SIMM2023 Toward Empowering Technological Transformation 51 Paper ID: SIMM2023: 028-024 Title: A Soft Continuum Manipulator for Multi-Environmental Inspection. First Author: Mohamed Tahir Shoani Co-Author: Mohamed Najib Ribuan; Ahmad Athif Mohd Faudzi Faculty of Electrical & Electronic Eng., UTHM The current methods for inspecting partially submerged structures, requires employing different types of robots to inspect the parts of the structure above and below the water surface. This increases the cost and complicates the operation. This paper proposes a solution in the form of a teleoperated soft continuum manipulator capable of inspecting a partially submerged structure above and below the water surface. The manipulator employs tendon actuation to control the posture of its thin arm, and a sliding mechanism to alter the length of the bending segment, thus change the location and attitude of the arms tip. The manipulators modular design facilitates the replacement of its arm with another of different length and material to suite different structures and environments. A camera placed at the manipulators arm tip was used to inspect several numeric labels, some below and others above the water surface. The experimental results demonstrated the manipulators ability to inspect the partially submerged structure and provide clear images above and below the water surface. The manipulators ability to use arms of different lengths and materials makes it suitable for operating in different environments unsuitable for current inspection robots and manipulators.
SIMM2023 Toward Empowering Technological Transformation 52 Paper ID: SIMM2023: 042-029 Title: Performance of Extreme Learning Machine. First Author: Fateh Alrahman Kamal Qasem Al-Nagashi Co-Author: Norasmadi Abdul Rahim ([email protected]) Classification is an important task in machine learning and has a wide range of applications in various fields. In this paper, we compare the performance of two popular classification algorithms, the Extreme Learning Machine (ELM) and kNearest Neighbors (KNN) classifiers, on a benchmark dataset. ELM is a feedforward neural network with a single hidden layer that uses random weights for fast training and high accuracy. KNN is a non-parametric classification algorithm that assigns the class label of a test sample based on the majority vote of its k nearest neighbors in the training set. We evaluate the classifiers on 5 datasets, we first preprocess the data by normalizing the features and converting the class labels to numerical values. We then train and test the classifiers on the dataset and report the training and testing accuracy as well as the confusion matrix to evaluate their performance. Our findings suggest that ELM is a promising alternative to KNN for classification tasks, especially in scenarios where the dimensionality of the data is high and the number of training samples is limited. The results also highlight the importance of careful parameter tuning and model selection in achieving high performance in classification tasks.
SIMM2023 Toward Empowering Technological Transformation 53 Paper ID: SIMM2023: 061-045 Title: Oil Palm Fresh Fruit Branch Ripeness Detection using YOLOV6. First Author: Alvi Khan Chowdhury Co-Author: Wan Zailah Binti Wan Said ([email protected]); Sarah Atifah Saruchi ([email protected]) Faculty of Engineering, Technology and Built Engineering, UCSI University, 56000 Kuala Lumpur Malaysia is one of the largest palm oil producers and exporters in the world. As a result, the efficient production of palm oil plays a major role in the Malaysian economy. Currently, oil palm fresh fruit branches (FFBs) are harvested by human graders at oil palm plantations based on the fruit surface colour and the number of loose fruits on the ground as a measure of the ripeness level. However, solely relying on human graders would result in misclassified bunches due to factors such as the position of the FFBs on the trees, the height of the trees and incorrect count of the number of loose fruits on the ground. Therefore, this study proposes an automated oil palm FFBs ripeness detection via a deep learning algorithm namely You Only Look Once Version 6 (YOLOV6). YOLOV6 is used to detect oil palm FFB ripeness levels based on the surface colour of the oil palm fruits. Real images of FFBs at oil palm plantations are taken to be classified into four classes; unripe, underripe, ripe and overripe. Each class differs from other classes of oil palm FFBs based on their colour. The datasets are trained using two YOLOV6 models which are YOLOV6s and YOLOV6m. The proposed algorithm performance is analysed based on the precision, recall, F1 score, mean average precision, training time, inference speed, frames per second and nonmaximum suppression time. The YOLOV6m model outperformed the YOLOV6s model, according to the investigation. For the YOLOV6m model using 100 training epochs, the achieved precision, recall, f1 score, mAP(50), and mAP(50- 95) is 36.9%, 30%, 33.1%, 36.9%, and 16.5%, respectively. The YOLOV6m model automatically detected oil palm FFBs accurately with a predicted bounding box including objectness score for all the four classes of palm oil fresh fruit branches which includes unripe FFB, underripe FFB, ripe FFB and overripe FFB, respectively. It is expected that the finding of this study can be applied as a real implementation at an oil palm plantation in the future.
SIMM2023 Toward Empowering Technological Transformation 54 Paper ID: SIMM2023: 064-047 Title: Classification of Malodor Gases in Greenhouse Environment Using Artificial Intelligence Technique. First Author: Abdul Syafiq Bin Abdull Sukor Co-Author: Mariam binti Majid ([email protected]); Muhamad Alif Aiman bin Jalaludin ([email protected]); Mohd Wafi bin Nasrudin ([email protected]) Faculty of Mechanical Engineering & Technology, Universiti Malaysia Perlis, 02600, Arau, Perlis Malodor gases is a presence of unwanted gases that can be harmful to our health and production of the plant in a greenhouse environment. It is considered an important environmental pollution issue that needs to be monitored. To overcome this problem, the development of instrumentation technology system to detect and classify the presence of malodor gases need to be thoroughly studied. In this paper, an intelligent technique is proposed based on the concept of electronic nose and environmental monitoring technology. The proposed system employs several of chemical gas sensor mounted with Arduino microcontroller to detect the presence of malodor gases. Then, various artificial intelligence methods such as Support Vector Machine (SVM) Artificial Neural Network (ANN) and Decision Tree Algorithm (DTA) are used to classify the types of gasses and compared their performance with each other. Our proposed system can enhance the reproducibility, reliability, and selectivity of odor sensor output. This system utilizes the function of artificial intelligence technique in environmental monitoring, its applications in greenhouse and detecting the smell of particular volatile organic compound like ammonia, carbon dioxide and methane at different concentrations.
