SlideShare a Scribd company logo

Introduction to ANN Principles and its Applications in Solar Energy Technology

Download as PPTX, PDF
A
Ali Al-Waeli

This document provides an overview of artificial neural networks (ANN) and their applications. It begins with introducing general concepts in artificial intelligence, machine learning, and ANN. It then discusses different ANN architectures like multilayer perceptrons and backpropagation. It also covers ANN learning paradigms such as supervised, unsupervised, and reinforcement learning. Finally, it provides examples of using ANN for applications in fields like solar energy prediction, structural analysis, and control systems.

Engineering
1 of 103
Download to read offline
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
Department of Engineering American University of Iraq, Sulaimani (AUIS) 1
Contents 2 ARTIFICIAL NEURAL NETWORKS REGRESSION PERFORMANCE PREDICTION ANN FOR RESEARCH SAMPLE EXERCISE
3 ARTIFICIAL NEURAL NETWORKS
 GENERAL FRAMEWORK OF AI  ARTIFICIAL INTELLIGENCE  MACHINE LEARNING  ARTIFICIAL NEURAL NETWORKS  DEEP LEARNING 4 ARTIFICIAL NEURAL NETWORKS
5 Image Source: aiso-lab.com
WHAT – WHY - HOW 6
Artificial Intelligence 7
Artificial Intelligence  What is AI ? Branch of computer science that focuses on creating systems that are intelligent and independent.  Why use AI ? To enhances the speed, precision and effectiveness of human efforts.  How is AI implemented ? AI can be developed through learning that is symbolic-based or data-based. 8
Machine learning  What is Machine Learning ? Using learning paradigms in machines to achieve pattern recognition and other human functions.  Why use Machine Learning ? Machines can utilize more data, and dimensions of data, to perform functions such as pattern recognition.  How is Machine Learning developed ? Machines learn through either statistical learning or deep learning. 9
The State-of-AI 10 Text AI Visual AI Interactive AI Functional AI Analytic AI text recognition, speech-to-text conversion machine translation, and content generation capabilities. Chatbots and smart personal assistants. Computer vision or augmented reality Sentiment analysis and supplier risk assessment Maintenance of breakdown in machines after spotting errors, etc. Examples: Alex Bekker. 5 Types of AI to Propel Your Business. Science soft. Link: https://www.scnsoft.com/blog/artificial-intelligence-types
11 Source: Alex Bekker,ScienceSoft. link: https://www.scnsoft.com/blog/artificial-intelligence-types The State-of-AI
The State-of-AI 12 Symbolic learning Machine learning Pattern recognition Computer vision Neural Networks Deep Learning learning capabilities (or) reasoning capabilities What about combining the two ? Able to give explanations and can manipulate complex data structures Richa Bhatia. Understanding the difference between symbolic AI & non symbolic AI. https://analyticsindiamag.com/understanding-difference-symbolic-ai-non-symbolic-ai/
13 Symbolic learning Symbolic learning See Process Computer vision Robotics Interact Symbolic learning
14 Human ability to interact with the world System that can learn Symbolic learning Machine learning Pattern recognition Computer vision Neural Networks Deep Learning Machine learning
15 Robotic system Sensor Actuators Non-AI programming AI algorithm AI and Robotics
ARTIFICIAL NEURAL NETWORKS  What are Artificial Neural Networks ?  Why use Artificial Neural Networks ?  How are Artificial Neural Networks developed ? Answers in the presentation > slides 15-41 16
Deep Learning  What is Deep Learning ?  Why use Deep Learning ?  How is Deep Learning achieved ? Answers in the presentation > slide 42 17
Artificial Neural Networks (ANN) C 18
Artificial Neural Networks (ANN) Motivation 19
Artificial Neural Networks (ANN) Advantages 20
Artificial Neural Networks (ANN) Components 21
Artificial Neural Networks (ANN) C 22
Artificial Neural Networks (ANN) C 23
Artificial Neural Networks (ANN) C 24
Artificial Neural Networks (ANN) Numerical Data Labeling Images Video Text Time-series Clustering Machine perception Classification Clustering C 25
Artificial Neural Networks (ANN) Classification SUPERVISED LEARNING Face detection Object Identification Text classification Voice detection Gesture Identification Transcript speech-text Create Numerical Data Assign to data TRAIN THE ANN Create correlation 26
Artificial Neural Networks (ANN) Ni, D. X. (2007). Application of neural networks to character recognition. Proceedings of students/faculty research day, CSIS, Pace University, May 4th. Classification C 27
Artificial Neural Networks (ANN) Classification https://www.analyticsindiamag.com/how-to-create-your-first-artificial-neural-network-in-python/ C 28
Artificial Neural Networks (ANN) Regression SUPERVISED LEARNING Hardware breakdown Employee turnover Customer churn Health breakdown Machine output C 29
Artificial Neural Networks (ANN) Regression https://dataaspirant.com/2017/03/02/how-logistic-regression-model-works/ C 30
Artificial Neural Networks (ANN) Regression https://dataaspirant.com/2017/03/02/how-logistic-regression-model-works/ C 31
Artificial Neural Networks (ANN) Regression https://dataaspirant.com/2017/03/02/how-logistic-regression-model-works/ C 32
Artificial Neural Networks (ANN) Clustering UNSUPERVISED LEARNING Comparing documents, images, etc. to surface similar items Detecting anomalies, or unusual behavior C 33
Artificial Neural Networks (ANN) Clustering C 34
Artificial Neural Networks (ANN) 35 https://quantumcomputingtech.blogspot.com/2019/02/machine-learning-animated-gif.html Clustering
Artificial Neural Networks (ANN) REINFORCEMENT Self-driving car Customer management Adaptive traffic signal control Games Reinforcement LEARNING C 36
37 Artificial Neural Networks (ANN) REINFORCEMENT This paradigm does not require a labelled input/output pairs to be presented and does not need sub-optimal actions to be explicitly corrected.
Artificial Neural Networks (ANN) Learning paradigms SUPERVISED LEARNING UNSUPERVISED LEARNING REINFORCEMENT LEARNING 38
Artificial Neural Networks (ANN) Learning paradigms Kaplan, S. (2017). Deep generative models for synthetic retinal image generation. C 39
