Search

Detection of single-phase induction motor failure through the use of accelerometer and temperature sensor via Arduino is accomplished in this paper. A prototype was developed to measure the vibration velocity and temperature of a single-phase motor with a nominal power of 60 W and a rated voltage of 230 V AC . Simulation via LabVIEW was also developed to characterize the motor whether it is in good, bad, or in critical condition. The system made use of Fast Fourier Transform (FFT) Analysis to convert distorted waveforms into a simpler graph, which makes it possible to determine the amplitude of the data to be able to identify the condition of the motor. The motor conditions were characterized using the ISO 10816-1 Std. and IEC 85 Std. for the vibration and temperature, respectively. The gathered data is processed through training and validation which obtained a total of 648 data sets.

The design of 300 watts LLC half-bridge resonant converter operating within an input range of 350 Vdc to 460 Vdc and output of 48–50 Vdc with the implementation of synchronous rectifier on the secondary side is presented. The design procedure and mathematical computation are likewise incorporated. Simulation result shows that the converter achieves the desired output with the ripple and noise voltage contained below 500mV pk–pk.The simulation for the frequency response also yielded an output of a gain margin of –18dB and a phase margin of 48 degrees which complies with the stability requirements. Likewise, the design achieves zero-volt switching and can fulfill the 20ms holdup time requirement. Meanwhile, the selection of components helps the design to achieve an efficiency of 97.30% at full load and further evaluated to meet the requirements of 80 Plus Gold Efficiency Standard.

Power supplies are evolving towards the future and are now moving towards the integration of digital control systems in their circuits. This research is in line with this development and is aimed to show what a digital system can do to protect a power supply. The purpose of this research is to design an auxiliary circuit that will power all auxiliary components and implement a housekeeping for protection against voltage and current abnormalities of the power supply. The auxiliary circuit uses a quasi-resonant flyback converter topology to improve the efficiency of the device while the housekeeping system uses a DSPIC as its core processor to make use of the different signals coming from the power supply. The DSPIC will control the turn-on and turn-off sequence as well as the different fault protections to properly power up and power down the whole system. The design achieved has inputs ranging from 300 to 450 Vdc with a regulated output of 12-volts ±4% and has achieved a phase margin of 47.37 degrees and a gain margin of 12.12dB. The housekeeping system was also successful in detecting and reacting against different faults.

The usual way of checking the attendance in a class has its own drawbacks. To be able to resolve it, automated attendance systems were introduced. In this paper, the design and development of an e-attendance checker using a facial recognition system were implemented. It can scan the faces of multiple students in a standard classroom setup. A commonly used approach for face detection called Histogram of Oriented Gradients (HOG) with Support Vector Machine (SVM) was applied to examine the effect of luminance of the surrounding, the facial orientation of the student and so as their distance from the camera in the facial detection and recognition. The obtained attendance will then be uploaded to a database with authentication. It was found that the system has an accuracy of 95.65% and can detect and recognize up to 37 students. It is suggested that the classroom should have a luminance level of about 217.39 lux or higher to achieve a better accuracy performance of the system. As for the analysis of the effect of distance in the system, it is claimed that the distance of the student does not affect the accuracy of the system. Lastly, it is suggested that the face angles of the subject should be directly facing the camera to achieve a more accurate recognition result.

Wireless communications become unavailable due to power losses and natural disasters that disrupt power distribution. The study aims to create a localized resilient communication system that can maintain adequate service in the midst of disasters and power outages. An appropriate solar-powered mechanism was designed to maintain system operation, having an average switching time of 0.849 seconds. To establish communication among users, an Android smartphone application was integrated into the network, consisting of a TP-LINK WR840N wireless router and TP-LINK CPE220 access points, through a Raspberry Pi acting as a chat server. Testing shows that the system can route messages with a 100% accuracy with 16 maximum users simultaneously connected within a 140-m radius. While simulating power losses, the system exhibited a 0% chance of going down. The developed system is resilient enough to withstand power outages as well as maintain its normal operation under standard and disaster conditions. The study would be beneficial in disaster and risk mitigation as well as providing wireless communications to areas with little-to-no coverage.

an approach in testing the absolute maximum radio frequency (RF) input power rating of an RF and microwave power amplifier is presented. A developed test method applying pretest, posttest, and destructive RF input power stress sweeps were used to verify the safe points and survivability limit of a gallium nitride microwave monolithic integrated circuit. The test method created incorporates actual DC and RF measuring equipment for collecting and logging current and power levels from the device under test. The developed test method and actual implementation proved to accurately measure both the absolute maximum rating safe datasheet specification and the actual RF input power where the part reaches actual breakdown. At these conditions, it is assured that the actual characterization and results are robust and is aligned to simulated absolute maximum rating specified in the datasheet. The approach proved to be a milestone in evaluating and characterizing RF active devices.

Air Pollution is an environmental and a social problem which has harmful effects on ecosystems, climate, life expectancy and quality of life. It is counted in the largest environmental health hazards in the Philippines especially in the nation’s urban areas. This study developed a platform that can monitor air quality in the Philippines wherein the environmental parameters being monitored are ambient temperature, humidity, particulate matter and carbon monoxide concentration levels. This study applies Dual Sensor Redundancy Mechanism to examine the validity of the sensor readings. Using Wi-Fi technology, the data obtained from the microcomputer is sent to the Cloud based Web Server. The output of the system which involves sensor readings and Graphical Representation with respect to time is displayed on the mobile application. After validating the prototypes, the percent difference obtained fall under 5% mark and by conducting Two-tailed T-test, it is determined that the prototypes are functioning correctly. The platform is also implemented in two remote areas and environmental conditions are compared using the data obtained during the monitoring process.

