly on the electrodes during operation is desirable. In this study a commercially available on-line heat transfer biofilm sensor is applied to a graphite-polypropylene (C-PP) pipe and compared to its standard version where the sensor is applied to a stainless-steel pipe. The aim was to investigate the transferability of the sensor to a carbonaceous material (C-PP), that are preferably used as electrode materials for bioelectrochemical systems, thereby enabling biofilm monitoring directly on the electrode surface. The sensor signal was correlated to the gravimetrically determined biofilm thickness in order to identify the sensitivity of the sensor for the detection and quantification of biofilm on both materials. Results confirmed the transferability of the sensor to the C-PP material, despite the sensor sensitivity being decreased by a factor of approx. 5 compared to the default biofilm sensor applied to a stainless-steel pipe.

Low-temperature preparation of cupric spinel for room-temperature removal of hydrogen sulfide to value-added product ammonia formation

Spinel-type compounds are extremely effective in the adsorptive removal of gas pollutants. However, they are prepared at high temperatures, and their lack of catalytic production of value-added products deserves further research. This study focused on the preparation of cupric spinel compounds at low temperatures and their application to H2S removal with value-added product formation. Cupric spinel (CuFe2O4) was prepared by the co-precipitation of bivalent copper and trivalent iron at 60 °C, which allowed the formation of CuFe2O4 with a high surface area of 202.72 m2/g. The obtained breakthrough capacity was 5.68 mg H2S/g in ambient temperature removal of H2S. The analysis of outflowing gas by online sensor demonstrated the formation of 73.75 mg NH3/g spinel compound. The sulfide compound presented in the sample analysis supported the NH3 formation at the end of H2S removal, that is, NH3 was synthesized at ambient temperature after the formation of sulfide compound. Successfully prepared the CuFe2O4 at a low temperature and its application on H2S removal with value added product NH3 formation achieved at ambient temperature.

Highly sensitive fiber-optic SPR sensor with surface coated TiO2/MWCNT composite film for hydrogen sulfide gas detection

Hydrogen sulfide (H2S) gas has a severe effect on the respiratory system of the human body and ambient environment, necessitating development of on-line H2S gas sensors with high performance for safety and health concerns. Here, we proposed a fiber-optic surface plasmon resonance sensor for H2S detection employing TiO2 nanoparticles and multilayer carbon nanotubes composite (TiO2/MWCNT) as sensing film, featuring desirable advantages of high sensitivity, selectivity, and real-time detection. Benefiting from special structure and large specific surface area of MWCNTs, the adsorption capacity of sensing surface to gas molecules can be significantly enhanced. Moreover, the high carrier mobility of MWCNTs can further promote the charge transfer between TiO2 and H2S. These unique features of TiO2/MWCNT composite film result in an obvious improvement of sensitivity for H2S detection. Experimental results show that the maximum sensitivity of 21.76 pm ppm−1 (picometer/part-per-million) and detection limit of 0.2 ppm can be obtained by appropriately optimizing the componential constitutions of TiO2/MWCNT composite. Such detection limit is strikingly lower than the threshold concentrations in workplace set by Federal Institute for Occupational Safety (10 ppm). In addition, the favorable selectivity, response/recovery times, repeatability and stability were demonstrated as well. This facile and cost-effective work provides a novel strategy for constructing high performance H2S gas sensor with fast response and real-time detection, which has prospective application in the fields of human health and environmental conservation.

Quali-quantitative characterization of biogas with the temporal behavior of organic load on wastewater treatment plant with upflow anaerobic sludge blanket reactors through measurement in full‑scale systems