SIMM2023 Toward Empowering Technological Transformation 55 Paper ID: SIMM2023: 071-063 Title: SUAS-based Infrared Thermography for Rapid Temperature Reading in Building and Solar Photovoltaic Panels Inspection. First Author: Ahmad Anas Bin Yusof Co-Author: Mohd Faid Yahya; Muhammad Fahmi Miskon; Mohd Khairi Mohamed Nor; Anuar Mohamed Kasim; Mohd Saifuzam Jamri Universiti Teknikal Malaysia Melaka This research investigates the rapid measurement of surface temperature distribution and the factors that affect it on buildings and photovoltaic panels at Universiti Teknikal Malaysia Melaka Main Campus in Durian Tunggal, Melaka. The study utilizes a small unmanned aerial system (SUAS)-based infrared thermography system to move and position the thermal camera to measure temperature in three selected locations: The Chancellery building, the Faculty of Electrical Engineering building, and the Solar Photovoltaic Panel Site. The research findings are presented through detailed thermographic documentation. Reflection, transmission, and emittance are identified as the main factors that can alter apparent temperature readings. The research concludes that the measured temperatures are impacted by the sky background, the wall color, the roofing materials, and the shading of the sun.
SIMM2023 Toward Empowering Technological Transformation 56 Paper ID: SIMM2023: 077-064 Title: Review on Wireless Control Methods of Air Conditioning System for Thermal Comfort in Rooms. First Author: Ismail Ishaq Bin Ibrahim Co-Author: Prof. Madya Dr Shazmin Aniza Binti Abdul Shukor ([email protected]); Prof. Karl Kohlhof ([email protected]); Dr. Muhajir Bin AB. Rahim ([email protected]) UNIVERSITI MALAYSIA PERLIS Thermal comfort plays a big part in everyday lives. Comfortability is always desired whether when working, studying, or even when relaxing at home. Comfort is needed to maintain a stable mental health during long hours of activities. In tropical countries like Malaysia, air conditioning system are often used to create thermal comfort indoors. However, it should be controlled accordingly in order to offer suitable comfort level. By utilizing wireless technologies, all of the parameters can be obtained without interfering with the occupants current activities. This will ensure a more natural and unstaged data accessed. Wireless technology can also be used to control the air conditioner after the data have been analyzed for a better and quick way to maintain the level of comfort without the need of the subject to keep adjusting the air conditioner. In this paper, the most suitable wireless control method for thermal comfort data acquisition and controlling air conditioning system is reviewed.
SIMM2023 Toward Empowering Technological Transformation 57 Paper ID: SIMM2023: 070-065 Title: An IoT Based System to Monitor Soil Parameter of Harumanis Tree. First Author: Muhammad Faiz Aiman Mohamed Zaini Co-Author: Muhamad Khairul Ali Hassan ([email protected]); Fathinul Syahir Ahmad Saad ([email protected]); Sukhairi Sudin ([email protected]); Shafriza Nisha Basah ([email protected]); Khairul Salleh Basaruddin ([email protected]); Muhammad Juhairi Aziz Safar ([email protected]); Haniza Yazid ([email protected]); Mohd Hanafi Mat Som ([email protected]); Muhammad Zunnurrainie Zulkifli ([email protected]) Faculty of Electrical Engineering & Technology, Universiti Malaysia Perlis The Internet of Things (IoT) is a paradigm-shifting technology that represents the future of communication and computing. IoT implementation is extensive and can be used in any industry. Agriculture is becoming an increasingly vital sector as the world's population expands. This project will create an Internet of Thingsbased system to monitor the soil parameters of the Harumanis tree. The main problem in the Harumanis sector is increasing Harumanis production and quality without relying on constant human monitoring. IoT reduces labour costs while improving farm management, cost efficiency, crop monitoring, and crop quantity and quality. This Internet of Things system includes a pH level sensor, a temperature and humidity sensor and soil moisture sensor for monitoring the Harumanis farm. The system is a simple IoT architecture in which sensors will collect data and send it to a cloud database through Wi-Fi. This project was tested at UniMAP's Faculty of Electrical Engineering Technology's Harumanis farm. Every piece of data collected during testing was saved in a cloud database using the Blynk IoT platform, which farmers can use to track the tree's daily growth from their smartphone or laptop. This research is expected to aid farmers in monitoring the development of Harumanis trees and producing the greatest quality Harumanis mangoes possible, as farmers will be able to respond instantly if any anomalies are found via the Blynk App's warning notification.
SIMM2023 Toward Empowering Technological Transformation 58 Paper ID: SIMM2023: 067-066 Title: Acoustic lmpulse and Ultrasound Wave for Measuring the Firmness and lnsidious Fruit Rot of Fruit. First Author: Muhammad Zunnurrainie Bin Zulkifli Co-Author: F. S. A. Saad- Fathinul Syahir Ahmad Saad ([email protected]); M.K.A.Hassan- Muhamad Khairul Bin Ali Hassan ([email protected]); M.F. Ibrahim- Mohd Firdaus Ibrahim ([email protected]); H Yazid- Haniza yazid ([email protected]); MJA Safar- Muhamad Juhairi Aziz Safar ([email protected]); M. Z. M. Faiz Aiman- Muhammad Faiz Aiman Bin Mohamed Zaini ([email protected]) Faculty of Electrical Engineering & Technology, Universiti Malaysia Perlis This project is about development of acoustic impulse response device to measure firmness and Insidious Fruit Rot (IFR) in mango fruit. Harumanis mango farmer face a problem which is an internal tissue break-down which have cost their fruit quality to drop. This problem also occurs in other countries which had different term such as Soft Nose and Jelly Seed. There is almost no device to detect the IFR and the farmer use traditional method. As there is no cure yet for the problem and to detect the IFR is by cut open the mango which is destructive and not productive. This project was developed to solve the problem using non-destructive acoustic impulse response which consist of mechanical excitation to produce vibration of sound and the response will be captured by output sensing medium such as microphone. The signal output will be analyzed and evaluate to differ which fruit have IFR and which do not. This method implements non-destructive method to prevent the device from destroying the fruit and provide a solution for the farmer to sorting their mango fruit. Each fruit sample will be tested three times and the average value will be calculated and classified to determine which fruit have IFR and which does not have IFR. By integrating microcontroller, mechanical excitation module and microphone and some other component and software will help accomplishment the objective of this project.