Artificial Neural Networks (ANN) Learning paradigms Unsupervised learning Supervised learning Clustering: An Introduction to Supervised Machine Learning and Pattern Classification: The Big Picture by Sebastian Raschka C 40
Artificial Neural Networks (ANN) SINGLE LAYER C 41
Artificial Neural Networks (ANN) MULTIPLE LAYER PERCEPTRON (MLP) C 42
Artificial Neural Networks (ANN) MULTIPLE LAYER PERCEPTRON (MLP) C 43
Artificial Neural Networks (ANN) MULTIPLE LAYER PERCEPTRON (MLP) C 44
Artificial Neural Networks (ANN) MULTIPLE LAYER PERCEPTRON (MLP) Jain & Mao, 1996 C Deep 45
Artificial Neural Networks (ANN) MULTIPLE LAYER PERCEPTRON (MLP) A MLP consists of at least three layers of nodes: an input layer, a hidden layer and an output layer. Except for the input nodes, each node is a neuron that uses a nonlinear activation function. MLP utilizes a supervised learning technique called backpropagation for training. Its multiple layers and non-linear activation distinguish MLP from a linear perceptron. It can distinguish data that is not linearly separable. 46
Artificial Neural Networks (ANN) MULTIPLE LAYER PERCEPTRON (MLP) Changing connection weights after each piece of data is processed, based on the amount of error in the output compared to the expected result. 47
Artificial Neural Networks (ANN) MULTIPLE LAYER PERCEPTRON (MLP) MLPs are useful in research for their ability to solve problems stochastically, which often allows approximate solutions for extremely complex problems like fitness approximation. MLPs are universal function approximators as shown by Cybenko's theorem, so they can be used to create mathematical models by regression analysis. As classification is a particular case of regression when the response variable is categorical, MLPs make good classifier algorithms. MLPs were a popular machine learning solution in the 1980s, finding applications in diverse fields such as speech recognition, image recognition, and machine translation software, but thereafter faced strong competition from much simpler (and related) support vector machines. Interest in backpropagation networks returned due to the successes of deep learning. 48
Artificial Neural Networks (ANN) MULTIPLE LAYER PERCEPTRON (MLP) 49
Artificial Neural Networks (ANN) FEED-FORWARD Connections between the nodes do not form a cycle. The first and simplest type of artificial neural network devised. 50
Artificial Neural Networks (ANN) BACKPROPAGATION 51
Artificial Neural Networks (ANN) BACKPROPAGATION 52
Artificial Neural Networks (ANN) BACKPROPAGATION  Supervised learning  Used at each layer to minimize the error between the layer’s response and the actual data  The error at each hidden layer is an average of the evaluated error 53
Artificial Neural Networks (ANN) BACKPROPAGATION  N is a neuron.  Nw is one of N’s inputs weights  Nout is N’s output.  Nw = Nw +Δ Nw  Δ Nw = Nout * (1‐ Nout)* NErrorFactor  NErrorFactor = NExpectedOutput – NActualOutput  This works only for the last layer, as we can know the actual output, and the expected output. 54
Artificial Neural Networks (ANN) Training stage Validating stage Testing stage 55
Contents 56 ARTIFICIAL NEURAL NETWORKS REGRESSION PERFORMANCE PREDICTION ANN FOR RESEARCH SAMPLE EXERCISE
57 REGRESSION A measure of the relation between the mean value of one variable (e.g. output) and corresponding values of other variables (e.g. time and cost). Dependent variable Independent variables
58 REGRESSION Dependent variable Independent variables Control variables
Regression and Correlation 59
Contents 60 ARTIFICIAL NEURAL NETWORKS REGRESSION PERFORMANCE PREDICTION ANN FOR RESEARCH SAMPLE EXERCISE
PERFORMANCE PREDICTION Methodology Performance Prediction Statistical methods Analysis 61
PV/T modeling and performance prediction • Why use artificial neural networks (ANN) in solar energy technologies? • How to use artificial neural networks (ANN) in solar energy technologies? • How to utilize regression and classification to invest in solar energy? 62
PV performance time-series prediction Short-term power forecasting Estimating power loss due to environment Estimating the energetic performance of PV/T PV fault detection Optimize PV array inclination Solar Energy Forecasting Optimize maximum power point tracking 63
PV performance time-series prediction C Senkal, O. & Kuleli, T. (2009). Estimation of solar radiation over Turkey using artificial neural network and satellite data. Applied Energy, Vol. 86, pp. 1222– 64
PV performance time-series prediction Almonacid, F., Rus, C., Hontoria, L., Fuentes, M. & Nofuentes G. (2009). Characterisation of Si-crystalline PV modules by artificial neural Networks. Renewable Energy, C 65
PV performance time-series prediction Firefly algorithm (FA) and Particle Swarm Optimization (PSO) applied to train NN EXAMPLE 1: SHORT TERM POWER FORECASTING Demirdelen, T., Aksu, I. O., Esenboga, B., Aygul, K., Ekinci, F., & Bilgili, M. (2019). A new method for generating short-term power forecasting based on artificial neural networks and optimization methods for solar photovoltaic power plants. In Solar Photovoltaic Power Plants (pp. 165-189). Springer, Singapore. Multi-Layer Feed Forward (MLFF) neural network structure The data of 1MWPV power plant in Turkey is used to estimate output power by real-time data mining for short time prediction. 66
PV performance time-series prediction Example 1 C 67
PV performance time-series prediction Example 1 Ambient temperature [◦C] Solar radiation [W/m2] PV Panel temperature [◦C] Input Layer Hidden Layer Output Layer 24 8 Interconnections Interconnections PV power (W) 68
PV/T performance time-series prediction Particle Swarm Optimization (PSO) applied to train NN EXAMPLE 2: ESTIMATING ENERGETIC PERFORMANCE Alnaqi, A. A., Moayedi, H., Shahsavar, A., & Nguyen, T. K. (2019). Prediction of energetic performance of a building integrated photovoltaic/thermal system thorough artificial neural network and hybrid particle swarm optimization models. Energy conversion and management, 183, 137-148. Multi-Layer Feed-Forward Back-Propagation (FFBP) neural network structure building integrated photovoltaic/thermal system 69
PV/T performance time-series prediction Example 2 C 70
PV/T performance time-series prediction Example 2 C 71
