DESIGN AND DEVELOPMENT OF AN AUTOMATED VIDEO CAMERA SYSTEM FOR PRESENTING AND ONLINE TEACHING
Abstract – The Covid-19 pandemic had a greater impact on the day-to-day life of people around the world. Due to the pandemic, most countries have started to change their education system into an ‘online teaching method’ and have tried to carry out the educational system unstopped. As a result of this pandemic, most of the higher education sector and secondary education sector in the world has started to offer courses in an online manner. In Sri Lanka, most of the teachers use online meeting software applications to communicate with children via the inbuilt camera in the instrument they are using. Sometimes teachers shared their screens with students and use slides or lecture notes while teaching. Usually, teachers or lecturers used several blackboards or whiteboards to explain the subjects in the face-to-face classroom session. In an online delivery system,some teachers teach using a board and film the lecture using a separate camera and a person to operate the camera. This method has several issues, cannot capture several boards at one time clearly, and so on. If the lecturer uses another board for teaching purposes without regular ones, the camera cannotmove directly in the board’s direction without an operator. Because the camera frame is focused on a single normal teachingboard and if the camera changes its capturing position manually, it causes unclear video capturing. Therefore, this research paper is presented to overcome those problems using an automated camera system. .
Keywords—Online Teaching, Computer Vision, Pan-tilt Mechanism.
Authors —K.K.B. Kariyawasam1 , K.A.V. Perera2 , M.P.H. Nuwan3 , D.D.D. Suraweera
AN ARTIFICIAL INTELLIGENCE-BASED ALGORITHM FOR THE VISUALLY IMPAIRED TO LISTEN TO THECONTENTS OF TEXT IN AN IMAGE
Abstract – This paper presents an algorithm that can be used in a smart Computer vision device for the visually impaired. Complete or virtually full loss of eyesight is referred to as blindness. Wordsin an image object can be difficult for the blind to decipher. This program allows the blind to read written and handwritten material with the same ease as sighted people. For the benefit of the visually impaired, this system’s primary purpose is to convert the written text included in an image into spoken language, and in order to create the algorithm, we use Artificial Intelligence. Those who are blind or have low vision will find this approach especially helpful to listen to necessary content.
Keywords—Artificial Intelligence; Visual Impairment; Algorithm; Computer Vision; Text to Speech.
Authors —E.A.G.A. Edirisinghe, J.A.S.Y. Jayasinghe
ENHANCING VOLTAGE REGULATION IN DISTRIBUTION FEEDERS WITHHIGH PHOTOVOLTAIC PENETRATION FROM ROOFTOP SOLAR
Abstract—In recent times, much attention has been given to developing renewable energy sources and utilizing them to their maximum capacity. In Sri Lanka, solar power has topped the list. One of the adverse effects of increased Photo-voltaic penetration on the low-voltage distribution system is a decrease in reliability which is the root cause of overvoltage. This research has examined the impacts of overvoltage in roof-top PV on the distribution network and evaluates possible mitigation actions in terms of voltage and power quality. A simulation is done, and a real-time implementation of a buck converter is designed for roof-top solar PV users to limit the power output to a value at the threshold of commencement of overvoltage without tripping due to high photovoltaic penetration. The results show that the developed controller can drive the output voltage of the buck converter to the desired voltage reference and reduce a pre-defined power output regardless of solar PV input voltage variations.
Keywords—rooftop solar PV, overvoltage, high photo-voltaic penetration, power electronics, system stability.
