Postural sway is a critical measure for evaluating postural control, and its analysis plays a vital role in preventing falls among the elderly. Typically, physiotherapists assess an individual's postural control using tests such as the Berg Balance Scale, Tinetti Test, and time up-and-go test. Sensor-based analysis is available based on devices such as force plates or inertial measurement units. Recently, machine learning methods have demonstrated promising results in the sensor-based analysis of postural control. However, these models are often complex, slow, and energy-intensive. To address these limitations, this study explores the design space of lightweight machine learning models deployable to microcontrollers to assess postural stability. We developed an artificial neural network (ANN) model and compare its performance to that of random forests, gaussian naive bayes, and extra tree classifiers. The models are trained using a sway dataset with varying input sizes and signal-to-noise ratios. The dataset comprises two feature vectors extracted from raw accelerometer data. The developed models are deployed to an ARM Cortex M4-based microcontroller, and their performance is evaluated and compared. We show that the ANN model has 99.03% accuracy, higher noise immunity, and the model performs better with a window size of one second with 590.96 us inference time.
The present invention relates to a torque sensor (1) comprising a body (2), at least one airtight chamber (3) provided in the body, a pressure sensor (4) measuring 1 the pressure in said at least one airtight chamber, and a pressure to torque converter (5) connected with the pressure sensor. Each airtight chamber is arranged to change its volume when the body is subjected to a torque, wherein the volume change causes a change of pressure of the enclosed air in the airtight chamber. The change of pressure is detected and converted to the corresponding torque.
A new control strategy for steam pressure control in the drying cylinders during web breaks is presented. The development is performed using a new physical model implemented in Matlab Simulink. The goal of the control strategy is to obtain the same drying properties after the web break as before the break as fast as possible.
Web breaks in the dryer section of a paper machine cause loss of production and quality problems. After a web break, the steam pressure in the cylinders must be reduced to avoid overheating. The goal of this project is to determine optimal steam pressure trajectories during web breaks, so that the production is restarted with the same drying properties of the cylinder as before the break. A detailed physical dynamic model of the drying cylinder has been developed. The model describes the relations between the steam valve position, the steam pressure, the cylinder temperature, and the paper temperature. The model is based on partial differential equations that describe heat conductivity for the cylinder and the paper web, and mass balances of water and dry material in the paper. The accuracy of the model has been verified through experiments made at the M-real paper mill in Husum, Sweden. Verifications are made both during normal operation and during web breaks. The dynamic model has been reduced in order to derive simple transfer functions between the steam pressure and the cylinder temperature, and between a logic signal that is active during web breaks and the cylinder temperature, respectively. The transfer functions obtained were used to find the optimal steam pressure trajectory during web breaks. A new feed-forward strategy for steam pressure control during web breaks is presented. The strategy has been tested on a paper machine with good results. The strategy is built on feed-forward compensation and has been well received at the mill.
A new control strategy for steam pressure control in the drying cylinders during web breaks is presented. The development is performed using a new physical model implemented in Simulink. The goal of the control strategy is that the cylinder temperatures are retained when the production is restarted after the web break.
This paper presents a new strategy for steam pressure control during web breaks in the paper machine. The aim was to restart paper production with the same drying properties of the cylinder as before the break. A detailed physical dynamic model of the drying cylinder has been developed. The accuracy of the model has been verified through experiments made at the M-real paper mill in Husum, Sweden. Verifications are made both during normal operation and during web breaks. The dynamic model has been reduced in order to derive simple transfer functions which were used to find the optimal steam pressure trajectory during web breaks. The resulting strategy has been tested on a paper machine with good results and it has been well received at the mill.
Rapporten beskriver driftsättningen av VKR hos NPI. Den redogör för installerad hård- och mjukvara, samt vilka tillämpningar som är aktuella.
The MondiPackaging Dynäs pulp & paper mill in Sweden produces strong sack and kraft paper from chemical softwood pulp. Despite good quality a project was initiated by a need for less variability in paper quality, and less energy consumption. It was known that there is variability in incoming stock, and that the HC and LC refining strategies could be improved. The aim of the project is to reduce energy consumption by 10%, and to reduce quality variation by installing a new online fibre analyzer, PulpEye, and implementing modern modelling and control methodology. We have used multivariate subspace models in combination with optimization to create an OPC based interactive application for on-line operator support and optimization of kraft paper production. The application is now installed and is under evaluation. The specific refiner energy has so far been reduced with at least 25%. The decision support application is indicating even more reduction potential.
In real-time industrial process applications, there is often a need to predict the effects of adjustments before actually effectuated. Multivariate prediction models based on historical data works fine for predictions as long as the new X data have the correlation structure preserved. This paper shows that such can anyway be used for �What if �� decision support in presence of correlated X variables, whenever it is possible to create a help model that predicts the influence of the manipulated X variables on the rest of the X variables.