SIMM2023 Toward Empowering Technological Transformation 59 Paper ID: SIMM2023: 064-080 Title: Performance Comparisons of GNB, RBF-SVM and NN for Stress Levels Classification using Discrete Wavelet Discrete Transform. First Author: Muhammad Rasydan bin Mazlan Co-Author: Abdul Syafiq bin Abdull Sukor ([email protected]); Abdul Hamid bin Adom ([email protected]); Latifah Munirah Kamarudin ([email protected]) Faculty of Electrical Engineering & Technology, Universiti Malaysia Perlis, 02600, Arau, Perlis Mental health disorders have risen rapidly in recent years due to the pandemic Covid-19. Normal daily life was impeded, and livelihood was disrupted causing stress to accumulate. Early recognition of stress has become imperative to avoid long-term exposure leading to mental health disorders. The application of Electroencephalography (EEG) facilitates the need for stress signal identification. By observing the brain wave pattern, stress-related features can be shown through a graphical representation of the Brain-Machine Interface (BMI) device. However, the complexity and huge amount of data recorded from the brains activity make it harder to determine the specific characteristics of stress signals. To overcome that, this study proposed a Discrete-Wavelet Transform (DWT) analysis to extract the stress-related features, and classification was conducted through Artificial Intelligence (AI) algorithms. The recorded EEG data were preprocessed with segmentation and normalization before being labelled into respective stress stages. After that, the decomposition was carried out using DWT before being classified using Gaussian Nave Bayes (GNB), Radial Basis Function Support Machine Vector (RBF-SVM) and Neural Network NN for performance analysis. The results show that the NN model achieved higher classification accuracy (79.62%) compared to others. In addition, the precision, sensitivity and F1-score of NN achieve a value of 80% for all features which is higher than the other classifiers.
SIMM2023 Toward Empowering Technological Transformation 60 6.Modelling & Simulation
SIMM2023 Toward Empowering Technological Transformation 61 Paper ID: SIMM2023: 014-012 Title: Investigation on Accessible CNC Simulation Approaches for Multi-Axis Milling Machining through CAD/CAM. First Author: Lailisa Nur Misman Co-Author: Mohd Salman Abu Mansor (Corresponding Authors: [email protected]) Universiti Sains Malaysia This paper presents accessible approaches of CNC simulation for CNC milling machines that available to be utilized. Before real machining is performed, several simulations are generated to identify any possibility to perform milling machining. Application of CAD/CAM is most often employed for machining the prototypes and real parts of products. Machining parameters can affect the surface quality of products in CNC machining. Unlike 5-axis CNC milling machine, the combination of synchronized movements on 3-axis CNC milling machine restrict user to machine parts with complex shapes. Therefore, the purposes of this investigation are to (i) identify machining parameters for CNC milling in order to machine a part with complex shape, (ii) employ any accessible machining simulation approaches by using three different CNC simulations and (iii) observe the outcome from each machining simulation approach for future improvement. CNC Keller, Mastercam and CATIA are used to identify the applicable methods for machining the part with complex shape by using 3-axis and 5-axis CNC milling machines. With dissimilar arrangements of setup orientations, the total machining times are recommended as the crucial for roughing and finishing processes. Although there are three simulations in this study, fully comparison is not suitable to be made between those three CNC simulations due to its accessibility and restriction.
SIMM2023 Toward Empowering Technological Transformation 62 Paper ID: SIMM2023: 030-015 Title: The Identification of The Modeling Based Estimation (Mbest) Technological Challenges in Engineering Industry Works Estimation. First Author: Norsyakilah Romeli Universiti Malaysia Perlis Modelling Based estimation (MBEST) is a comprehensive visualisation and simulation methods in line with the Building Information Modelling (BIM) based approach for predicting construction costs during the preconstruction phase of a project. Even though there are numerous advantages to use MBEST which equipped with IR 4.0 thrust, the local construction sector is still hesitant to use the technology to deliver its services. This study aims to identify the level of MBEST technological adoption in engineering industrial based construction projects. This study is focusing on discover a performance improvements strategy for MBEST adoption among the industrial practitioner in completing the projects. The mixed method strategy was adopted in this study. The approach of purposive sampling has been selected. This type of sampling, also known as judgement sampling, involves the researcher utilizing their expertise to select the most appropriate sample to the research's objectives. The industrial players were chosen among the civil engineers participating in pre-contract cost estimation will be the sample chosen since their engagement has greatly affected cost projections and estimations for infrastructure activities. The surveys will be administered online. The questionnaire was produced, and the link was shared and emailed to the civil engineering consultant in Malaysia. Data were analyzed using descriptive analysis in IBM SPSS software. For qualitative method, the data were analyzed using thematic analysis. Later, the data produced from the quantitative technique was validated by a semi-structured interview with ACEM civil consulting engineers and triangulated. Furthermore, this research may add to the body of knowledge in the academic industry on MBEST issues. The results will be used to develop future features of modelling-based estimate software, ensuring that the MBEST fits the demands of Malaysian construction professionals.
SIMM2023 Toward Empowering Technological Transformation 63 Paper ID: SIMM2023: 032-017 Title: Development of Smart Intelligent Medicine Monitoring System. First Author: Azian Azamimi Abdullah Co-Author: Kwan Janson; Wei Chern Ang Universiti Malaysia Perlis The medication plays an important role in everyday life for a person suffering from various diseases. If a person who fails to follow taking medication as prescribed or as schedule may increase the risks of health and in the worst cause of death. This situation can consider medication errors such as medication mistakes, incorrect dosage, double dose, caused by forgetfulness, confusion, improper timing in the medication self-management. This project aims to enhance the monitoring system as a reminder for a person to take medication and to monitor the elderly to take the medication. To make the monitoring system smart, the system connects to the wireless network, the system can preset the alarm time interval to take medication by the guardian. The smart monitoring utilizes reminder system using a buzzer and the light from LED is to create the alarm to remind the elderly or user to take the medication. The system able to monitor in real-time by receive the notification and trace the patient if they are taken the wrong medication.
SIMM2023 Toward Empowering Technological Transformation 64 Paper ID: SIMM2023: 034-020 Title: Modeling of Portable Water Purification System by Using III-V Multijunction Solar Cell and Cationic Polyacrylamide Alum to Improve the Efficiency. First Author: Muhammad Shehram Co-Author: Aeizaal Azman A.wahab, [email protected]; Uzma Rani, [email protected] University Sains Malaysia In most developing country municipally gave water is not microbiologically ensured. This evaluation looked into drinking water quality and the effect of home water purification attempts. Commonwealth countries facing a shortage of potable water and a deficiency of storage systems for agriculture and electricity generation. Water purification is over with by using renewable energy resources especially solar system which is cheap as compare to other energy resources. Multijunction PV cells are utilized to get better performance of the water purification system as compared to the existing system. In our research work, cationic polyacrylamide (CPAM) and alum are used for drinking water treatment. Cationic polyacrylamide (CPAM) and alum were then applied to oculate the untreated water. The performance of the system water treatment depends upon the condition of the flocculation. Emphasis framework of the untreated water like PH, temperature, time taken for the settling of residues, total amount of addition of flocculation, and the type of flocculant by optimizing this understand the performance of flocculation. The removal of efficiency (more than 98%) was outstanding regarding turbidity and color removal. Cationic polyacrylamide CPAM and alum are initiates to be environmentally good, ultra-efficient and fast settling in potable water treatment. By using these things efficiency, of the water system improved on both sides energy usage and purification quality.
