R. Srimathi
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
R. Srimathi has a total of 23 co-authors. Here are the top co-authors by number of shared publications:
- Hemamalini Sathish
VIT University
3 shared publications K. Suresh
Christ University
1 shared publicationK. V. Gunalan
Elcot SEZ
1 shared publicationKishore Eswaran
Vellore Institute of Technology
1 shared publication- Kishore Eswaran
Developer, KPIT, Bangalore, Karnataka, India
1 shared publication - Meenakshi Jayaraman
School of Electrical Engineering Vellore Institute of Technology Chennai Tamil Nadu India
1 shared publication
Latest Publications
- Open Access
- 9 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 29018-29047
Data-Driven Load Prediction for Hospitals: Evaluating ML Models for Energy Management
Hospitals need reliable long-term electrical load forecasting and energy management to maintain an unlimited power supply, avoid power outages that put the lives of patients at risk, and manage energy in critical healthcare facilities. Conventional statistical tools like ARIMA are unable to defend the nonlinearity and the highly dynamic nature of hospital energy consumption by HVAC systems, medical equipment, emergency cases, and 24/7 operations. In this work, seven machine learning and deep learning models are compared, namely Multi-Layer Perceptron (MLP), XG Boost, Random Forest (RF), Support Vector Regression (SVR), Long Short-Term Memory (LSTM), CNN (Convolutional Neural Network) -LSTM hybrid, and Transformer on one year of actual hourly hospital load data. The standard measures are used to evaluate the models: Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and coefficient of determination (R ${}^{2}$ ). Statistical analysis, convergence stability, the Friedman test, and post-hoc Wilcoxon signed-rank tests with Holm error correction all demonstrate significant performance differences. The transformer model has a better ability to capture long-range dependencies through self-attention, and the overall predictive accuracy and stability of the model are improved. The LSTM and MLP models have a close-to-similar performance, and ensemble and hybrid models are competitive. The suggested solution will allow planning energy more efficiently, which can potentially save 10-15 percent of energy use and improve the ability of healthcare facilities to overcome power outages.
- 9 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 29018-29047
Open AccessData-Driven Load Prediction for Hospitals: Evaluating ML Models for Energy Management
Hospitals need reliable long-term electrical load forecasting and energy management to maintain an unlimited power supply, avoid power outages that put the lives of patients at risk, and manage energy in critical healthcare facilities. Conventional statistical tools like ARIMA are unable to defend the nonlinearity and the highly dynamic nature of hospital energy consumption by HVAC systems, medical equipment, emergency cases, and 24/7 operations. In this work, seven machine learning and deep learning models are compared, namely Multi-Layer Perceptron (MLP), XG Boost, Random Forest (RF), Support Vector Regression (SVR), Long Short-Term Memory (LSTM), CNN (Convolutional Neural Network) -LSTM hybrid, and Transformer on one year of actual hourly hospital load data. The standard measures are used to evaluate the models: Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and coefficient of determination (R ${}^{2}$ ). Statistical analysis, convergence stability, the Friedman test, and post-hoc Wilcoxon signed-rank tests with Holm error correction all demonstrate significant performance differences. The transformer model has a better ability to capture long-range dependencies through self-attention, and the overall predictive accuracy and stability of the model are improved. The LSTM and MLP models have a close-to-similar performance, and ensemble and hybrid models are competitive. The suggested solution will allow planning energy more efficiently, which can potentially save 10-15 percent of energy use and improve the ability of healthcare facilities to overcome power outages.
- 22 April 2025
- Published by Springer Nature
- p. 247-257
Quasi Z-Source Inverter with Simple Boost and Maximum Boost Pulse Width Modulation Techniques for PV Grid Connection
- 22 April 2025
- Published by Springer Nature
- p. 247-257
Quasi Z-Source Inverter with Simple Boost and Maximum Boost Pulse Width Modulation Techniques for PV Grid Connection
- 19 May 2023
- Published by Springer Nature
- p. 273-288
Enhanced ATM Security Using Facial Recognition, Fingerprint Authentication, and WEB Application
- 19 May 2023
- Published by Springer Nature
- p. 273-288
Enhanced ATM Security Using Facial Recognition, Fingerprint Authentication, and WEB Application
- Open Access
- 28 September 2022
- Published by Hindawi Limited in International Journal of Photoenergy
- Vol. 2022, 1-13
Performance Evaluation and Estimation of Energy Measures of Grid-Connected PV Module
In this paper, the effectiveness of two grid-connected photovoltaic (PV) techniques up of copper indium selenium (CIS) and monocrystalline silicon (m-Si) arrays has been examined. In order to determine whether the technology is suitable for the actual winter and summer climatic conditions in Thoothukudi, Tamil Nadu, the observed and calculated performances have been compared. The final yield, photovoltaic (PV) effectiveness, array yield, performance ratio, and capacity utilisation factor seem to be the variables used to evaluate performance. Using recorded meteorological data at the selected location, PVsyst software predicts both PV systems’ year-round performances. These predictions are then contrasted to the outcomes of the actual measurements. The outcome showed that with a maximal observed performance ratio, both PV systems function marginally better in the winters than those in the summers. The performance indicators of the PV mechanisms are contrary with those of other PV systems with comparable capacities that are located in different places.