Statistical methods • Coefficient of determination • Mean Absolute Percentage Error • Root Mean Square Error • Mean Absolute Error • Mean Square Error difference between two continuous variables how well a regression model is capable of describing a data set a measure of prediction accuracy of a forecasting method in statistics, a measure of the quality of an estimator measure of the differences between an estimator and the values observed 72
Statistical methods COEFFICIENT OF DETERMINATION a measure for how well a regression model is capable of describing a data set. used to evaluate the validity of the predictive results compared to the actual (experimental) real model results 𝒚𝒊= experimental value of (y) 𝒇𝒊= predicted value of (y) 𝒚𝒊= the mean of the experimental values N = the number of values 73
Statistical methods MEAN ABSOLUTE ERROR a measure of difference between two continuous variables. the average vertical distance between each point and the identity line. MAE is also the average horizontal distance between each point and the identity line. 74
Statistical methods MEAN ABSOLUTE PERCENTAGE ERROR a measure of prediction accuracy of a forecasting method in statistics, 75
Statistical methods MEAN SQUARE ERROR measures the average of the squares of the errors—that is, the average squared difference between the estimated values and what is estimated The MSE is a measure of the quality of an estimator—it is always non- negative, and values closer to zero are better. 76
Statistical methods ROOT MEAN SQUARE ERROR a measure for the difference between a data set and a corresponding fit. The RMS asymptotically converges to the standard deviation from the model's predicted value for sufficiently large sizes of data sets. measure of the differences between values (sample or population values) predicted by a model or an estimator and the values observed. 77
Applications in engineering disciplines Civil engineering 78 Suryanita, R. (2016). The application of artificial neural networks in predicting structural response of multistory building in the region of Sumatra island. KnE Engineering. Reinforced Concrete Building Models
Applications in engineering disciplines Electrical engineering 79 Hiyama, T., Kouzuma, S., Imakubo, T.:‘Identification of optimal operating pointof PV modules using neural network for real time maximum power trackingcontrol’,IEEE Trans. Energy Convers., 1995,10, (2), pp. 360–367
Applications in engineering disciplines Mechanical engineering 80 Ranasinghe, R. A. T. M., Jaksa, M. B., Kuo, Y. L., & Nejad, F. P. (2017). Application of artificial neural networks for predicting the impact of rolling dynamic compaction using dynamic cone penetrometer test results. Journal of Rock Mechanics and Geotechnical Engineering, 9(2), 340-349. Predicting the impact of rolling dynamic compaction using dynamic cone penetrometer test results
Applications in engineering disciplines Control engineering 81 KACHULKOVA, S. Applications of Artificial Neural Networks in Control Systems. transfer, 1(1), 5.
82 ARTIFICIAL NEURAL NETWORKS REGRESSION PERFORMANCE PREDICTION ANN FOR RESEARCH SAMPLE EXERCISE
Tips to start utilizing machine learning and ANN techniques for your previous, current and future studies in your field… PAST – PRESENT - FUTURE 83 ANN FOR RESEARCH
84 TIP #1 CLASSIFY THE VARIABLES DEPENDENT VARIABLES (DV) INDEPENDENT VARIABLES (IV) CONTROL VARIABLES (CV)
85 TIP #2 ENSURE HIGH-QUALITY OF DATA GIGO GARBAGE IN = GARBAGE OUT
GIGO, VIGO, VIVO 86
87 TIP #2 ENSURE HIGH-QUALITY OF DATA WHAT IS YOUR UNCERTAINTY THRESHOLD? WHAT MODEL ARE YOU USING FOR UNCERTAINTY ANALYSIS? WHAT IS YOUR UNCERTAINTY THRESHOLD?
88 Example of uncertainty model and analysis The uncertainty associated with the result is WR and those associated with the independent variables are W1, W2, …, Wn. Source: Kline, S. J. (1953). Describing uncertainty in single sample experiments. Mech. Engineering, 75, 3-8.
89 Example of uncertainty model and analysis Source: Sarafraz, M. M., Safaei, M. R., Leon, A. S., Tlili, I., Alkanhal, T. A., Tian, Z., ... & Arjomandi, M. (2019). Experimental investigation on thermal performance of a PV/T-PCM (photovoltaic/thermal) system cooling with a PCM and nanofluid. Energies, 12(13), 2572.
90 Data filtration Height of an adult male by fixing the typo or possibly remeasuring the item or person. If that’s not possible, you must delete the data point because you know it’s an incorrect value. Source: Statistics by Jim. Link: https://statisticsbyjim.com/basics/remove-outliers/ (access date: 25th of July 2022)
91 COLLECT AS MUCH DATA AS POSSIBLE. TIP #3 COLLECT BULK DATA Once testing station, or setup, is established. Record as much data as possible. With a variety of design configurations.
92 Number of input, outputs and hidden nodes. TIP #4 Select the components Utilize the identified variables (IV, DV, CV)
93 How much (%) of the data are you allocating for training? Validating? Testing? TIP #5 Choose the data splitting percentage
94 Use MATLAB/Python to create a simple MLP Neural Network to process the data TIP #6 Start with a simple MLP ANN
95 ARTIFICIAL NEURAL NETWORKS REGRESSION PERFORMANCE PREDICTION ANN FOR RESEARCH SAMPLE EXERCISE
96 Simple exercise Perform tasks 1-5
Food for Thought • What is the difference between Artificial Intelligence and computational intelligence? • How to interpret skewness and kurtosis? • What is the significance of the activation function? • What is the significance of the bias? 97
 Identify the dependent and independent variables in your work.  Assign the input and output nodes to your network.  Conduct data pre-processing.  Decide the type of problem you have.  Choose an activation function.  Choose the number of hidden nodes.  Assign weights and biases.  Download MATLAB and implement a simple MLP code.  Carryout the training, validation and testing. Homework 98
Textbook in PV/T https://www.amazon.com/Photovoltaic-Thermal-Systems-Principles- Applications/dp/3030278239/ref=sr_1_1?keywords=PV%2FT+al-waeli&qid=1563014343&s=gateway&sr=8-1 99
Review article in PV/T 100
Original article on ANN use for PV/T Performance Prediction 101
It’s play time… • https://playground.tensorflow.org/ 102
Thank you 103