SIMM2023 Toward Empowering Technological Transformation 65 Paper ID: SIMM2023: 041-026 Title: Data Detection of Blind Selective Mapping Using Soft Output Viterbi Algorithm (SOVA). First Author: Aeizaal Azman Abdul Wahab Co-Author: Nur Qamarina Muhammad Adnan University Sains Malaysia Orthogonal Frequency Division Multiplexing (OFDM) is a well know multiplexing modulation scheme and widely used in telecommunication applications due to huge advantages given. Unfortunately, OFDM also possess a huge drawback which is high Peak to Average Power Ratio (PAPR) that can affect the system badly. Selective Mapping (SLM) is one of the most reliable reduction techniques but it requires the transmission of side information to the receiver which can reduce the quality of the signal transmission. In this paper, Soft Output Viterbi Algorithm (SOVA) was proposed as blind estimation and compared to the other estimation method. SOVA has proven to provide the best Bit Error Rate (BER) performance when compared to the other methods.
SIMM2023 Toward Empowering Technological Transformation 66 Paper ID: SIMM2023: 041-028 Title: Particle Swarm Optimization for PAPR Reduction of OFDM Systems. First Author: Aeizaal Azman Abdul Wahab Co-Author: Nur Qamarina Muhammad Adnan; Syed Sahal Nazli Alhady; Wan Amir Fuad Wajdi Othman University Sains Malaysia Orthogonal frequency division multiplexing (OFDM) is a multicarrier modulation technique used in wireless networks for high-speed data transfer. It also offers greater immunity to multipath fading, impulse noise and power efficiency. Despite having those advantages, OFDM does have a high peak to average power ratio (PAPR) problem. One of the most popular method for reducing the PAPR in OFDM systems is partial transmit sequences (PTS). Nonetheless, the partial transmits sequence (PTS) technique usually requires an exhaustive search over all possible combinations of phase weighting factors and the search complexity grows exponentially with the number of sub-blocks. Besides, from previous researches, the Particle Swarm Optimization (PSO) is able to reduce the computational complexity of the PTS technique. Thus, in order to solve this problem, PSO-based PTS is used in OFDM systems to reduce the computational complexity of both PTS and the original PSO, as well as the PAPR. The modulation scheme used is Quadrature Phase Shift Keying (QPSK). At the end of this process, by using PSO-based PTS, PAPR values are successfully reduced around 9dB, however for PTS Technique, its only able to reduce the PAPR values around 3dB to 6dB.
SIMM2023 Toward Empowering Technological Transformation 67 Paper ID: SIMM2023: 027-033 Title: Water Quality Modelling of Langat River Basin Using HEC-RAS. First Author: Nor Faiza Abd Rahman Co-Author: Tatenda Mazarire; Vin Cent Tai ([email protected]); Munira Mohammad ([email protected]); Mohamad Shakri Mohmad Shariff ([email protected]); Khairi Khalid ([email protected]) SEGi University This study aimed to evaluate the water quality of Langat River, Malaysia by analyzing the concentrations of BOD and ammonium-nitrate and their relationship with rainfall patterns. Water samples were collected from Langat River on a monthly basis from 2016 to 2018 and analyzed for BOD and ammonium-nitrate concentrations. The HEC-RAS program was used to assess the river flow, and the relationship between rainfall and water quality was analyzed. The study found that BOD and ammonium-nitrate concentrations in Langat River were influenced by rainfall patterns. The highest BOD concentration was observed in January and June 2017, while the highest ammonium-nitrate concentration was recorded in January 2017. Both pollutants showed a direct relationship with rainfall intensity, with the NE monsoon season having the highest concentrations. However, strict government policies on pollution prevention and control were found to have reduced the concentrations of both pollutants in the river. This study provides valuable information on the water quality of Langat River and highlights the importance of monitoring and managing water resources to prevent pollution. The use of the HEC-RAS program in assessing river flow provides a useful tool for future studies on water resource management.
SIMM2023 Toward Empowering Technological Transformation 68 Paper ID: SIMM2023: 022-034 Title: Evaluating the Effectiveness of Exponentially Weighted Moving Average Filter in Enhancing Landslide Detection from Accelerometer Data. First Author: Suardi Kaharuddin Co-Author: Prof Dr MOHD FADZIL AIN ([email protected]); MOHD NADZRI MAMAT ([email protected]); MOHAMAD NAZIR ABDULLAH ([email protected]) Universiti Sains Malaysia Landslides are a major natural hazard that can cause significant damage to infrastructure, loss of life, and economic disruption. Early detection of landslides can greatly reduce the impact of these events and is crucial for effective risk management and mitigation. In this study, we propose a method of landslide detection using accelerometer-based soil movement sensing and exponentially weighted moving average (EWMA) digital filtering. The accelerometer data is processed using exponentially weighted moving average digital filtering to reduce the noise in the data, which is then used to detect ground motion. The results demonstrate a statistically significant improvement with the use of the EWMA method, as indicated by the reduced root mean square error (RMSE) values. Specifically, the raw data RMSE was 0.1103, whereas the EWMA RMSE was 0.0331. The standard deviation of the raw data was 0.1072, while that of the EWMA was 0.0201. The proposed method has the potential to improve the early warning capabilities of landslide monitoring systems and enhance the safety of people living in landslide-prone areas.
SIMM2023 Toward Empowering Technological Transformation 69 Paper ID: SIMM2023: 027-035 Title: Analysis of Annual Cumulative Rainfall Events in Cameron Highlands, Malaysia. First Author: Nor Faiza Abd Rahman Co-Author: EeLi Siew ([email protected]); Vin Cent Tai ([email protected]); Munira Mohammad ([email protected]); Mohamad Shakri Mohmad Shariff ([email protected]); Khairi Khalid ([email protected]); Yannick Mondelly ([email protected]) SEGi University This paper presents an analysis of the annual cumulative rainfall pattern and probability of extreme rainfall events in Cameron Highlands, Malaysia. The study was conducted using 30 years of rainfall data from three data stations. The results show that the annual cumulative rainfall pattern of each station is consistent, with the lowest recorded in 1979 and the maximum recorded in 1994 or 1999. Two monsoon seasons exist in Malaysia, and Cameron Highlands is not affected by the North-East Monsoon. The study also analyzed the probability of extreme rainfall events occurring over different time periods using the Weibull distribution and Average Recurrence Interval (ARI). The results can be useful for the design of drainage systems in Malaysia, particularly for major drainage systems that are designed to withstand storms with ARI of 50 to 100 years. This paper provides valuable information for hydrological design, enabling engineers and planners to make informed decisions about the design of drainage systems, taking into consideration the probability of extreme rainfall events occurring over different time periods.