Data is everyone’s need in each field of human life, either individuals or groups and whether in each personal, company, business, organization, government, etcetera. Because it is a factual source of information and knowledge that one can use for a more accurate, reliable and intelligent decision-making along each struggle, challenges and needs. These data are either machine-readable, also known as structured data that requires computer programs to process or a human-readable which is also known as unstructured data that needs a person to interpret the information. This study will be dealing with the High Frequency Doppler Radar (HFDR) data, a structured data processed and analyzed using the algorithms for Ship Detection and Tracking. Moreover, processing and analyzing these data are not enough because human brain tends to find patterns more efficiently when data is presented. Thus, a visual representation of these processed data was carried out as a result of this study so that decision makers could easily figure out the complete picture of the system-generated information consisting the plotted detections and track records.

This study aims to cover the gap between the geometry and topology in drainage network classification by creating a model produced using a weighted flow function. The proposed methodology focused on building a model from Object-Based Image Analysis (OBIA) based from the sophisticated parameter to be used as the weight input to morphological analysis. This study performed segmentation evaluation using Area Fit Index. Further, to improve classification performance, this study performed feature selection using InfoGain based from the stratified random sampling in python.

With the use of Light Detection and Ranging (LiDAR) Data, this study focuses on the processing of the LiDAR derived data through different software tools to generate a map that can classify forest types. A 20 × 20 meter plot in the selected forest area was identified in this study for the field validation of the classified leaf type. Leaf recognition is performed using Neural Network in Matlab. The leaf statistics were measured through the prototype developed using leaf extraction algorithms T-test is used for the comparative measurement between the perimeter of the extracted data and the actual perimeter of a sample leaf. The result shows that for the specie, the actual perimeter is statistically the same with the perimeter measured by the developed prototype. The accuracy of classification was calculated as 91.25%. The overall minimum and maximum precision of the prototype is computed to be 90.40% and 99.14%, respectively.

In classifying plant leaves' spectral signature, different approach were already taken to achieve the general results. There are different Spectral Signatures that a leaf may possess on each reading using the standard and calibrated in the field. These are Spectral properties of different tropical plants found in tropical regions. This study focus on the spectral readings and variations in the spectral signatures found on leaf specifically plants that grows on Arid lands, Wet lands, and Near the Sea. The study uses a mechanism that shows how Spectral Signatures of these plants may vary in readings based on the location where they can be found. This study applies the algorithm known as Fast Fourier Transform in analyzing the gathered spectral data from tropical plants using MATLAB. The processed Reflectance through FFT shows how the Spectral Signature of each plant per location are different from each other. The following values is the peak values of Reflectance — FFT results for each plant: Banana 0.9466 @190Hz (Arid land), 1.112 @190Hz (Near the Sea) and 1.5408 @188Hz (Wet land); Mango: 0.723 @189Hz (Arid land), 1.2401 @189Hz (Near the Sea) and 1.634 @183Hz (Wet land); Coconut: 1.1278 @194Hz (Arid land), 0.8246 @194Hz (Near the Sea) and 1.0816 @194Hz (Wet land); Avocado: 0.8164 @194Hz (Arid land), 0.3852 @194Hz (Near the Sea) and 0.9661 @190Hz (Wet land); Pineapple: 0.654 @189Hz (Arid land), 1.3586 @189Hz (Near the Sea) and 1.524 @183Hz (Wet land).

There are several ways to determine if low-lying regions are prone to flooding, one is to set threshold values that will serve as basis to determine if the area is below sea level, and then that area could be a flood prone area depending on the measured level. Another situation is that not all area situated below sea level is flood prone and vice-versa. In this study, threshold was used to serve as reference in processing LiDAR data to determine the condition of each region under data processing. The method provided in this study provided an efficient way in looking at particular map regions to identify the elevations based on the Delaunay Triangulation Technique. The said method also allows the determination the different levels under a particular region.

The major disadvantage of Support Vector Machine (SVM) happens in its training phase as it requires to solve a quadratic programming problem, making computation very costly. With the integration of LiDAR data and high spatial resolution orthophoto, more input data layers are available for object-based Support Vector Machine classification. Initially, confusion among classes arises because of the presence of irrelevant and redundant data layers. Hence, this study applies Decision Tree (DT), a popular data mining technique, as a pre-classification process in SVM to select the relevant features from the input variables. We assessed the effectiveness of seven vegetation indices, two vegetation index combinations and 14 LiDAR height metrics for mapping agricultural resources in Calatagan, Batangas. We were able to filter the input variables and subsequently achieve at least 73% reduction of training features. With the DT-based feature selection, we were able to reduce the number of input features as well as make the SVM training and classification time shorter by more than 90%. Importantly, the overall accuracy and kappa index of agreement both increased when DT-based SVM was used in contrast with using all the variables for SVM classification.

The use of unmanned aerial vehicles (UAV) propelled the growing interest of small to medium farmers in precision agriculture applications. In this study, a methodology was developed to automatically map the land use in a vegetable farm with a very high resolution aerial image (5cm/pixel) taken at an altitude of 100 m. Using color index of vegetation extraction and Otsu's thresholding method, the soil was delineated from vegetation by object-based technique, and subsequently segmented the vegetation using multi-resolution algorithm. The Support Vector Machine classification on various vegetation indices produced agreeable results, where eggplant, corn, string beans and grass/weeds were classified. The classification was further improved by including Gray Level Co-Occurrence Measures or textural features, as indicated by the increase in overall accuracy and kappa index of agreement. The output map can serve as guide for farmers and other management agencies to know the correct farm interventions such as weed control. The output map can also be used in the periodic updating of the initial land cover map produced from high resolution LiDAR data.