More Related Content

What's hot (20)

PPTX
Optimization Simulated Annealing
Uday Wankar
 
PDF
NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis
taeseon ryu
 
PDF
CCS355 Neural Network & Deep Learning UNIT III notes and Question bank .pdf
Guru Nanak Technical Institutions
 
PDF
SSII2022 [SS1] ニューラル3D表現の最新動向〜 ニューラルネットでなんでも表せる？？ 〜​
SSII
 
PPTX
Graph Neural Networks
tm1966
 
PDF
An introduction to Machine Learning
butest
 
PPTX
Simulated Annealing
Joy Dutta
 
PPTX
Dilation and erosion
Aswin Pv
 
PDF
Reinforcement learning
DongHyun Kwak
 
PPT
rnn BASICS
Priyanka Reddy
 
PPT
hidden surface elimination using z buffer algorithm
rajivagarwal23dei
 
PPTX
Soft computing
Arvind sahu
 
PDF
Knowledge based agent
Shiwani Gupta
 
PDF
2-Approximation Vertex Cover
Kowshik Roy
 
PDF
Introduction to 3D Computer Vision and Differentiable Rendering
Preferred Networks
 
PPTX
CNN-SLAMざっくり
EndoYuuki
 
PPTX
Deep Learning A-Z™: Self Organizing Maps (SOM) - Reading an Advanced SOM
Kirill Eremenko
 
PPTX
FUZZY LOGIC
VanishriKornu
 
PPTX
Evolutionary computing - soft computing
SakshiMahto1
 
PDF
Simulated Annealing
kellison00
 
Optimization Simulated Annealing
Uday Wankar
 
NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis
taeseon ryu
 
CCS355 Neural Network & Deep Learning UNIT III notes and Question bank .pdf
Guru Nanak Technical Institutions
 
SSII2022 [SS1] ニューラル3D表現の最新動向〜 ニューラルネットでなんでも表せる？？ 〜​
SSII
 
Graph Neural Networks
tm1966
 
An introduction to Machine Learning
butest
 
Simulated Annealing
Joy Dutta
 
Dilation and erosion
Aswin Pv
 
Reinforcement learning
DongHyun Kwak
 
rnn BASICS
Priyanka Reddy
 
hidden surface elimination using z buffer algorithm
rajivagarwal23dei
 
Soft computing
Arvind sahu
 
Knowledge based agent
Shiwani Gupta
 
2-Approximation Vertex Cover
Kowshik Roy
 
Introduction to 3D Computer Vision and Differentiable Rendering
Preferred Networks
 
CNN-SLAMざっくり
EndoYuuki
 
Deep Learning A-Z™: Self Organizing Maps (SOM) - Reading an Advanced SOM
Kirill Eremenko
 
FUZZY LOGIC
VanishriKornu
 
Evolutionary computing - soft computing
SakshiMahto1
 
Simulated Annealing
kellison00
 

Similar to Introduction to ANN Principles and its Applications in Solar Energy Technology (20)

PDF
Neural networking this is about neural networks
v02527031
 
PDF
Artificial Neural Networking
shresthapragyan063
 
DOCX
ABSTRACT.docxiyhkkkkkkkkkkkkkkkkkkkkkkkkkkkk
suriyakalavinoth
 
PPT
SET-02_SOCS_ESE-DEC23__B.Tech%20(CSE-H+NH)-AIML_5_CSAI300
dhruvkeshav123
 
PPTX
02 Fundamental Concepts of ANN
Tamer Ahmed Farrag, PhD
 
PPTX
Artificial Neural Network
NainaBhatt1
 
PPTX
Artificial neural network
nainabhatt2
 
PPTX
Artificial Neural Network in Medical Diagnosis
Adityendra Kumar Singh
 
PPTX
ANN - UNIT 1.pptx
SRM Institute of Science and Technology
 
PPTX
artificialneuralnetwork-200611082546.pptx
olisahchristopher
 
PPTX
The Backbone of Modern AI Models" The architecture of Transformers
JoshiniM2
 