Authors —J.R.Lucas1 , P.S.N.de Silva , C.D. Pathirana ,
THE CONCEPTUAL DESIGN FOR ENHANCING CONSTRUCTION EFFICIENCY AND PRODUCTIVITY THROUGH THE USE OF AN AUTOMATIC FLOOR LEVELLINGMACHINE
Abstract—In today’sfast-paced society, efficiency and speed are highly valued, especially in the construction industry.To address the challenges of irregularities and slowness in the construction process, various tools and machines have been developed to improve productivity and streamline workflows. One such innovative machine is designed to automate the manual leveling task, which is known to be both tedious and complex. By automating this process, the construction industry can achieve higher precision and ease of use, improving efficiency. The proposed machine utilizes the Arduino method for its programming, offering several benefits. It allows for better control of mechanical parts, efficient time management, ease of programming, increased profitability, and space efficiency. These features enhance productivity and make the machine an attractive investment for the construction industry. The construction industry can save time and money by replacing traditional manual methods, which are time- consuming and require skilled labor, with automated floor leveling technology. The automation reduces labor costs, increases precision, and improves overall workflow efficiency, leading to higher-quality work and timely project completion. In summary, the development of machines designed for automated floor leveling represents a significant advancement in improving the efficiency and productivity of the constructionindustry. Using sensitive, high-tech devices and Arduinoprogramming provides several advantages over traditional manual methods. These advantages include increased precision,reduced labor costs, and improved workflow efficiency. As the construction industry continually embraces new technological advancements, this technology represents a valuable investment to enhance construction processes further.
Keywords—Flow levelling, Robotics, Efficiency, Productivity, Arduino.
Authors —B.U.G.C. Gunawardhana, G.G.V.P. Rajakaruna, H.M.R.G. Herath, Sanjeewa D.A Sondarangalla
ANOMALY DETECTION IN TRADING BY MODELING THE UNDERLYING STOCHASTIC PROCESSES
Abstract—We present a new algorithm to estimate a profile of trading and detect instances of market manipulation. The algorithm estimates a profile of the changes to the demand and supply of a stock. The algorithm estimates the pattern of trading as a sequence of dependent events with each event contributing to the outcome of the market. The events are grouped in the order in which they are generated to form sequences. The events segment trading into sequences of a fixed number of events delimited as windows. The time interval of the sequence and length of window in time will vary depending on the activity in the market. The window controls the resolution of the analysis. The amount of information about the variation lost from averaging. Each event is an order of a specific type, price and volume in time. The sequence of orders placed is modeled as a Markov process with probability of transitioning from one type of order to another estimated from past observations in the sequence. The sequences can be compared by comparing the transition matrices. The volume of an order is modeled as being composed of a series of orders of the same type but having a predefined unit volume to make the total effect of the series equivalent to the volume of the original market order. The algorithm estimates the distribution in the data adapting to changes. It can adapt to new markets and conditions. It operates by comparing trading sessions modeled as sequences to detect those that are outliers. The outliers can be detected onthe variables used to describe the trading and include those patterns that are trading anomalies. K-Means clustering, principal component analysis and Mahalanobis distance methods of analysis are applied to variables defined on the sequence to identify outliers corresponding to similarity, correlation in types of orders observed during a window, and the distribution in the data respectively. The three kinds of outliers are classified into segments by the degree of overlap in a Venn diagram. This model of trading relates prices to the sequence in which the orders are placed. The estimated profile of trading can be used to provide information to trade by using past patterns to estimate future trends. A windowed technique where the performance scales with the number of calculations needed to process a sequence of a fixed number of events.
Keywords—machine learning; survelliance; markov random process; market manipulation; anomaly detection; outlier detection; Mahalanobis distance; principal component analysis; K-Means cluster analysis.
Authors —Asoka Korale
ANALYSIS OF APPLICATIONS OF LINE DROP COMPENSATION FOR LOW-COST SUBSTATIONS INSRI LANKA
Abstract—Voltage drop at distant locations of distribution lines create problems in many ways. Since most distribution transformers are equipped with Off-Circuit Tap Changers (OCTC), consumer voltage drops below the recommended limits. In Sri Lanka, this phenomenon has caused under voltages below 400 V at consumer side at long distances. So far, the solution has been to construct new substations at suchlocations. However, it is possible to use the “Line drop compensation” (LDC) option available in power transformer OLTC to mitigate this issue for some extent. In this study, we present up to which lengths line drop compensation option could be used to regulate line voltage at nominal level using a simulation model in DIgSILENT PowerFactory. Actual line, and equipment parameters were used in this study. Results show that load power factor has a direct correlation with the line length. Further, Automatic Voltage Regulator (AVR) and capacitor bank combination could be used to go for increased line distances. In conclusion, these options could be implemented in the Sri Lankan transmission network to increase the coverage of transmission grid substations. We can avoid costs on new transmission grid substations with this.