SIMM2023 Toward Empowering Technological Transformation 70 Paper ID: SIMM2023: 027-037 Title: Rainfall Data Analysis in Kerian River Basin Using Hyfran-Plus Model, Malaysia. First Author: Nor Faiza Abd Rahman Co-Author: Yannick Mondelly ([email protected]); Munira Mohammad ([email protected]); Vin Cent Tai ([email protected]); Mohamad Shakri Mohmad Shariff ([email protected]); Khairi Khalid ([email protected]); EeLi Siew ([email protected]) SEGi University This study aimed to analyze 11 years of rainfall data from stations within the Kerian River Basin using HYFRAN-PLUS software. The statistical values generated for each station were used to test for independence and stationarity, and the results showed that most p-values were below 0.05, indicating potential non-independence and non-stationarity in the data. Four stations, namely Pusat Kesihatan Kecil, Kolam Air Jkr, Terap, and Kawasan Sg. Acheh, underwent independent and stationary analysis, and it was found that the annual maximum rainfall data remained within the lower and upper control bands of 95% confidence intervals. This suggests that the best-fitted Probability Density Function accurately describes the rainfall. The study emphasizes the importance of validating data to ensure the accuracy and re-liability of recorded rainfall data. It is crucial to identify and address non-independence and non-stationarity in the recorded data before using it for further analysis or decision-making. The study also highlights the usefulness of Non-exceedance probability (NEP) plots in assessing the fit of Probability Density Functions and estimating the return period of extreme rainfall events.
SIMM2023 Toward Empowering Technological Transformation 71 Paper ID: SIMM2023: 052-039 Title: A Study of Micro Shock Hazard in EMIS-based pH Sensor during Fetal Acidosis Assessment. First Author: Siti Fatimah Abdul Halim Co-Author: Zulkarnay Zakaria ([email protected]); Engku Ismail Engku-Husna ([email protected]); Ahmad Nasrul Norali ([email protected]); Jaysuman Pusppanathan ([email protected]); Anas Mohd Noor ([email protected]); Mohd Hafiz Fazalul Rahiman ([email protected]); Siti Zarina Mohd Muji ([email protected]); Ruzairi Abdul Rahim ([email protected]); Aiman Abdulrahman Ahmed ([email protected]) Biomedical Electronic Engineering, Faculty of Electronic Engineering & Technology, Universiti Malaysia Perlis, Arau, 02600 Perlis, Malaysia Fetal acidosis assessment is crucial during labor. Diagnosis of fetal acidosis will help to strengthen decision for c-section delivery in fetal with oxygen deficiency where pH 7.20 or SpO2 35%. This can be determine using electromagnetic induction spectroscopy technique (EMIS) where it has a bright future to be implemented as a non-invasive fetal monitoring device for fetal acidosis assessment. Development of EMIS pH sensor is important in order to replace the existing invasive fetal blood sampling (FBS) which introduced hazards to the fetus. How-ever, there is still lack of research associated with this application. Here we pre-sent the simulation study of micro shock in EMIS pH sensor using COMSOL Multiphysics. In EMIS application, the excitation current must be enough to in-duce current in low conductivity fetal scalp, at the same time harmless to the fetus. The result shows that 1A is the best excitation current with excitation frequency range 1-5MHz that produce V range induced voltage at the receiver which can be further converted to pH values.
SIMM2023 Toward Empowering Technological Transformation 72 Paper ID: SIMM2023: 063-046 Title: Application of Deep Learning in Building Digital Twin: A Review. First Author: F. A. Ismail Universiti Malaysia Perlis Digital twin technology is a rapidly developing domain that has garnered substantial interest in recent years. It involves the creation of virtual representations of physical assets and systems, such as buildings, to enable real-time monitoring, control, and optimization of building performance. digital twin technology provides a comprehensive and data-driven model of the building, incorporating information from various sources, including design information, operational data, and sensory data. This empowers building managers and operators to conduct analysis and optimize building performance, enhance energy efficiency, and reduce operational costs. The implementation of the deep learning (DL) method will provide more accurate results for the usage of digital twin. The focus of this review is on the various DL for 3 main applications of digital twin which is predictive maintenance, energy optimization, and 3D building model generations. Based on the reviews, the common DL that is being implemented for digital twin has been analyzed, also some of the limitations are discussed.
SIMM2023 Toward Empowering Technological Transformation 73 Paper ID: SIMM2023: 032-048 Title: Brain Tumor Radiogenomic Classification using Deep Learning Algorithms. First Author: Azian Azamimi Abdullah Co-Author: Nur Balqis Hanum Zaharuddin ([email protected]); Nur Farahiyah Mohammad ([email protected]); Latifah Mohamed ([email protected]) Universiti Malaysia Perlis Medical images of brain tumors are essential for identifying the pathology of tumors and making a prompt diagnosis. There are numerous medical imaging methods for brain tumors. The nature of brain tumors can be precisely determined by combining the specific characteristics of each magnetic resonance imaging (MRI) scan modality. The current method of genetic analysis is timeconsuming and necessitates surgical removal of brain tissue samples. According to the World Health Organization, glioblastoma (GBM) is the most severe form of brain cancer (grade five). Following surgical removal of a tumor, the standard treatment combines chemotherapy and radiotherapy. Since radiation can kill both normal and malignant cells, radiotherapy can induce significant adverse effects. The amount of MGMT promoter methylation is determined by analyzing tumor DNA from a biopsy or surgical resection. It is currently employed as a prognostic and predictive marker; nevertheless, intra-tumoral heterogeneity of MGMT promoter methylation can impede whole-tumor characterization, resulting in varied survival outcomes. It has been demonstrated that the level of O6-methylguanine-DNA methyl-transferase (MGMT) promoter methylation affects the efficacy of alkylating drugs, such as Temozolomide (TMZ), the most widely used chemotherapeutic treatment for glioblastoma. As a result, epigenetic suppression of MGMT promoter expression has been utilized as a crucial molecular marker in therapeutic practice. Compared to patients with a low level of MGMT promoter methylation, glioblastoma patients with a high level of MGMT promoter methylation are more sensitive to TMZ and have a superior overall survival (OS). In this study, the MGMT promoter methylation was predicted using deep learning algorithms, which are ResNet-50 and Convolutional Neural Network (CNN). Data acquisition, Exploratory Data Analysis (EDA), image preprocessing and performance evaluation of the models proposed has been done in this project. The accuracy was calculated for evaluating the performance of MGMT prediction. The proposed deep learning algorithms which are ResNet-50 and CNN, have 86% and 85% accuracy, respectively. As a result, ResNet-50 shows a high level of accuracy in predicting MGMT promoter methylation.