DOCX
Neural networks of artificial intelligence
alldesign
 
PPTX
Unit one ppt of deeep learning which includes Ann cnn
kartikaursang53
 
DOCX
Neural networks report
ChiradipBhattacharya
 
PDF
Machine learningiwijshdbebhehehshshsj.pdf
arewho557
 
DOCX
Artificial Intelligence.docx
ashumar
 
PPTX
Deep learning Techniques JNTU R20 UNIT 2
EXAMCELLH4
 
PPT
AI-CH5 (ANN) - Artificial Neural Network
abrahadawit101
 
PPTX
Karan ppt for neural network and deep learning
KathiriyaParthiv
 
PPTX
Artificial Neural Network
Burhan Muzafar
 
Neural networking this is about neural networks
v02527031
 
Artificial Neural Networking
shresthapragyan063
 
ABSTRACT.docxiyhkkkkkkkkkkkkkkkkkkkkkkkkkkkk
suriyakalavinoth
 
SET-02_SOCS_ESE-DEC23__B.Tech%20(CSE-H+NH)-AIML_5_CSAI300
dhruvkeshav123
 
02 Fundamental Concepts of ANN
Tamer Ahmed Farrag, PhD
 
Artificial Neural Network
NainaBhatt1
 
Artificial neural network
nainabhatt2
 
Artificial Neural Network in Medical Diagnosis
Adityendra Kumar Singh
 
ANN - UNIT 1.pptx
SRM Institute of Science and Technology
 
artificialneuralnetwork-200611082546.pptx
olisahchristopher
 
The Backbone of Modern AI Models" The architecture of Transformers
JoshiniM2
 
Neural networks of artificial intelligence
alldesign
 
Unit one ppt of deeep learning which includes Ann cnn
kartikaursang53
 
Neural networks report
ChiradipBhattacharya
 
Machine learningiwijshdbebhehehshshsj.pdf
arewho557
 
Artificial Intelligence.docx
ashumar
 
Deep learning Techniques JNTU R20 UNIT 2
EXAMCELLH4
 
AI-CH5 (ANN) - Artificial Neural Network
abrahadawit101
 
Karan ppt for neural network and deep learning
KathiriyaParthiv
 
Artificial Neural Network
Burhan Muzafar
 
Ad

Recently uploaded (20)

PPT
Oxygen Co2 Transport in the Lungs(Exchange og gases)
SUNDERLINSHIBUD
 
PPTX
Heart Bleed Bug - A case study (Course: Cryptography and Network Security)
Adri Jovin
 
PDF
Ethics and Trustworthy AI in Healthcare – Governing Sensitive Data, Profiling...
AlqualsaDIResearchGr
 
PDF
PORTFOLIO Golam Kibria Khan — architect with a passion for thoughtful design...
MasumKhan59
 
DOC
MRRS Strength and Durability of Concrete
CivilMythili
 
PPTX
Green Building & Energy Conservation ppt
Sagar Sarangi
 
PPTX
site survey architecture student B.arch.
sri02032006
 
PPTX
Pharmaceuticals and fine chemicals.pptxx
jaypa242004
 
PPTX
Destructive Tests corrosion engineer (1).pptx
zeidali3
 
PDF
Statistical Data Analysis Using SPSS Software
shrikrishna kesharwani
 
PPTX
Day2 B2 Best.pptx
helenjenefa1
 
PDF
GTU Civil Engineering All Semester Syllabus.pdf
Vimal Bhojani
 
PDF
monopile foundation seminar topic for civil engineering students
Ahina5
 
PPTX
artificial intelligence applications in Geomatics
NawrasShatnawi1
 
PDF
UNIT-4-FEEDBACK AMPLIFIERS AND OSCILLATORS (1).pdf
Sridhar191373
 
PPTX
Innowell Capability B0425 - Commercial Buildings.pptx
regobertroza
 
PPTX
265587293-NFPA 101 Life safety code-PPT-1.pptx
chandermwason
 
PDF
6th International Conference on Machine Learning Techniques and Data Science ...
ijistjournal
 
PPTX
UNIT DAA PPT cover all topics 2021 regulation
archu26
 
PDF
Set Relation Function Practice session 24.05.2025.pdf
DrStephenStrange4
 
Oxygen Co2 Transport in the Lungs(Exchange og gases)
SUNDERLINSHIBUD
 
Heart Bleed Bug - A case study (Course: Cryptography and Network Security)
Adri Jovin
 
Ethics and Trustworthy AI in Healthcare – Governing Sensitive Data, Profiling...
AlqualsaDIResearchGr
 
PORTFOLIO Golam Kibria Khan — architect with a passion for thoughtful design...
MasumKhan59
 
MRRS Strength and Durability of Concrete
CivilMythili
 
Green Building & Energy Conservation ppt
Sagar Sarangi
 
site survey architecture student B.arch.
sri02032006
 
Pharmaceuticals and fine chemicals.pptxx
jaypa242004
 
Destructive Tests corrosion engineer (1).pptx
zeidali3
 
Statistical Data Analysis Using SPSS Software
shrikrishna kesharwani
 
Day2 B2 Best.pptx
helenjenefa1
 
GTU Civil Engineering All Semester Syllabus.pdf
Vimal Bhojani
 
monopile foundation seminar topic for civil engineering students
Ahina5
 
artificial intelligence applications in Geomatics
NawrasShatnawi1
 
UNIT-4-FEEDBACK AMPLIFIERS AND OSCILLATORS (1).pdf
Sridhar191373
 
Innowell Capability B0425 - Commercial Buildings.pptx
regobertroza
 
265587293-NFPA 101 Life safety code-PPT-1.pptx
chandermwason
 
6th International Conference on Machine Learning Techniques and Data Science ...
ijistjournal
 