This study aims to present the time behavior of wastewater flow parameters, organic matter, biogas flow, biogas composition, and its relations, measured through online sensors, in a municipal wastewater treatment plant (WWTP) operating full-scale upflow anaerobic sludge blanket (UASB) reactors, installed in the south of Brazil. WWTP has online measurement devices to evaluate some physicochemical variables of the sewage and the biogas. The COD analyzer (UV– Vis probe), ultrasonic flow meter, biogas flow meter, and biogas composition analyzer were the equipment used. The monitoring occurred for two time periods each of 72 h and one time period for 48 h in the year 2018. Data were checked with descriptive statistics, data independence was checked through the autocorrelation Box– Ljung test, normality behavior was checked with several tests (Shapiro– Wilk, Kolmogorov–Smirnov, Lilliefors, Anderson–Darling, D’Agostino K2, and Chen–Shapiro), and Spearman’s correlation coefficient was used to evaluate the correlations among the parameters. The mean sewage flow was 345 ± 120 L.s-1; removed organic load was, in average, 48%; biogas quality values were 82.32% ± 3.62% v/v (CH4), 2.66% ± 1.19% v/v (CO2), and 3453 ± 1268 ppm (H2S); and the production per capita obtained was 4.51 ± 1.65 NL.hab-1.d-1. It was estimated an electric power generation of 3118.6 kWh.d-1, which is equivalent to an installed power of 130 KW. The behavior of removed organic load and biogas flow (Nm3.h-1), produced in the treatment plant, showed variable, periodic, and nonstationary time behavior.

Intelligent Sensors for Sustainable Food and Drink Manufacturing

Food and drink is the largest manufacturing sector worldwide and has significant environmental impact in terms of resource use, emissions, and waste. However, food and drink manufacturers are restricted in addressing these issues due to the tight profit margins they operate within. The advances of two industrial digital technologies, sensors and machine learning, present manufacturers with affordable methods to collect and analyse manufacturing data and enable enhanced, evidence-based decision making. These technologies will enable manufacturers to reduce their environmental impact by making processes more flexible and efficient in terms of how they manage their resources. In this article, a methodology is proposed that combines online sensors and machine learning to provide a unified framework for the development of intelligent sensors that work to improve food and drink manufacturers' resource efficiency problems. The methodology is then applied to four food and drink manufacturing case studies to demonstrate its capabilities for a diverse range of applications within the sector. The case studies included the monitoring of mixing, cleaning and fermentation processes in addition to predicting key quality parameter of crops. For all case studies, the methodology was successfully applied and predictive models with accuracies ranging from 95 to 100% were achieved. The case studies also highlight challenges and considerations which still remain when applying the methodology, including efficient data acquisition and labelling, feature engineering, and model selection. This paper concludes by discussing the future work necessary around the topics of new online sensors, infrastructure, data acquisition and trust to enable the widespread adoption of intelligent sensors within the food and drink sector.

Evanescent Field Absorption-Based Fiber Optic Sensor for Detecting Power Transformer Oil Degradation

ABSTRACT Observing the transformer oil quality is essential for the transformer’s health and system. The aging state of the transformer oils and the average remaining life accordingly can be determined using the breakdown voltage (BDV) test, ultraviolet-visible (UV-Vis) spectroscopy, and the refractive index (RI) standard methods. The subject of this study is to design and manufacture a much simpler and low-cost, evanescent field absorption-based fiber optic sensor (EFA-FOS) compatible with these proven standard methods and to perform a comparative performance analysis by determining a critical voltage threshold. Apart from the expensive and cumbersome standard methods mentioned above, the usability of the oil is carried out, thanks to the online sensor system, with the pretest using this critical voltage value. EFA-FOS measurement results of oil samples can determine the samples’ aging degree with high accuracy. The samples’ RI is determined with EFA-FOS with a sensitivity of −70 V/RIU and a linearity of 0.98 (R2). A strong relationship has also been determined between the sensor output and measured BDV with a regression constant of 0.86 (R2).

Domain Adaptation and Federated Learning for Ultrasonic Monitoring of Beer Fermentation

Beer fermentation processes are traditionally monitored through sampling and off-line wort density measurements. In-line and on-line sensors would provide real-time data on the fermentation progress whilst minimising human involvement, enabling identification of lagging fermentations or prediction of ethanol production end points. Ultrasonic sensors have previously been used for in-line and on-line fermentation monitoring and are increasingly being combined with machine learning models to interpret the sensor measurements. However, fermentation processes typically last many days and so impose a significant time investment to collect data from a sufficient number of batches for machine learning model training. This expenditure of effort must be multiplied if different fermentation processes must be monitored, such as varying formulations in craft breweries. In this work, three methodologies are evaluated to use previously collected ultrasonic sensor data from laboratory scale fermentations to improve machine learning model accuracy on an industrial scale fermentation process. These methodologies include training models on both domains simultaneously, training models in a federated learning strategy to preserve data privacy, and fine-tuning the best performing models on the industrial scale data. All methodologies provided increased prediction accuracy compared with training based solely on the industrial fermentation data. The federated learning methodology performed best, achieving higher accuracy for 14 out of 16 machine learning tasks compared with the base case model.