Keywords—Line drop compensation; voltage drop; OLTC; DIgSILENT; substations.
Authors —B.L.E.A. Balasuriya, G.W.I.M. Gunathilaka
ENERGY CONSERVING ECONOMICAL AUTOMATED SMART CONTROL SYSTEM FOR STREET LIGHTING
Abstract—In a set of street lighting systems, there are switches fixed to the street light poles that are situated along the roads, usually, people neglect to switch off and wasting energy during the day. A set of street lights with a single switch to turn on and off is another type and this system requires additional wiring for controlling. The street lights that are connected to a photodiode operates according to the sunlight level is another type of street lighting and cannot be operated remotely as it automatically turns the lights on and off. Thus, a massive amount of energy is wasted. The street light system of wireless operated also not suitable for street lights in developing countries such as Sri Lanka, because the wireless automation unit cost is high. Hence it is conclusive that a smart control system for street lighting is a necessity. This research scrutinizes the feasibility of an economical automated system for the street lights which subsequently will reduce power wastage and increase efficiency. By using the planned system, every light can be controlled whenever needed, without additional cost for wiring to control the lights and switches for every pole. For the experiment, home electrical system automation protocol X-10 has been selected as the power line communication protocol. With the proposed system, the control signal can be sent over the power lines. Instead of the requirement of wiring, another line for controlling is advantageous than wired and wireless systems. Fixing the implemented receiver unit for every light is sufficient. The cost of the receiver unit also will be less compared to the existing systems and the cost of energy wastage.
Keywords—: X-10 protocol, power line Communication protocol, energy conservation.
Authors —N.D.G.Thathsara Nanayakkara, W.K.M.Nuwan Bandara, L.G.P.Kasunika Guruge
A NOVAL MODEL FOR OPTIMIZING PUMPING ENERGY CONSUMPTIONIN BIOMASS HIGH PRESSURE HOT WATER GENERATOR
Abstract—High-Pressure Hot Water (HPHW) for a variety of applications in the manufacturing industry, including leaching, curing, and drying. Nowadays HPHW is primarily generated using Biomass Hot Water Generators that require centrifugal pumps to maintain minimum hot water flow and pressure. Attributable to the criticalness in maintain pressure above saturation and evade steaming in the system, designers may allow for an excessive margin of error when deciding the capacity and developed head of centrifugal pumps used for HPHW circulation. An oversized pump will deliver more flow and discharge pressure than is required for the system and therefore impact the smooth synchronization of the HPHW system. To overcome such a scenario, throttling is used to influence the flow rate and discharge pressure in the pumping system by varying system resistance. It is the simplest way to shift the operating point of a centrifugal pump because it uses existing valves fitted inside, upstream, or downstream of the HPHW system. However, control by throttling is the least energy-efficient control method as it is based on the principle of control by generating losses. Existing Lanka BIOMASS 2 HPHW GENERATOR circulation pumps operate in Star-Delta Circuit, while discharge pressure and flow rate are regulated by throttling manual valves at the pump discharge. This study is on the possibility of reducing pumping losses by changing the speed of the circulation pump used in the HPHW generator. The possibility of using Variable Frequency Drives (VFD) to reduce energy consumption was studied based on calculations and verified with pump OEM datasheets. Finally, the technical and economic viability of using VSD for improving pumping efficiency in existing BIOMASS 2 HPHW GENERATOR is presented and discussed.
Keywords—: component; Energy Management, Centrifugal pumps energy conservation, VFD Flow Control, Valve Throttling.