UNIT DAA PPT cover all topics 2021 regulation
archu26
 
Set Relation Function Practice session 24.05.2025.pdf
DrStephenStrange4
 
Ad

Introduction to ANN Principles and its Applications in Solar Energy Technology

  • 1. Department of Engineering American University of Iraq, Sulaimani (AUIS) 1
  • 4.  GENERAL FRAMEWORK OF AI  ARTIFICIAL INTELLIGENCE  MACHINE LEARNING  ARTIFICIAL NEURAL NETWORKS  DEEP LEARNING 4 ARTIFICIAL NEURAL NETWORKS
  • 6. WHAT – WHY - HOW 6
  • 8. Artificial Intelligence  What is AI ? Branch of computer science that focuses on creating systems that are intelligent and independent.  Why use AI ? To enhances the speed, precision and effectiveness of human efforts.  How is AI implemented ? AI can be developed through learning that is symbolic-based or data-based. 8
  • 9. Machine learning  What is Machine Learning ? Using learning paradigms in machines to achieve pattern recognition and other human functions.  Why use Machine Learning ? Machines can utilize more data, and dimensions of data, to perform functions such as pattern recognition.  How is Machine Learning developed ? Machines learn through either statistical learning or deep learning. 9
  • 10. The State-of-AI 10 Text AI Visual AI Interactive AI Functional AI Analytic AI text recognition, speech-to-text conversion machine translation, and content generation capabilities. Chatbots and smart personal assistants. Computer vision or augmented reality Sentiment analysis and supplier risk assessment Maintenance of breakdown in machines after spotting errors, etc. Examples: Alex Bekker. 5 Types of AI to Propel Your Business. Science soft. Link: https://www.scnsoft.com/blog/artificial-intelligence-types
  • 11. 11 Source: Alex Bekker,ScienceSoft. link: https://www.scnsoft.com/blog/artificial-intelligence-types The State-of-AI
  • 12. The State-of-AI 12 Symbolic learning Machine learning Pattern recognition Computer vision Neural Networks Deep Learning learning capabilities (or) reasoning capabilities What about combining the two ? Able to give explanations and can manipulate complex data structures Richa Bhatia. Understanding the difference between symbolic AI & non symbolic AI. https://analyticsindiamag.com/understanding-difference-symbolic-ai-non-symbolic-ai/
  • 13. 13 Symbolic learning Symbolic learning See Process Computer vision Robotics Interact Symbolic learning
  • 14. 14 Human ability to interact with the world System that can learn Symbolic learning Machine learning Pattern recognition Computer vision Neural Networks Deep Learning Machine learning
  • 16. ARTIFICIAL NEURAL NETWORKS  What are Artificial Neural Networks ?  Why use Artificial Neural Networks ?  How are Artificial Neural Networks developed ? Answers in the presentation > slides 15-41 16
  • 17. Deep Learning  What is Deep Learning ?  Why use Deep Learning ?  How is Deep Learning achieved ? Answers in the presentation > slide 42 17
  • 19. Artificial Neural Networks (ANN) Motivation 19
  • 20. Artificial Neural Networks (ANN) Advantages 20
  • 21. Artificial Neural Networks (ANN) Components 21
  • 25. Artificial Neural Networks (ANN) Numerical Data Labeling Images Video Text Time-series Clustering Machine perception Classification Clustering C 25
  • 26. Artificial Neural Networks (ANN) Classification SUPERVISED LEARNING Face detection Object Identification Text classification Voice detection Gesture Identification Transcript speech-text Create Numerical Data Assign to data TRAIN THE ANN Create correlation 26
  • 27. Artificial Neural Networks (ANN) Ni, D. X. (2007). Application of neural networks to character recognition. Proceedings of students/faculty research day, CSIS, Pace University, May 4th. Classification C 27
  • 28. Artificial Neural Networks (ANN) Classification https://www.analyticsindiamag.com/how-to-create-your-first-artificial-neural-network-in-python/ C 28
  • 29. Artificial Neural Networks (ANN) Regression SUPERVISED LEARNING Hardware breakdown Employee turnover Customer churn Health breakdown Machine output C 29
  • 30. Artificial Neural Networks (ANN) Regression https://dataaspirant.com/2017/03/02/how-logistic-regression-model-works/ C 30
  • 31. Artificial Neural Networks (ANN) Regression https://dataaspirant.com/2017/03/02/how-logistic-regression-model-works/ C 31
  • 32. Artificial Neural Networks (ANN) Regression https://dataaspirant.com/2017/03/02/how-logistic-regression-model-works/ C 32
  • 33. Artificial Neural Networks (ANN) Clustering UNSUPERVISED LEARNING Comparing documents, images, etc. to surface similar items Detecting anomalies, or unusual behavior C 33
  • 34. Artificial Neural Networks (ANN) Clustering C 34
  • 35. Artificial Neural Networks (ANN) 35 https://quantumcomputingtech.blogspot.com/2019/02/machine-learning-animated-gif.html Clustering
  • 36. Artificial Neural Networks (ANN) REINFORCEMENT Self-driving car Customer management Adaptive traffic signal control Games Reinforcement LEARNING C 36
  • 37. 37 Artificial Neural Networks (ANN) REINFORCEMENT This paradigm does not require a labelled input/output pairs to be presented and does not need sub-optimal actions to be explicitly corrected.
  • 38. Artificial Neural Networks (ANN) Learning paradigms SUPERVISED LEARNING UNSUPERVISED LEARNING REINFORCEMENT LEARNING 38
  • 39. Artificial Neural Networks (ANN) Learning paradigms Kaplan, S. (2017). Deep generative models for synthetic retinal image generation. C 39
  • 40. Artificial Neural Networks (ANN) Learning paradigms Unsupervised learning Supervised learning Clustering: An Introduction to Supervised Machine Learning and Pattern Classification: The Big Picture by Sebastian Raschka C 40
  • 41. Artificial Neural Networks (ANN) SINGLE LAYER C 41
  • 42. Artificial Neural Networks (ANN) MULTIPLE LAYER PERCEPTRON (MLP) C 42
  • 43. Artificial Neural Networks (ANN) MULTIPLE LAYER PERCEPTRON (MLP) C 43
  • 44. Artificial Neural Networks (ANN) MULTIPLE LAYER PERCEPTRON (MLP) C 44
  • 45. Artificial Neural Networks (ANN) MULTIPLE LAYER PERCEPTRON (MLP) Jain & Mao, 1996 C Deep 45
  • 46. Artificial Neural Networks (ANN) MULTIPLE LAYER PERCEPTRON (MLP) A MLP consists of at least three layers of nodes: an input layer, a hidden layer and an output layer. Except for the input nodes, each node is a neuron that uses a nonlinear activation function. MLP utilizes a supervised learning technique called backpropagation for training. Its multiple layers and non-linear activation distinguish MLP from a linear perceptron. It can distinguish data that is not linearly separable. 46
  • 47. Artificial Neural Networks (ANN) MULTIPLE LAYER PERCEPTRON (MLP) Changing connection weights after each piece of data is processed, based on the amount of error in the output compared to the expected result. 47
  • 48. Artificial Neural Networks (ANN) MULTIPLE LAYER PERCEPTRON (MLP) MLPs are useful in research for their ability to solve problems stochastically, which often allows approximate solutions for extremely complex problems like fitness approximation. MLPs are universal function approximators as shown by Cybenko's theorem, so they can be used to create mathematical models by regression analysis. As classification is a particular case of regression when the response variable is categorical, MLPs make good classifier algorithms. MLPs were a popular machine learning solution in the 1980s, finding applications in diverse fields such as speech recognition, image recognition, and machine translation software, but thereafter faced strong competition from much simpler (and related) support vector machines. Interest in backpropagation networks returned due to the successes of deep learning. 48
  • 49. Artificial Neural Networks (ANN) MULTIPLE LAYER PERCEPTRON (MLP) 49
  • 50. Artificial Neural Networks (ANN) FEED-FORWARD Connections between the nodes do not form a cycle. The first and simplest type of artificial neural network devised. 50
  • 51. Artificial Neural Networks (ANN) BACKPROPAGATION 51
  • 52. Artificial Neural Networks (ANN) BACKPROPAGATION 52
  • 53. Artificial Neural Networks (ANN) BACKPROPAGATION  Supervised learning  Used at each layer to minimize the error between the layer’s response and the actual data  The error at each hidden layer is an average of the evaluated error 53
  • 54. Artificial Neural Networks (ANN) BACKPROPAGATION  N is a neuron.  Nw is one of N’s inputs weights  Nout is N’s output.  Nw = Nw +Δ Nw  Δ Nw = Nout * (1‐ Nout)* NErrorFactor  NErrorFactor = NExpectedOutput – NActualOutput  This works only for the last layer, as we can know the actual output, and the expected output. 54
  • 55. Artificial Neural Networks (ANN) Training stage Validating stage Testing stage 55