Fusing NIR and Process Sensors Data for Polymer Production Monitoring.

Process analytical technology and multivariate process monitoring are nowadays the most effective approaches to achieve real-time quality monitoring/control in production. However, their use is not yet a common practice, and industries benefit much less than they could from the outcome of the hundreds of sensors that constantly monitor production in industrial plants. The huge amount of sensor data collected are still mostly used to produce univariate control charts, monitoring one compartment at a time, and the product quality variables are generally used to monitor production, despite their low frequency (offline measurements at analytical laboratory), which is not suitable for real-time monitoring. On the contrary, it would be extremely advantageous to benefit from predictive models that, based on online sensors, will be able to return quality parameters in real time. As a matter of fact, the plant setup influences the product quality, and process sensors (flow meters, thermocouples, etc.) implicitly register process variability, correlation trends, drift, etc. When the available spectroscopic sensors, reflecting chemical composition and structure, consent to monitor the intermediate products, coupling process, and spectroscopic sensor and extracting/fusing information by multivariate analysis from this data would enhance the evaluation of the produced material features allowing production quality to be estimated at a very early stage. The present work, at a pilot plant scale, applied multivariate statistical process control (MSPC) charts, obtained by data fusion of process sensor data and near-infrared (NIR) probes, on a continuous styrene-acrylonitrile (SAN) production process. Furthermore, PLS regression was used for real-time prediction of the Melt Flow Index and percentage of bounded acrylonitrile (%AN). The results show that the MSPC model was able to detect deviations from normal operative conditions, indicating the variables responsible for the deviation, be they spectral or process. Moreover, predictive regression models obtained using the fused data showed better results than models computed using single datasets in terms of both errors of prediction and R 2. Thus, the fusion of spectra and process data improved the real-time monitoring, allowing an easier visualization of the process ongoing, a faster understanding of possible faults, and real-time assessment of the final product quality.

Comparison of H2S Gas Sensors: A Sensor Management Procedure for Sewer Monitoring

Hydrogen sulphide (H2S) emissions are one of the major problems associated with sewer networks. This gas, with its characteristic smell of rotten eggs is highly toxic and leads to the corrosion of sewer infrastructures. To protect cities and ensure the safety of sewer workers, sewers are commonly monitored using H2S gas sensors. In this work, three commercial H2S gas sensors for air quality monitoring were compared at two different sites in Berlin, Germany. Two of the sensors provide online access to data, while the other one is a data logger. Moreover, based on statistical measures (RMSE, MAE, MB, and a graphical analysis), we evaluated whether a rotation/exchange between data logger (reference) and online sensors is possible without significant differences in the gas measurements. Experimental evaluation revealed that measurement differences are dependent on the H2S concentration range. The deviation between sensors increases as the H2S concentration rises. Therefore, the interchange between reference and online sensors depends on the application site and the H2S levels. At lower ranges (0–10 ppm) there were no observed problems. Finally, to support practitioners on-site, a management procedure in the form of a decision-making tool is proposed for assessing whether gas sensors should be exchanged/rotated.