SIMM2023 Toward Empowering Technological Transformation 74 Paper ID: SIMM2023: 066-049 Title: Investigating the Impact of Process Parameters on Quality of Injection Molded Plastic Axial Fan Blades: A Moldflow Simulation Study. First Author: M. U. Rosli Co-Author: Muhammad Afiq Mazlan; C.Y.Khor ([email protected]) Universiti Malaysia Perlis The injection molding process is widely used in the manufacturing of plastic components, including axial fan blades. To achieve optimal performance and quality, it is crucial to carefully select and optimize process parameters. In this study, a parametric investigation of the injection molding process for plastic axial fan blades was conducted using Moldflow simulation. Solidworks software was utilized for part modeling, and the simulation results were analyzed to determine the impact of various process parameters on the final product's quality. From results, injection time affects shrinkage and warpage of material, as increasing injection time leads to unstable shrinkage and affects the speed of shrinkage cooling during the chilling process, resulting in warpage. Melt temperature also affects shrinkage and warpage. Increasing temperature creates space for defects, but if the temperature is too high, viscosity increases and the mold may not be filled in the set time. Based on these findings, recommendations for process parameter optimization were made to improve the overall quality of the plastic axial fan blade. The results of this study have important implications for the optimization of the injection molding process for various plastic components.
SIMM2023 Toward Empowering Technological Transformation 75 Paper ID: SIMM2023: 066-050 Title: One Factor at A Time Investigation of Injection Molding Process for ThinWalled Laptop Component via Simulation. First Author: Muhammad Afiq Mohamad Jailani Co-Author: M.U.Rosli ([email protected]); C.Y.Khor ([email protected]) Universiti Malaysia Perlis This study aims to determine the optimal processing parameters for the injection molding process used to produce plastic laptop bottom cases. Thin-walled parts, such as laptop components, can be difficult to produce due to warpage and shrinkage defects. Proper selection of process parameters is essential to achieving high-quality, functional parts. The researchers used Moldflow simulation soft-ware to conduct One Factor at A Time (OFAT) investigations of mold temperature, fill time, fill pressure, and melt temperature to determine their effects on warpage and shrinkage. SolidWorks software was used for part modeling. The study found that injection time significantly impacts warpage, shrinkage, and short shot defects. Mold temperature should be at or below 75°C to control warpage, while melt temperature should be between 255-260°C to prevent short shot defects and control warpage. Higher injection pressure can reduce warpage, but it can cause longer freezing times, which can be addressed by changing gate location and cooling channel design. Increasing injection time can decrease warp-age, but times over 1.2 seconds may cause short shot defects. These findings can be useful for improving the quality of thin-walled plastic parts and reducing manufacturing costs.
SIMM2023 Toward Empowering Technological Transformation 76 Paper ID: SIMM2023: 078-068 Title: Archimedes Optimization Algorithm for PID Controller Design of Buck Converter. First Author: Mohd Ashraf Ahmad Co-Author: Ling Kuok Fong Universiti Malaysia Pahang A closed-loop DC-DC buck converter is a type of power converter that uses a feedback loop to regulate the output voltage in response to changes in load or input voltage. A closed-loop DC-DC buck converter needs a PID controller to maintain its design voltage and safeguard any connected loads. PID controllers need to be fine-tuned in every different load situation to provide the best possible response time with minimum voltage error and voltage overshoot. With that, metaheuristic optimization algorithms can help to tune the PID controller. Accordingly, this paper introduces a well-known optimization algorithm, which is archimedes optimization algorithm (AOA) for tuning the PID controller. The tuning results of AOA are compared to those of other popular optimizers, including the improved Nelder-Mead method (AEONM), artificial ecosystembased optimization (AEO), differential evolution (DE), and particle swarm optimizer (PSO). Comparative findings of statistical boxplot, error-integral performance indexes, and transient response are used to verify the PID controller's performance.
SIMM2023 Toward Empowering Technological Transformation 77 Paper ID: SIMM2023: 006-072 Title: Modelling an Accurate ANN Model with Multiple Inputs to Predict Dimensional Accuracy. First Author: Hani Nasuha Co-Author: Mohd Sazli Saad ([email protected]); Mohamad Ezral Baharudin ([email protected]); Azuwir Mohd Nor ([email protected]); Mohd Zakimi Zakaria ([email protected]) Faculty of Mechanical Engineering & Technology, Universiti Malaysia Perlis, 02600, Arau, Perlis, Malaysia Conventional modelling such as regression models is unable to fully capture underlying relationships between the input and outputs of complicated processes such as Fused Deposition Modelling. This study aims to develop an Artificial Neural Network (ANN) model, a non-conventional model better suited to communicating the relationship between for most influential FDM process parameters such as layer height, infill density, printing speed and printing temperature to the single response of dimensional accuracy. 78 samples were generated using Face Centered Central Composite Design (FCCCD), printed using the Ender 3 V2 printer and measured for errors using Vernier calipers. Utilizing the data, several networks were trained with varying hidden layers and hidden layer neurons to identify the best-performing models. The most accurate models were developed and compared based on the lowest mean squared error (MSE), root mean squared error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE) and highest coefficient of determination (R2). Results deduced that the best performing ANN model structure is 4-12-12-1 with the lowest overall MSE of 0.002898 and highest overall R2 of 0.955769.
SIMM2023 Toward Empowering Technological Transformation 78 7.Industrial Engineering
SIMM2023 Toward Empowering Technological Transformation 79 Paper ID: SIMM2023: 024-010 Title: Exploring the Link between Green Manufacturing and Financial and Social Performance: A Structural Equation Modeling Study. First Author: Siti Norhafiza Binti Abdul Razak University Kuala Lumpur Malaysia France Institute This research paper explores the relationship between green manufacturing and both financial performance and social performance. The study uses a quantitative methodology, specifically structural equation modeling, to analyze the data. The survey questionnaire used in this study was designed to investigate the attitudes and perceptions of the employees in the electric and electronics sectors towards sustainable practices. The results suggest that green manufacturing practices have a positive impact on both financial performance and social performance. This highlights the importance of companies implementing green manufacturing strategies, not only for financial gain but also for the benefits they bring to society and the environment. The findings of this study contribute to the growing body of literature on the relationship between sustainability and business performance and can inform decisions made by companies and policymakers.