- 28 September 2022
- Published by Hindawi Limited in International Journal of Photoenergy
- Vol. 2022, 1-13
Open AccessPerformance Evaluation and Estimation of Energy Measures of Grid-Connected PV Module
In this paper, the effectiveness of two grid-connected photovoltaic (PV) techniques up of copper indium selenium (CIS) and monocrystalline silicon (m-Si) arrays has been examined. In order to determine whether the technology is suitable for the actual winter and summer climatic conditions in Thoothukudi, Tamil Nadu, the observed and calculated performances have been compared. The final yield, photovoltaic (PV) effectiveness, array yield, performance ratio, and capacity utilisation factor seem to be the variables used to evaluate performance. Using recorded meteorological data at the selected location, PVsyst software predicts both PV systems’ year-round performances. These predictions are then contrasted to the outcomes of the actual measurements. The outcome showed that with a maximal observed performance ratio, both PV systems function marginally better in the winters than those in the summers. The performance indicators of the PV mechanisms are contrary with those of other PV systems with comparable capacities that are located in different places.
- Open Access
- 1 November 2020
- Published by IOP Publishing in IOP Conference Series: Materials Science and Engineering
- Vol. 995 (1) , 012003
Iterative Learning Control for a Repetitive 2D Process
To control the systems which operate in a repetitive mode is known to be a difficult task. For example, systems like robotic manipulators are used for doing the repetitive assigned task. This can be overcome by the iterative learning control (ILC) method. In this article, the DC motor model is discussed for controlling the speed of the motor by the design of an ILC algorithm, and its performance is compared with the conventional PID algorithm. The ILC algorithm is implemented for the minimum and non-minimum phase systems. The resulting response, absolute and integral absolute errors are plotted. The Z-N tuning method is used to implement PID control. The performance of the DC motor has been evaluated by comparing the Integral Square Error (ISE), Integral Absolute Error (IAE), and Integral Time Absolute Error (ITAE) with the conventional PID controller.
- 1 November 2020
- Published by IOP Publishing in IOP Conference Series: Materials Science and Engineering
- Vol. 995 (1) , 012003
Open AccessIterative Learning Control for a Repetitive 2D Process
To control the systems which operate in a repetitive mode is known to be a difficult task. For example, systems like robotic manipulators are used for doing the repetitive assigned task. This can be overcome by the iterative learning control (ILC) method. In this article, the DC motor model is discussed for controlling the speed of the motor by the design of an ILC algorithm, and its performance is compared with the conventional PID algorithm. The ILC algorithm is implemented for the minimum and non-minimum phase systems. The resulting response, absolute and integral absolute errors are plotted. The Z-N tuning method is used to implement PID control. The performance of the DC motor has been evaluated by comparing the Integral Square Error (ISE), Integral Absolute Error (IAE), and Integral Time Absolute Error (ITAE) with the conventional PID controller.