Highly sensitive on-line detection of trace Pb2+ based on tapered fiber integrated with black phosphorus

In this paper, a high-sensitivity Pb2+ on-line sensor based on black phosphorus (BP) integrated tapered fiber is proposed. The BP nanosheets which are deposited on the surface of the tapered fiber by light deposition technology are obtained by one-pot solvothermal reaction and ultrasonic peeling in the laboratory. Due to the electrostatic adsorption between BP and Pb2+, the effective refractive index (RI) of the tapered fiber surface is changed by the adsorption complexes, which eventually causes the interference wavelength to shift. The experimental results show that the sensor has a sensitivity of 0.03714 nm/ppb when the range of Pb2+ solution concentration is 0.1–105 ppb, and its limit of detection (LOD) is 0.0206 ppb. The response time of 1 ppb and 106 ppb Pb2+ solution is 44 s and 49 s, respectively. Meanwhile, the sensor has good stability after repeated measurement of the same concentration solution. In general, the sensor has the characteristics of simple preparation, high sensitivity, low detection limit, integration, and low-cost, and has potential application prospects to detecting trace Pb2+ in water quality monitoring.

On the Use of Copula for Quality Control Based on an AR(1) Model

Manufacturing for a multitude of continuous processing applications in the era of automation and ‘Industry 4.0’ is focused on rapid throughput while producing products of acceptable quality that meet customer specifications. Monitoring the stability or statistical control of key process parameters using data acquired from online sensors is fundamental to successful automation in manufacturing applications. This study addresses the significant problem of positive autocorrelation in data collected from online sensors, which may impair assessment of statistical control. Sensor data collected at short time intervals typically have significant autocorrelation, and traditional statistical process control (SPC) techniques cannot be deployed. There is a plethora of literature on techniques for SPC in the presence of positive autocorrelation. This paper contributes to this area of study by investigating the performance of ‘Copula’ based control charts by assessing the average run length (ARL) when the subsequent observations are correlated and follow the AR(1) model. The conditional distribution of yt given yt−1 is used in deriving the control chart limits for three different categories of Copulas: Gaussian, Clayton, and Farlie-Gumbel-Morgenstern Copulas. Preliminary results suggest that the overall performance of the Clayton Copula and Farlie-Gumbel-Morgenstern Copula is better compared to other Archimedean Copulas. The Clayton Copula is the more robust with respect to changes in the process standard deviation as the correlation coefficient increases.

Qualitative and quantitative recognition method of drug-producing chemicals based on SnO2 gas sensor with dynamic measurement and PCA weak separation

With the rampant drug crime, the detection of drug-producing chemicals has put forward the great demand of multi-type and multi-concentration on-line rapid detection. The rapid development of dynamic measurement for semiconductor gas sensors provides a solution to this problem. However, the mutual blend of type and concentration information negatively affects sensor performance. In this paper, principal component analysis (PCA) was used for weak separation of type and concentration; k-Nearest Neighbor (KNN) was used for qualitative recognition; polynomial regression was used for quantitative recognition. The physical meaning of the dynamic response signal after PCA transformation was first proposed: PC1 has a weak concentration meaning; the combination of PC2, PC3, and PC4 has a weak type meaning. Based on the weak separation, the stepwise recognition method of qualitative classification and quantitative regression was first used to improve the recognition rate, the resolution and the generalization performance of the sensor. Using the inverse transformation of PCA, the principle of PCA and the method of ideal data verified the feasibility of this method. The qualitative and quantitative recognition of various drug-producing chemicals had been realized, which is a new way of on-line rapid sensor detection for drug-producing chemicals.

Lubricating oil deterioration modeling and remaining useful life prediction based on hidden semi-Markov modeling

Lubricating oil, which carries information about machine’s health condition, is of great importance to the performance of machines in the full life cycle. The main purpose of oil deterioration modeling and its remaining useful life prediction is to determine the exact time that the lubricating oil has degraded and it is no longer able to maintain its functions. Generally, lubricating oil deterioration can be partially detected by condition monitoring based on wear debris analysis, and thus can be categorized into three states. In our paper, vector data, which contain wear debris concentration and carry information about the state of lubricating oil, are obtained by an on-line visual ferrograph sensor from a four-ball tester at regular sampling epochs. The oil’s state process is described by a hidden semi-Markov model, and its sojourn times in each state are assumed to be Erlang distributed. A vector autoregressive method based on time series modeling is presented to obtain residual observations, which are regarded as the observable process of oil information in the hidden semi-Markov model framework. The unknown parameters of the hidden semi-Markov model are then estimated by using expectation-maximization algorithm. Afterward, a Bayesian updating approach is presented to derive the explicit formulas of the conditional reliability and mean residual life. To validate the proposed approach, a real case study of lubricating oil deterioration is demonstrated and a comparison with the hidden Markov model is given to illustrate the effectiveness of the new developed remaining useful life prediction approach for lubricating oil.

