IOT BASED ONLINE LOAD FORECASTING OF ELECTRICAL LABORATRIES LOADS USING MACHINE LEARNING APPROACH A PROPOSED SCHEME
Deficient load management is major concern in our country as it leads to the load shedding and poor planning to the entire power systems. In this project, the load of six labs of the superior university is taken into account in a data set which trained in a machine learning (ML) algorithm. This proj
2025-06-28 16:28:11 - Adil Khan
IOT BASED ONLINE LOAD FORECASTING OF ELECTRICAL LABORATRIES LOADS USING MACHINE LEARNING APPROACH A PROPOSED SCHEME
Project Area of Specialization Electrical/Electronic EngineeringProject SummaryDeficient load management is major concern in our country as it leads to the load shedding and poor planning to the entire power systems. In this project, the load of six labs of the superior university is taken into account in a data set which trained in a machine learning (ML) algorithm. This project has the opportunity of STLF online with accurate prediction models by using ML algorithms. In the university labs, electrical load consumption data is implemented from which the different techniques of ML are trained. By using cloud computing and machine learning auto regression forecasting models are developed. Autoregression has less information required and self-variable series can be used to forecast the data. It also forecasts any recurring patterns in the data. Online forecasting is more refined and effective because of its ability to use the recent data logs for training and forecasting data online. So source selections are derived to assess the electricity consumption. The IOT based project module consists of a Led display which shows the forecasted and run-time values of load. So the load is managed according to the requirement of the labs and future expansion of university labs is also taken easily. Effective university lab load forecasts can help to improve and properly plan the load management of the entire unit. Accurate models for electric power load forecasting is proposed for the operation and planning of a university labs load. Similarly this can be extended in future for load management of residential, commercial and industrial units in our country.
Project Objectives- The main objective of our project is load management and planning of the power generation networks. Ultimately it overcomes the load shedding effects and deficient planning for the power houses.
- University lab load expansion planning starts with a forecast of anticipated future load requirements. The accurate estimation of demand and power for the university is very effective in system planning.
- This model is also used to establish procurement policies for the construction of new labs, and forecasting of capital energy.
- Our model is designed for Less Power consumption by circuit components.
- It plays a vital role in load switching and infrastructure development of the university labs.
Design parameters and combinational forecasting model based on machine learning autoregression (AR). In this project, a time collection and autoregression algorithm for analyzing the greater complex and large statistics. The time series AR model is in addition verified utilizing the ML-based remarks in which 5 levels are computed. Methods for achieving the pleasant feature choice are evaluated and attained greater optimized consequences. A remarkable improvement in MAPE (Mean Absolute Percentage Error) and average MAPE have been performed utilizing a proposed integration strategy which gives a 20% annual improvement in comparison to previous ones. This version can optimize the performance of university labs by predicting the optimized STLF and can triumph over the issues related to the planning and running of university labs. Simulation methods on MATLAB software are approached. The project is simply integrated with Raspberry Pi for machine learning and sending data to the cloud.
Benefits of the Project- Through this project, accurate load forecasting can ensure the economics of power system operation, and also contribute to the development of the university economy. As an important part of power planning, power load forecasting affects the planning operation mode and power exchange plan to manage the university labs load.
- In addition, as a basis for assisting power planning and construction, load forecasting can also play a role in improving energy efficiency and reducing power generation costs. Moreover, it is also possible to implement power system operations of the university more economically.
- In this project, basically, we forecast the university lab’s power load this will help us plan and manage the load in a better way. The university lab load is essentially uncontrollable although minor variations can be affected by frequency control and more drastically by load shedding.
- The variation in university lab load does exhibit certain daily and monthly pattern repetitions and analysis of these forms the basis of several load prediction techniques. This concern can be overcome by using our model.
- Wastage due to misplanning like the purchase of equipment in university labs, which is not immediately required, can be avoided by implementing the proposed model.
The implementation methodology is to first understand the lab’s load, schedules, and what problems are faced during the whole process. The load forecasting project will be realistic and implementable for university labs’ load management. The six lab data is first collected according to the timetable of labs and this is a data set trained on raspberry by machine learning algorithm. The module has capacity for the ease of integration and machine learning on the Raspberry Pi. The forecasting module is a box of 10?
15inch. Which contain the CT(current transformer) and PT (potential transformer) for the real-time values of voltage and current. For the power sources selection (solar and conventional) relay is connected. To perform the data acquisition, Arduino Uno is used which displays the real-time reading on the GUI (graphic user interface), and then the system is integrated with the Raspberry Pi which can forecast data by machine learning and transfer it to the cloud. The ubidot is the free-of-cost platform for displaying the values and graphs anywhere on the cloud IOT based).
| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 78900 | |||
| Ardunio nano | Equipment | 1 | 1100 | 1100 |
| Raspberry Pi 4 | Equipment | 1 | 38000 | 38000 |
| Non-invasive current sensor | Equipment | 6 | 1200 | 7200 |
| ZMPT101 B | Equipment | 2 | 350 | 700 |
| LCD | Equipment | 1 | 500 | 500 |
| I2c lcd mode | Equipment | 2 | 150 | 300 |
| Memory Card | Equipment | 1 | 2000 | 2000 |
| Soldering Iron Kavya | Equipment | 1 | 1000 | 1000 |
| Digital Multimeter 33-B | Equipment | 1 | 1400 | 1400 |
| Clocking board | Equipment | 1 | 1500 | 1500 |
| LED HDMI | Equipment | 1 | 12000 | 12000 |
| Capacitor, Resistance, Vero Board, etc | Miscellaneous | 1 | 3200 | 3200 |
| Cloud unit | Equipment | 1 | 4000 | 4000 |
| Material required for t0he box making. | Miscellaneous | 1 | 6000 | 6000 |