- 23 September 2020
- Published by Springer Nature
- p. 337-351
Lighting Design for a Campus Hostel Building Using Single-Stage Efficient Buck LED Driver
- 23 September 2020
- Published by Springer Nature
- p. 337-351
Lighting Design for a Campus Hostel Building Using Single-Stage Efficient Buck LED Driver
- 5 November 2019
- Published by Taylor & Francis in IETE Journal of Research
- Vol. 68 (3) , 1943-1955
Performance Analysis of Single-Stage LED Buck Driver Topologies for Low-Voltage DC Distribution Systems
The demand for lean protein is the main reason for raising domesticated birds in poultry farms for the purpose of farming meat or eggs for food. In these poultry farms, artificial lighting system is an essential factor for the success of the commercial production of egg layers and broilers. Lighting loads connected to the existing AC system requires DC power. Nearly 25% of the total power consumed in India is by lighting loads. Therefore, low-voltage DC distribution system is the new requirement of electrical network to improve energy conservation. In this paper, an energy efficient single-stage LED driver topology is proposed as a better alternative to fluorescent lights used in poultry farms. The proposed topology is a phase-shifted buck converter with reduced size of passive components. The converter is mathematically modelled and designed for a power of 24 W LED lights. An experimental prototype is developed for the proposed converter and the results are validated in terms of cost, energy savings, and energy efficiency with various other LED driver topologies like conventional buck converter and interleaved buck converter. Further to justify the outcome of the proposed LED driver, a case study is also done for poultry farms.
- 5 November 2019
- Published by Taylor & Francis in IETE Journal of Research
- Vol. 68 (3) , 1943-1955
Performance Analysis of Single-Stage LED Buck Driver Topologies for Low-Voltage DC Distribution Systems
The demand for lean protein is the main reason for raising domesticated birds in poultry farms for the purpose of farming meat or eggs for food. In these poultry farms, artificial lighting system is an essential factor for the success of the commercial production of egg layers and broilers. Lighting loads connected to the existing AC system requires DC power. Nearly 25% of the total power consumed in India is by lighting loads. Therefore, low-voltage DC distribution system is the new requirement of electrical network to improve energy conservation. In this paper, an energy efficient single-stage LED driver topology is proposed as a better alternative to fluorescent lights used in poultry farms. The proposed topology is a phase-shifted buck converter with reduced size of passive components. The converter is mathematically modelled and designed for a power of 24 W LED lights. An experimental prototype is developed for the proposed converter and the results are validated in terms of cost, energy savings, and energy efficiency with various other LED driver topologies like conventional buck converter and interleaved buck converter. Further to justify the outcome of the proposed LED driver, a case study is also done for poultry farms.
- 15 June 2018
- Published by Taylor & Francis in Electric Power Components and Systems
- Vol. 46 (10) , 1134-1146
LED Boost Driver Topologies for Low Voltage DC Distribution Systems in Smart Structured Buildings
Most of the loads connected to the existing AC distribution system operate with low voltage DC power. Hence, there is a requirement to convert AC into DC power. Often there are many inefficiencies associated with AC to DC power conversion. To enhance the energy conservation, low voltage DC distribution system (LVDCDS) is suggested in this paper. Forty percent of the total power consumed in India is by lighting loads which make use of DC power. LVDCDS is an ideal choice to power lighting loads in residential and commercial buildings. This paper presents a detailed study of four different single stage, interleaved, non-isolated light emitting diode (LED) drivers for lighting load. All the topologies are developed in hardware and tested experimentally with controller implementation to provide a constant power of 70 W to power LED lights in the campus building. The installation cost, payback period, and energy savings for the converters connected to LVDCDS are analyzed and discussed in particular.
- 15 June 2018
- Published by Taylor & Francis in Electric Power Components and Systems
- Vol. 46 (10) , 1134-1146
LED Boost Driver Topologies for Low Voltage DC Distribution Systems in Smart Structured Buildings
Most of the loads connected to the existing AC distribution system operate with low voltage DC power. Hence, there is a requirement to convert AC into DC power. Often there are many inefficiencies associated with AC to DC power conversion. To enhance the energy conservation, low voltage DC distribution system (LVDCDS) is suggested in this paper. Forty percent of the total power consumed in India is by lighting loads which make use of DC power. LVDCDS is an ideal choice to power lighting loads in residential and commercial buildings. This paper presents a detailed study of four different single stage, interleaved, non-isolated light emitting diode (LED) drivers for lighting load. All the topologies are developed in hardware and tested experimentally with controller implementation to provide a constant power of 70 W to power LED lights in the campus building. The installation cost, payback period, and energy savings for the converters connected to LVDCDS are analyzed and discussed in particular.
- Open Access
- 1 June 2017
- Published by Elsevier in Energy Procedia
- Vol. 117, 224-235
High Efficiency Buck LED Driver Using SiC
- 1 June 2017
- Published by Elsevier in Energy Procedia
- Vol. 117, 224-235
Open AccessHigh Efficiency Buck LED Driver Using SiC