A SENSOR NETWORK BASED ON IOT AND ISTSOS TO ANALYSE THE CATCH BASIN ENVIRONMENT IN THE CONTEXT OF THE AE. ALBOPICTUS EXPANSION IN SWITZERLAND

Abstract. The continuous expansion of invasive Asian tiger mosquito, Aedes albopictus, combined to its ability to transmit arboviruses (e.g. dengue, chikungunya) is raising major public health concern in Europe. In Switzerland, the mosquito is firmly established in most urban areas of the Canton of Ticino, south of the Alps, and there is a real risk that it will colonize also urban areas north of the Alps in the next years. The spatial distribution and colonization of new areas by Ae. albopictus depends on several environmental parameters, such as winter and summer temperatures, and precipitation patterns. A key factor for Ae. albopictus to establish at higher latitudes is the capability to develop cold-tolerant overwintering diapausing eggs under specific environmental conditions. Weather-driven abundance models are used to map the areas of potential distribution and to predict temporal dynamics of Ae. albopictus and the transmission potential of arboviruses. This contribution presents the designed system that integrates low-cost and on-line IoT sensors to monitor temperature, humidity and light with istSOS an OGC Sensor Observation Service server implementation with a user friendly interface and rich feature collection to easily manage this sensor network and distribute data in a standard way (www.istsos.org).

Development of a soil fertility index using on-line Vis-NIR spectroscopy

Soil fertility index (SFI) is commonly used for soil fertility assessment, which is critical for managing in-field variabilities and maximizing crop production with minimum environmental impacts. However, the majority of earlier SFIs were laboratory-based soil analyses. This study developed a novel SFI using on-line collected visible and near-infrared (vis-NIR) spectra. Six agricultural fields were scanned using an on-line vis-NIR sensor (CompactSpec, Tec5 Technology, Germany), when 139 soil samples were collected and analyzed for soil pH, organic carbon, available- phosphorous (P), potassium, magnesium (Mg), calcium, sodium, moisture content (MC) and cation exchange capacity. A minimum dataset was developed comprising the fertility attributes that showed pairwise correlation (r) smaller than 0.75. This was followed by a principal component analysis to calculate the weight factor of each parameter to be used in the SFI formulation using a double-weighted additive function. The data matrix consisting of the SFI and soil spectra was divided into calibration (70 %) and prediction (30 %) datasets. The former set was subjected to a partial least squares regression to calibrate SFI model, whose accuracy was validated using the prediction set. Results showed that the derived SFI was moderately to highly correlated with P (r = 0.57), pH (r = 0.75), and Mg (r = 0.74) and weakly correlated with MC (r = 0.26). The on-line vis-NIR sensor predicted SFI with very good accuracy [coefficient of determination (R2) = 0.75 and ratio of prediction to deviation (RPD) = 2.01]. Therefore, it is concluded that the vis-NIR can accurately predict SFI directly from on-line scanned soil spectra, which can effectively assess soil fertility and manage in-field variability.

Nanoparticle Determination in Water by LED-Excited Surface Plasmon Resonance Imaging

The increasing popularity of nanoparticles in many applications has led to the fact that these persistent materials pollute our environment and threaten our health. An online sensor system for monitoring the presence of nanoparticles in fresh water would be highly desired. We propose a label-free sensor based on SPR imaging. The sensitivity was enhanced by a factor of about 100 by improving the detector by using a high-resolution camera. This revealed that the light source also needed to be improved by using LED excitation instead of a laser light source. As a receptor, different self-assembled monolayers have been screened. It can be seen that the nanoparticle receptor interaction is of a complex nature. The best system when taking sensitivity as well as reversibility into account is given by a dodecanethiol monolayer on the gold sensor surface. Lanthanide-doped nanoparticles, 29 nm in diameter and with a similar refractive index to the most common silica nanoparticles were detected in water down to 1.5 µg mL−1. The sensor can be fully regenerated within one hour without the need for any washing buffer. This sensing concept is expected to be easily adapted for the detection of nanoparticles of different size, shape, and composition, and upon miniaturization, suitable for long-term applications to monitor the quality of water.