  • 57. 57 REGRESSION A measure of the relation between the mean value of one variable (e.g. output) and corresponding values of other variables (e.g. time and cost). Dependent variable Independent variables
  • 62. PV/T modeling and performance prediction • Why use artificial neural networks (ANN) in solar energy technologies? • How to use artificial neural networks (ANN) in solar energy technologies? • How to utilize regression and classification to invest in solar energy? 62
  • 63. PV performance time-series prediction Short-term power forecasting Estimating power loss due to environment Estimating the energetic performance of PV/T PV fault detection Optimize PV array inclination Solar Energy Forecasting Optimize maximum power point tracking 63
  • 64. PV performance time-series prediction C Senkal, O. & Kuleli, T. (2009). Estimation of solar radiation over Turkey using artificial neural network and satellite data. Applied Energy, Vol. 86, pp. 1222– 64
  • 65. PV performance time-series prediction Almonacid, F., Rus, C., Hontoria, L., Fuentes, M. & Nofuentes G. (2009). Characterisation of Si-crystalline PV modules by artificial neural Networks. Renewable Energy, C 65
  • 66. PV performance time-series prediction Firefly algorithm (FA) and Particle Swarm Optimization (PSO) applied to train NN EXAMPLE 1: SHORT TERM POWER FORECASTING Demirdelen, T., Aksu, I. O., Esenboga, B., Aygul, K., Ekinci, F., & Bilgili, M. (2019). A new method for generating short-term power forecasting based on artificial neural networks and optimization methods for solar photovoltaic power plants. In Solar Photovoltaic Power Plants (pp. 165-189). Springer, Singapore. Multi-Layer Feed Forward (MLFF) neural network structure The data of 1MWPV power plant in Turkey is used to estimate output power by real-time data mining for short time prediction. 66
  • 67. PV performance time-series prediction Example 1 C 67
  • 68. PV performance time-series prediction Example 1 Ambient temperature [◦C] Solar radiation [W/m2] PV Panel temperature [◦C] Input Layer Hidden Layer Output Layer 24 8 Interconnections Interconnections PV power (W) 68
  • 69. PV/T performance time-series prediction Particle Swarm Optimization (PSO) applied to train NN EXAMPLE 2: ESTIMATING ENERGETIC PERFORMANCE Alnaqi, A. A., Moayedi, H., Shahsavar, A., & Nguyen, T. K. (2019). Prediction of energetic performance of a building integrated photovoltaic/thermal system thorough artificial neural network and hybrid particle swarm optimization models. Energy conversion and management, 183, 137-148. Multi-Layer Feed-Forward Back-Propagation (FFBP) neural network structure building integrated photovoltaic/thermal system 69
  • 70. PV/T performance time-series prediction Example 2 C 70
  • 71. PV/T performance time-series prediction Example 2 C 71
  • 72. Statistical methods • Coefficient of determination • Mean Absolute Percentage Error • Root Mean Square Error • Mean Absolute Error • Mean Square Error difference between two continuous variables how well a regression model is capable of describing a data set a measure of prediction accuracy of a forecasting method in statistics, a measure of the quality of an estimator measure of the differences between an estimator and the values observed 72
  • 73. Statistical methods COEFFICIENT OF DETERMINATION a measure for how well a regression model is capable of describing a data set. used to evaluate the validity of the predictive results compared to the actual (experimental) real model results 𝒚𝒊= experimental value of (y) 𝒇𝒊= predicted value of (y) 𝒚𝒊= the mean of the experimental values N = the number of values 73
  • 74. Statistical methods MEAN ABSOLUTE ERROR a measure of difference between two continuous variables. the average vertical distance between each point and the identity line. MAE is also the average horizontal distance between each point and the identity line. 74
  • 75. Statistical methods MEAN ABSOLUTE PERCENTAGE ERROR a measure of prediction accuracy of a forecasting method in statistics, 75
  • 76. Statistical methods MEAN SQUARE ERROR measures the average of the squares of the errors—that is, the average squared difference between the estimated values and what is estimated The MSE is a measure of the quality of an estimator—it is always non- negative, and values closer to zero are better. 76
  • 77. Statistical methods ROOT MEAN SQUARE ERROR a measure for the difference between a data set and a corresponding fit. The RMS asymptotically converges to the standard deviation from the model's predicted value for sufficiently large sizes of data sets. measure of the differences between values (sample or population values) predicted by a model or an estimator and the values observed. 77
  • 78. Applications in engineering disciplines Civil engineering 78 Suryanita, R. (2016). The application of artificial neural networks in predicting structural response of multistory building in the region of Sumatra island. KnE Engineering. Reinforced Concrete Building Models
  • 79. Applications in engineering disciplines Electrical engineering 79 Hiyama, T., Kouzuma, S., Imakubo, T.:‘Identification of optimal operating pointof PV modules using neural network for real time maximum power trackingcontrol’,IEEE Trans. Energy Convers., 1995,10, (2), pp. 360–367
  • 80. Applications in engineering disciplines Mechanical engineering 80 Ranasinghe, R. A. T. M., Jaksa, M. B., Kuo, Y. L., & Nejad, F. P. (2017). Application of artificial neural networks for predicting the impact of rolling dynamic compaction using dynamic cone penetrometer test results. Journal of Rock Mechanics and Geotechnical Engineering, 9(2), 340-349. Predicting the impact of rolling dynamic compaction using dynamic cone penetrometer test results
  • 81. Applications in engineering disciplines Control engineering 81 KACHULKOVA, S. Applications of Artificial Neural Networks in Control Systems. transfer, 1(1), 5.
  • 83. Tips to start utilizing machine learning and ANN techniques for your previous, current and future studies in your field… PAST – PRESENT - FUTURE 83 ANN FOR RESEARCH
  • 84. 84 TIP #1 CLASSIFY THE VARIABLES DEPENDENT VARIABLES (DV) INDEPENDENT VARIABLES (IV) CONTROL VARIABLES (CV)
  • 85. 85 TIP #2 ENSURE HIGH-QUALITY OF DATA GIGO GARBAGE IN = GARBAGE OUT
  • 87. 87 TIP #2 ENSURE HIGH-QUALITY OF DATA WHAT IS YOUR UNCERTAINTY THRESHOLD? WHAT MODEL ARE YOU USING FOR UNCERTAINTY ANALYSIS? WHAT IS YOUR UNCERTAINTY THRESHOLD?
  • 88. 88 Example of uncertainty model and analysis The uncertainty associated with the result is WR and those associated with the independent variables are W1, W2, …, Wn. Source: Kline, S. J. (1953). Describing uncertainty in single sample experiments. Mech. Engineering, 75, 3-8.
  • 89. 89 Example of uncertainty model and analysis Source: Sarafraz, M. M., Safaei, M. R., Leon, A. S., Tlili, I., Alkanhal, T. A., Tian, Z., ... & Arjomandi, M. (2019). Experimental investigation on thermal performance of a PV/T-PCM (photovoltaic/thermal) system cooling with a PCM and nanofluid. Energies, 12(13), 2572.
  • 90. 90 Data filtration Height of an adult male by fixing the typo or possibly remeasuring the item or person. If that’s not possible, you must delete the data point because you know it’s an incorrect value. Source: Statistics by Jim. Link: https://statisticsbyjim.com/basics/remove-outliers/ (access date: 25th of July 2022)
  • 91. 91 COLLECT AS MUCH DATA AS POSSIBLE. TIP #3 COLLECT BULK DATA Once testing station, or setup, is established. Record as much data as possible. With a variety of design configurations.
  • 92. 92 Number of input, outputs and hidden nodes. TIP #4 Select the components Utilize the identified variables (IV, DV, CV)
  • 93. 93 How much (%) of the data are you allocating for training? Validating? Testing? TIP #5 Choose the data splitting percentage
  • 94. 94 Use MATLAB/Python to create a simple MLP Neural Network to process the data TIP #6 Start with a simple MLP ANN
  • 97. Food for Thought • What is the difference between Artificial Intelligence and computational intelligence? • How to interpret skewness and kurtosis? • What is the significance of the activation function? • What is the significance of the bias? 97
  • 98.  Identify the dependent and independent variables in your work.  Assign the input and output nodes to your network.  Conduct data pre-processing.  Decide the type of problem you have.  Choose an activation function.  Choose the number of hidden nodes.  Assign weights and biases.  Download MATLAB and implement a simple MLP code.  Carryout the training, validation and testing. Homework 98
  • 100. Review article in PV/T 100
  • 101. Original article on ANN use for PV/T Performance Prediction 101
  • 102. It’s play time… • https://playground.tensorflow.org/ 102