SIMM2023 Toward Empowering Technological Transformation 80 Paper ID: SIMM2023: 050-038 Title: TRIZ Patented Literature Review on Automated Guided Vehicle Technology for Systematic Innovation. First Author: Zulhasni Bin Abdul Rahim Co-Author: Muhammad Saqib Iqbal Universiti Teknologi Malaysia AGVs offer many benefits in terms of increased efficiency, reduced labor costs, and improved safety, they also have some limitations that should be carefully considered before implementing them in an industrial setting. Therefore, it is critical to improve the level of technology impact through promoting the development of invention and secured as patent. From the academic perspective, Systematic literature review (SLR) can include pa-tented literature, but there are some limitations and challenges associated with the inclusion of such literature. The concept of Patented literature re-view (PLR) can be considered a specialised form of SLR, with patents being a unique and valuable source of information in many fields. Furthermore, this study incorporates the concepts of TRIZ (Theory of Inventive Problem Solving) in the PLR concept to support innovation performance. TRIZ pro-vides a model for understanding the evolution of technological systems, and for identifying patterns and contradictions that can be used to generate innovative solutions. A case study of AGV is presented by using the pro-pose TRIZ-PLR model that involves defining the aim of the study, systematically reviewing related patents, and establishing the results of the study. The use of TRIZ in PLR can help to promote innovation by providing a systematic and structured approach to analysing patent literature and identifying opportunities for improvement and new ideas.
SIMM2023 Toward Empowering Technological Transformation 81 Paper ID: SIMM2023: 059-051 Title: Optimization of Injection Molding Process Parameters for Carabineer Lock using the Taguchi Method. First Author: Priscilia Annie Anak Eddy Co-Author: C.Y. Khor; M.U. Rosli; Muhammad Ikman Ishak Faculty of Mechanical Engineering & Technology, Universiti Malaysia Perlis, 02600 Arau, Perlis, Malaysia The injection molding process is critical to manufacturing various plastic products, including carabineer locks. A simulation-based optimization approach is conducted in this study using the Taguchi method to achieve optimal process parameters for carabineer lock production. The Taguchi meth-od identified the key factors that affect the process output, which can be used to optimize the process and minimize defects. Main effect analysis and analysis of variance are also included in this study. The results revealed that shrinkage is significantly affected by packing pressure (E) and melt temperature (B). However, the most influential factors for warpage are packing pressure (E), melt temperature (B) and cooling time (A). The verification test showed that the optimized factor and level reduced the shrinkage and warp-age in the injection molding. The results from the simulation-based optimization are expected to guide the selection of appropriate parameter settings for the injection molding process, leading to improved product quality and reduced manufacturing costs.
SIMM2023 Toward Empowering Technological Transformation 82 Paper ID: SIMM2023: 082-073 Title: Sustainable Manufacturing Through Materials Machining Innovations - A review. First Author: Ainur Munira Rosli Faculty of Manufacturing and Mechatronic Engineering Technology, Universiti Malaysia Pa-hang, 26600, Pekan, Pahang, Malaysia Sustainable manufacturing practices are essential to reduce the environmental impact of manufacturing operations, increase efficiency, and improve social responsibility. Materials machining, which is a key process in manufacturing, plays a crucial role in sustainable manufacturing practices. The review explores sustainable materials machining innovations and their potential to pro-mote sustainable manufacturing practices. Sustainable materials machining techniques includes green machining, dry machining, and cryogenic machining, among others. The advantages and disadvantages of each technique are shown. The review also covers sustainable materials selection, including criteria for sustainable materials selection, such as environmental, social, and economic impact, and methods for sustainable materials selection, such as life cycle assessment and sustainable material databases. Case studies demonstrating the implementation of sustainable machining processes are also included. Furthermore, the review explores sustainable manufacturing supply chain management, including green logistics, sustainable packaging, and other sustain-able practices. The implications for sustainable manufacturing practices and a call to action for the implementation of sustainable materials machining innovations are also discussed. In conclusion, potential of sustainable materials machining innovations to promote sustainable manufacturing practices and encourages stakeholders to take collective action to implement these practices.
SIMM2023 Toward Empowering Technological Transformation 83 Paper ID: SIMM2023: 058-078 Title: Manufacturing Changeover Time Reduction for Dry Etching Tools in Wafer Fabrication Industry Using Single-Minute-Exchange-of-Dies (SMED). First Author: Tan Chan Sin Co-Author: Chong Chao Ming; Nur Aishah Mohd Sahar; Ahmad Humaizi bin Hilmi; Rosmaini Ahmad; Shaliza Azreen Mustafa Advanced Manufacturing System Research Group, Faculty of Mechanical Engineering Technology, Universiti Malaysia Perlis, Pauh Putra, 02600, Arau, Perlis, Malaysia The thesis presents on how integration and utilization of Single-MinuteExchange-Dies (SMED) helps to reduce dry etching tools changeover time in wafer fabrication industry through optimization of process changeover planning as previously encountered inconsistent output due to numerous changeover activities performed in the factory. Performance of etching tools are analysed using live data collection according to number of usage as well as availability rate together with changeover element identification using X-site software. The data will be then used to create foundation according to SMED techniques, starting with clearly defining internal and external elements along with turning internal element into external element. By comparing the results before and after SMED implementation, successful rate of the implementation on three main etching tools used in dry etching process will be validated. From the result, positive outcome was shown as SMED system able to helps to re-duce dry etching tools changeover time about 42% - 62%. Tool availability rate and utilization also shown improvement as it increased to about 3% to 6.6% and 8% to 12% respectively.
SIMM2023 Toward Empowering Technological Transformation 84 Paper ID: SIMM2023: 085-084 Title: Manual Handling Risk Assessment using Manual Handling Chart (MAC) Among Mechanics at Tyre Service Center. First Author: Mazlina binti Kamarudzaman Co-Author: Mohd Nasrull bin Abdol Rahman ([email protected]); Nurul Elyna Asyiqen binti Mohd Zaki ([email protected]) Department of Manufacturing Engineering, Faculty of Manufacturing & Mechanical Engineering Universiti Tun Hussein Onn Malaysia, Batu Pahat, Johor, Malaysia MSDs are significant contributors to work-related disability and time lost caused by illness. As manual material handlers, mechanics are frequently exposed to the physical risk factors for MSDs. The objective of this study was to investigate the relationship between the Manual Handling Risk Assessment (MAC) tool and the Nordic musculoskeletal questionnaire (NMQ) among mechanics at tyre service centers. This study was conducted in Taiping, Perak. There are 239 mechanics from various workshops who participated in a cross-sectional survey using questionnaires, focusing mostly on lifting, carrying, and team-handling activities. Data was collected using a structured questionnaire and interview session which are Nordic Musculoskeletal Questionnaire (NMQ) and Manual Handling Chart (MAC) tools. NMQ was used to evaluate the exposure of the activities and related physical risk factors that had correlated to one of the region's bodies within a 12-month period while MAC is required to evaluate common risk factors of lifting, carrying, and team handling with a collection of 11 risk factors mainly. In this research, most mechanics have suffered musculoskeletal symptoms in the low back (64.4%), shoulder (54.0%), and neck pain (46.8%). According to a MAC analysis, most mechanics are at a medium risk of developing musculoskeletal symptoms as a result of their lifting activity (56.49%), carrying activity (46.03%) team handling activity (30.0%). Several definitions of pain showed a strong correlation (p).