Authors —Hasanka S. Haputhantiri , Damith Dilhan, Malinga Hewawitharana
MALWARE DETECTION OF IOT NETWORKS USING MACHINE LEARNING: AN EXPERIMENTALSTUDY WITH EDGE IIOT DATASET
Abstract—Artificial Intelligence (AI) is a powerful tool that assists security operation analysts in addressing threats in sectors like IIoT and IoT. As cyber-attacks increase, proactive security measures are crucial to protect critical infrastructure and sensitive data. AI-based learning models provide deeper insights into cyber threats, aiding in the development of effective mitigation strategies. This work contributes towards using supervised Machine Learning (ML) algorithms for Internet of things(IoT) sensor data to identify vulnerabilities. Supervised ML algorithms, including KNN, DTC, LR, SVM, and RFC, are evaluated using the “Edge-IIoTset” dataset for malware detection. Results show that RFC achieves 94% accuracy, outperforming other algorithms.
Authors —Nizwan Hamza, MWP Maduranga, H.K.I.S Lakmal, R.P.S Kathriarachchi
IOT BASED SMART PLANT MONITORING SYSTEM WITH AUTOMATED WATERING AND FERTILIZING
Abstract—It is vital in gardening and agriculture to main-tain the plants properly to ensure proper growth and health.Therefore, it is important to continuously monitor whether the plants get water, sunlight, and fertilizers properly. To achieve this “IOT Based Smart Plant Monitoring System with AutomatedWatering and Fertilizing” keeps tracking about important factors surrounding the plants such as soil moisture level, sunlight, humidity, and temperature. By keep tracking above parameters, “IOT Based Smart Plant Monitoring System with Automated Watering and Fertilizing” will be able to automatically water the plants, capable of applying liquid fertilizer automatically under a user-defined time cycle and a horticulture led grow light can also be automatically provided if the system is used in an indoor environment, based on the indoor lighting conditions. Also, these collected sensor data and automated activities can be monitored in real time, manually configured, and operate using a mobile device.
Keywords—IOT, Plant Monitoring, Automated, Fertilizing, Watering, Automated
Authors —Anuradha Jayakody ,Shashika Lokuliyana , Pipuni Wijesiri , I.S.Hathurusinghe , L.D.S.Perera ,P.T.L.Dilshan , W.A.K.M.Perera
Critical Assessment of Machine Learning-Based Approaches for Predicting System Inertia in Power Systems
Abstract—The increasing integration of renewable energy sources in power systems has led to declining system inertia, making power grid stability a significant challenge. Machine learning (ML) techniques have emerged as a promising approach to predicting system inertia and enhancing grid stability. This paper critically assesses various ML-based methods for predicting system inertia in power systems. We will discuss the current state of research, the challenges and limitations of existing ML approaches, and potential future directions for improving prediction accuracy and real-world implementation. This paper aims to provide researchers and practitioners with a comprehensive understanding of ML-based system inertia prediction techniques and their applicability in modern power systems.
Keywords—system inertia, machine learning, power grid stability, renewable energy integration, inertia prediction
Authors —Wijekoon W.M.K.G.V.B.
DESIGN AND IMPLEMENT OF WASTE WATER HEAT RECOVERY UNIT FOR GLOVEMANUFACTURING INDUSTRY
Abstract—High-Pressure Hot Water (HPHW) is used for variety of applications glove manufacturing industry, including leaching, curing, and drying. During leaching process, HPHW is used to heat up liquids inside process tanks and majority of this thermal energy is lost to the drain and considered as an industrial waste. In glove manufactrurig, there are two types of leaching called pre leaching and post leaching in each dip machine. Both processes involves increasing the temperature of fresh water or other chemical in an open tank upto 90 C using HPHW and heat exchangers. The heated content inside these leaching tanks are continously replaced with new liuids and the excess is drained out and transfered to the efluent treatment plant. The drained content usually consist a temperature between 55C to 90C and considered as waste thermal energy lost in the industrial process. The implemented tank is designed based on recovering heat from drain water and preheating incoming fresh water. In addition, special considerations was taken to minimize fouling which may contribute to the loss of efficiency in the WWHRU.
Keywords—component; Waster Water Hear Recovery System ,Energy Management, Centrifugal pumps energy conservation, VFD Flow Control, Valve Throttling.
Authors —Jayashani Ridmika