Immense amounts of data are collected into today's modern process monitoring systems. There are, however, few methods that have the capability to grasp the essentials in these, usually heavily correlated, data. The multivariate statistical techniques, principal components (PC) modelling and modelling by projection to latent structures (PLS) are two methods that have a great potential for process monitoring and forecasting in these situations
PID is a prevalent tool of automatic control in both industry and home environment, and PID parameters are often forced to modify because of systematic service on the machines or systems, which is time-costing. The project aims to investigate the possibility of applying neural network and reducing PID configuration in controlling industry process, by means of establishing control models and comparing control performance between conventional PID method and improved PID control based on neural network where two built neural networks are considered as cores to adjust weights which result in the suggested PID parameters. Adaptive learning rate is also applied which is adjusted by the algorithm based on the error changes. Algorithm program is written in Siemens TIA Portal and simulated in Factory I/O. In general, the simulations after analysis have shown that the proposed model has a better performance than conventional PID in terms of steady state, deviations and consistency of control value except tuning time. In the future the author is dedicated to continue improving the mentioned model through quickening learning process, applying better activation function and modifying variable structure and so on.
The importance of preventing failures in bearings has led to a large amount of research being conducted to find methods for fault diagnostics and prognostics. Many of these solutions, such as deep learning methods, require a significant amount of data to perform well. This is a reason why publicly available data are important, and there currently exist several open datasets that contain different conditions and faults. However, one challenge is that almost all of these data come from a laboratory setting, where conditions might differ from those found in an industrial environment where the methods are intended to be used. This also means that there may be characteristics of the industrial data that are important to take into account. Therefore, this study describes a completely new dataset for bearing faults from a pulp mill. The analysis of the data shows that the faults vary significantly in terms of fault development, rotation speed, and the amplitude of the vibration signal. It also suggests that methods built for this environment need to consider that no historical examples of faults in the target domain exist and that external events can occur that are not related to any condition of the bearing.
The present invention relates to a sensor arrangement (100) suitable for determining a condition, for example moisture, comprising a first RFID-unit 5 (110) and a second RFID-unit (120) being subjected to said condition. The sensor arrangement is characterized in that the second RFID-unit (120) is at least partly provided with a degradation means (130) having such properties that, when subjected to said condition, the second RFID-unit (120) is functionally degraded to a greater extent than the first RFID-unit (110). The 10 invention also relates to a sensor arrangement product (199) comprising at least one sensor arrangement.
Knowledge of the road status and specifically knowledge of the freezing point of the road surface fluid is crucial in order to perform effective and environmentally safe road maintenance. Road status sensors installed in the road can be passive conductivity sensors or active freezing point sensors. In this paper the output from a passive and an active sensor has been studied when the sensors has been contaminated with common chemicals that can be present on the road surface such as oil, alcohol and glycol. The results indicated that only intelligent active sensors reliably can detect freezing points on the road surface.
Multivariate time series analysis is applied to understand and model the dynamics of an electrolytic process manufacturing copper. Here, eight metal impurities were measured, twice daily, over a period of one year, to characterize the quality of the copper. In the data analysis, these eight variables were summarized by means of principal component analysis PCA.. Two principal component PC.scores were sufficient to well summarize the eight measured variables R2s0.67.. Subse-quently, the dynamics of these PC-scores latent variables.were investigated using multivariate time series analysis, i.e., par-tial least squares PLS.modelling of the lagged latent variables. Stochastic models of the auto-regressive moving average ARMA.family were appropriate for both PC-scores. Hence, the dynamics of both scores make the exponentially weighted moving average EWMA.control chart suitable for process monitoring.
In chemical kinetics and batch processes K variables are measured on the batches at regular time intervals. This gives a J×K matrix for each batch (J time points times K variables). Consequently, a set of N normal batches gives a three-way matrix of dimension (N×J×K). The case when batches have different length is also discussed. In a typical industrial application of batch modelling, the purpose is to diagnose an evolving batch as normal or not, and to obtain indications of variables that together behave abnormally in batch process upsets. Other applications giving the same form of data include pharmaco-kinetics, clinical and pharmacological trials where patients (or mice) are followed over time, material stability testing and other kinetic investigations. A new approach to the multivariate modelling of three-way kinetic and batch process data is presented. This approach is based on an initial PLS analysis of the ((N×J)×K) unfolded matrix ((batch×time)×variables) with ‘local time' used as a single y-variable. This is followed by a simple statistical analysis of the resulting scores and results in multivariate control charts suitable for monitoring the kinetics of new experiments or batches. ‘Upsets' are effectively diagnosed in these charts, and variables contributing to the upsets are indicated in contribution plots. In addition, the degree of ‘maturity' of the batch can be as predicted vs. observed local time. The analysis of batch data with respect to various questions is discussed with respect to typical objectives, overview and summary, classification, and quantitative modelling. This is illustrated by an industrial example of yeast production.
In this study, a convolutional neural network (CNN) model was developed for non-destructive damage classification of concrete materials based on acoustic emission techniques. The raw acoustic emission signal is used as the network model input, while the damage type is used as the output. In the study, 15,000 acoustic emission signals were used as the dataset, of which 12,000 signals were used for training, 1,500 signals for validation, and 1,500 signals for testing. Adaptive moment estimation (Adam) was used as the learning algorithm. Batch normalization and dropout layers were used to solve the overfitting problem generated in earlier versions of the model. The proposed model achieves an accuracy of 99.70% with 20,243 parameters, which provides a significant improvement over previous models. As a result, the classification of damages and decisions based upon them in non-destructive structural health monitoring applications can be improved.