Soft Sensing for On-Line Fault Detection of Ammonium Sensors in Water Resource Recovery Facilities.

The increasing demand for online sensors applied to advanced control strategies in water resource recovery facilities has resulted in the increasing investigation of fault-detection methods to improve the reliability of sensors installed in harsh environments. The study herein focuses on the fault detection of ammonium sensors, especially for effluent monitoring, given their potential in ammonium-based aeration control applications. An artificial neural network model was built to predict the ammonium content in the effluent by employing the information from five other sensors installed in the activated sludge tank: NH4+, pH, ORP, DO, and TSS. The residual between the model prediction and the effluent ammonium sensor signal was utilized in a fault-detection mechanism based on principal component analysis and Shewhart monitoring charts. In contrast to previous studies, the present work utilizes typical faults collected from a 1 year historic dataset of an actual sensor setup. Treatment process anomalies, calibration bias faults, and fouling drifts were the most common issues identified from the historic dataset, and they were promptly identified by the proposed fault-detection methodology. Once a fault is detected, the model prediction can be actively used in place of the sensor for process control without affecting the treatment process by utilizing faulty datasets.

Comparison of Online Sensors for Liquid Phase Hydrogen Sulphide Monitoring in Sewer Systems

Hydrogen sulfide (H2S) related to wastewater in sewer systems is known for causing significant problems of corrosion and odor nuisance. Sewer systems severely affected by H2S typically rely on online H2S gas sensors for monitoring and control. However, these H2S gas sensors only provide information about the H2S emission potential at the point being monitored, which is sometimes inadequate to design control measures. In this study, a comparison of three market-ready online sensors capable of liquid-phase H2S detection in sewer systems was assessed and compared. Two of the three sensors are based on UV/Vis spectrophotometry, while the other adapted the design and principles of a Clark-type electrochemical microsensor. The H2S measurements of the sensors were statistically compared to a standard laboratory method at first. Following that, the performance of the online sensors was evaluated under realistic sewer conditions using the Berlin Water Company (BWB) research sewer pilot plant. Test applications representing scenarios of typical H2S concentrations found in sulfide-affected sewers and during control measures were simulated. The UV/Vis spectrometers showed that the performance of the sensors was highly dependent on the calibration type and measurements used for deriving the calibration function. The electrochemical sensor showed high sensitivity by responding to alternating anaerobic/anoxic conditions simulated during nitrate dosing. All sensors were prone to measurement disturbances due to high amounts of sanitary solids in wastewater at the study site and required continuous maintenance for reliable measurements. Finally, a summary of the key attributes and limitations of the sensors compared for liquid phase H2S detection is outlined.

Technology transfer of a monitoring system to predict product concentration and purity of biopharmaceuticals in real-time during chromatographic separation.

Technological developments require the transfer to their location of application to make use of them. We describe the transfer of a real‐time monitoring system for lab‐scale preparative chromatography to two new sites where it will be used and developed further. Equivalent equipment was used. The capture of a biopharmaceutical model protein, human fibroblast growth factor 2 (FGF‐2) was used to evaluate the system transfer. Predictive models for five quality attributes based on partial least squares regression were transferred. Six out of seven online sensors (UV/VIS, pH, conductivity, IR, RI, and MALS) showed comparable signals between the sites while one sensor (fluorescence) showed different signal profiles. A direct transfer of the models for real‐time monitoring was not possible, mainly due to differences in sensor signals. Adaptation of the models was necessary. Then, among five prediction models, the prediction errors of the test run at the new sites were on average twice as high as at the training site (model‐wise 0.9–5.7 times). Additionally, new prediction models for different products were trained at each new site. These allowed monitoring the critical quality attributes of two new biopharmaceutical products during their purification processes with mean relative deviations between 1% and 33%.