Editor's Notes

  • #8: Self driving car , GPS navigation ,
  • #13: Symbolists firmly believed in developing an intelligent system based on rules and knowledge and whose actions were interpretable
  • #14: https://www.youtube.com/watch?v=2ePf9rue1Ao Symbolic learning uses symbols to represent certain objects and concepts, and allows developers to define relationships between them explicitly.
  • #16: Non-AI programs simply carry out a defined sequence of instructions.
  • #23: Neuron (unit / node / cell)
  • #24: Activation function A sigmoid function is a bounded, differentiable, real function that is defined for all real input values and has a non-negative derivative at each point and exactly one inflection point. A sigmoid "function" and a sigmoid "curve" refer to the same object. hyperbolic tangent function The rectified linear activation function
  • #27: All classification tasks depend upon labeled datasets; that is, humans must transfer their knowledge to the dataset in order for a neural network to learn the correlation between labels and data. This is known as supervised learning.
  • #30: Hardware breakdowns (data centers, manufacturing, transport) Health breakdowns (strokes, heart attacks based on vital stats and data from wearables) Customer churn (predicting the likelihood that a customer will leave, based on web activity and metadata) Employee turnover (ditto, but for employees)
  • #33: Hardware breakdowns (data centers, manufacturing, transport) Health breakdowns (strokes, heart attacks based on vital stats and data from wearables) Customer churn (predicting the likelihood that a customer will leave, based on web activity and metadata) Employee turnover (ditto, but for employees)
  • #37: The agent learns to behave i36 environment depending on these rewards. To understand the limitations and merits of an algorithm and to develop efficient learning algorithms is the goal in reinforcement learning. Sparse: random labelling of individual cells
  • #43: One or more hidden layers
  • #45: This is regression
  • #46: Influence of the number of layers on the pattern recognition ability of MLP
  • #56: Stage (1) and (2) same input / (3) different input
  • #71: Y = m (slope) x + B (intersection)
  • #75: interpretation is clear
  • #77: of an estimator (of a procedure for estimating an unobserved quantity)