SIMM2023 Toward Empowering Technological Transformation 85 8.Intelligent Manufacturing
SIMM2023 Toward Empowering Technological Transformation 86 Paper ID: SIMM2023: 031-016 Title: Application of the Bees Algorithm to Find Optimal Drill Path Sequence. First Author: Shafie Kamaruddin Co-Author: Muhammad Harith bin Zainal Abidin ([email protected]); Nor Aiman Sukindar ([email protected]); Afiqah Adam Malek ([email protected]) Department of Manufacturing and Materials Engineering, Faculty of Engineering, International Islamic University Malaysia, 53100 Jalan Gombak, Selangor, Malaysia Optimisation is about finding the best solution to a particular problem. It is applied in many fields, especially in engineering problems. Drilling holes is a machining process that uses a tool with a pointed end or cutting edges to create circular holes in a material. One of the most common problems with drilling multiple holes is machining time. When drilling multiple holes, 70% of the machining time is spent moving and changing tools. Therefore, optimising the drilling sequence is important to reduce the machining time and increase the production of the company. Therefore, this study is conducted to find the optimal drilling sequence of multiple holes. The Bees Algorithm is applied to find the optimal drilling sequence for some benchmark problems including the 5x5, 7x7, and 9x9 array matrices of holes. The Bees Algorithm was run using R Software. The results found are compared with the results of other algorithms in terms of the drill path length and machining time. The main finding of the study is that the Bees Algorithm found optimal drill path length and minimum machining time comparable to the results of the other algorithms for the 5x5, 7x7 and 9x9 problems. These results show that the Bees Algorithm can be an alternative approach to find the optimal drilling sequence.
SIMM2023 Toward Empowering Technological Transformation 87 Paper ID: SIMM2023: 033-018 Title: Product Design and Development of an Endotracheal Tube Connector (ETC) for medical Intubation using Fused Deposition Modeling (FDM). First Author: WAY Yusoff Co-Author: Nazirul Muzzamil The rapid prototyping process has undergone outstanding advancement. Recently, many sectors, such as the health care sector, have utilized rapid prototyping in medical prototyping. They have also used it to produce low-volume production parts. This paper aims to show how rapid prototyping can be used in the medical field by redesigning a prototype for an endotracheal tube connector. A survey was conducted. The purpose of this survey is to identify the common problem of ETT and to gain some ideas for redesigning the task of new ETT connectors. The prototypes were designed based on Quality Function Deployment (QFD). The basic specifications of the new design of ETT parts were determined. After analysis of the user’s needs and requirements, concept designs were presented. The Pugh Chart and the concept scoring matrix were used to evaluate the designs. Fused Deposition Modeling (FDM) was used to fabricate the prototypes.
SIMM2023 Toward Empowering Technological Transformation 88 Paper ID: SIMM2023: 056-042 Title: The Patented Technology Innovation Portfolio on Remanufacturing in Circular Economy Using TRIZ. First Author: Muhammad Saqib Iqbal Malaysia-Japan International Institute of Technology, Universiti Teknologi Malaysia, Kuala Lumpur, Malaysia Many have proposed a "circular economy" to sustain economic expansion without damaging ecosystems or nonrenewable resources. A product's or a component's useful life can be prolonged by remanufacturing. Though it positively affects the economy, the environment, and society, remanufacturing needs to be improved by supply, production time, and quality issues. Also, a standard reference list can only include a limited number of patent-ed materials, but the literature can still be observed. This research combines the PLR method with TRIZ (Theory of Inventive Problem Solving). To come up with new ideas, TRIZ analyses the development of technical innovations for trends and discrepancies. By systematically reviewing the patent library for fresh ideas and enhancements, TRIZ in PLR can stimulate originality.
SIMM2023 Toward Empowering Technological Transformation 89 Paper ID: SIMM2023: 054-055 Title: An Improved LSTM Text Classification Model for Factory Report. First Author: Nurul Hannah Mohd Yusof Co-Author: Nurul Adilla Mohd Subha UNIVERSITI TEKNOLOGI MALAYSIA In text classification, preprocessing is a crucial stage. Using text preprocessing techniques, several word forms are condensed into a single form. Moreover, text preprocessing methods are given a great deal of significance and are extensively researched in machine learning. Preprocessing features, extracting significant characteristics, and comparing them to features in a database are the fundamental steps in text classification. Most studies considered three main steps; tokenization, stemming, stop words removal. However, stemmed words can sometimes be out of context and affect the accuracy of the training model. This paper investigates the effect of the preprocessing tasks on English manufacturing report for text classification using LSTM network due to its ability to classify sequential da-ta with memory cells. Phase I shows that there is a significant increment on aver-age 3.12% to 3.13% of accuracy on every preprocessing tasks added, while in Phase II, the accuracy difference between lemmatization and stemming demonstrates for 5.21%. Lemmatization outperforms stemming due to importance of context in the factory report data set.
SIMM2023 Toward Empowering Technological Transformation 90 Paper ID: SIMM2023: 038-082 Title: 3D print: Stringing and Warping Detection using MobileNet-SSD. First Author: Safwan Hisham Co-Author: Shah Fenner Khan ([email protected]); Kamarulzaman Kamarudin ([email protected]) Universiti Malaysia Perlis Defect detection is an essential process in additive manufacturing, especially for 3D printing. This feature can potentially help in quality control, production efficiency, waste product and cost reduction. In this paper, we propose a deep learning-based approach using the MobileNet-SSD algorithm for detecting and differentiating faults on 3D artefacts during printing. The model was trained using Google Colab and implemented in a Raspberry Pi 4 for real-time detection. The dataset used was added after the initial training, allowing the model to be retrained to improve its accuracy. The research flow includes selecting the best deep learning algorithm for defect detection, data collection, creating a neural network model, image pre-processing, hyperparameter tuning, and model training and validation. The accuracy of the proposed model was evaluated, and the results achieved a mean average precesion (mAP) of 28%. The proposed approach is effective in detecting defects compare to ResNet-SSD with mAP 21.4%.
SIMM2023 Toward Empowering Technological Transformation 91 5th International Symposium on Intelligent Manufacturing &Mechatronics (SIMM2023)