Artificial intelligence for building energy analysis : towards high performance computing /: towards high performance computing. (2015)

• Record Type: Book
• Further Information: Note: Frederic Magoules, Hai-Xiang Zhao.
• Authors: Magoulès, F (Frédéric); Zhao, Hai-Xiang
• Contents: List of figures v List of tables vii Nomenclature vii 1 Overview of building energy analysis 1 1.1 Introduction 1 1.2 Definitions 2 1.3 Physical models 2 1.4 Grey models 5 1.5 Statistical models 5 1.5.1 Regression models 5 1.5.2 Auto-regressive model with extra inputs 5 1.5.3 Auto-regressive integrated moving average 5 1.5.4 Conditional demand analysis 5 1.6 Artificial intelligence models 5 1.6.1 Neural networks 5 1.6.2 Support vector machines 5 1.7 Concluding remarks 12 2 Data acquisition for building energy analysis 13 2.1 Introduction 13 2.2 Data measurement 14 2.3 Data simulation 14 2.3.1 Simulation process 14 2.3.2 Simulation software 14 2.3.3 Simulation of multiple buildings 19 2.4 Data uncertainty 20 2.5 Data calibration 20 2.6 Concluding remarks . 20 3 Artificial intelligence models 23 3.1 Introduction 23 3.2 Overview of artificial intelligence models 24 3.3 Neural networks 24 3.3.1 Feed forward neural network 24 3.3.2 Radial basis functions network 24 3.3.3 Reccurent neural network 24 3.3.4 Recursive deterministic perceptron 24 3.3.5 Applications of neural networks 24 3.4 Support vector machines 24 3.4.1 Support vector classification 24 3.4.2 "-support vector regression 28 3.4.3 One-class support vector machines 30 3.4.4 Multi-class support vector machines 31 3.4.5 v-support vector machines 32 3.4.6 Transductive support vector machines 33 3.4.7 Quadratic problem solvers 34 3.4.8 Applications of support vector machines 39 3.5 Concluding remarks 41 4 Artificial intelligence for building energy analysis 43 4.1 Introduction 43 4.2 Support vector machines for building energy prediction 43 4.2.1 Energy prediction definition 43 4.2.2 Practical issues 44 4.2.2.1 Operation flow 44 4.2.2.2 Experimental environment 44 4.2.2.3 Data pre-processing 45 4.2.2.4 Model selection 46 4.2.2.5 Model evaluation 47 4.2.3 Support vector machines for prediction 47 4.2.3.1 Prediction of single building energy 47 4.2.3.2 Extensive model evaluation 49 4.2.3.3 Prediction of multiple buildings energy 51 4.3 Neural networks for building energy fault detection 53 4.3.1 Description of faults 53 4.3.2 Recursive deterministic perceptron in fault detection 53 4.4 Neural networks for building energy fault diagnosis 57 4.4.1 Fault diagnosis definition 57 4.4.2 Practical issues 57 4.4.3 Recursive deterministic perceptron in fault diagnosis 57 4.5 Concluding remarks 59 5 Model reduction for support vector machines 61 5.1 Introduction 61 5.2 Overview of model reduction 62 5.2.1 Filter methods 62 5.2.2 Wrapper methods 62 5.2.3 Embedded methods 62 5.3 Model reduction for energy consumption 63 5.3.1 Algorithm 63 5.3.2 Feature set description 64 5.3.3 Feature set selection 65 5.4 Model reduction for single building energy 65 5.5 Model reduction for multiple buildings energy 69 5.6 Concluding remarks 71 6 Parallel computing for support vector machines 73 6.1 Introduction 73 6.2 Overview of parallel support vector machines 74 6.3 Parallel quadratic problem solver 75 6.4 MPI-based parallel support vector machines 77 6.4.1 Message Passing Interface programming model 77 6.4.2 Pisvm 77 6.5 MapReduce-based parallel support vector machines 77 6.5.1 MapReduce programming model 77 6.5.2 Caching technique 78 6.5.3 Sparse data representation 79 6.5.4 Comparison of MRPsvm with Pisvm 79 6.6 MapReduce-based parallel "-support vector regression 82 6.6.1 Implementation aspects 82 6.6.2 Energy consumption datasets 83 6.6.3 Evaluation for building energy prediction 84 6.7 Concluding remarks 86 7 Future of building energy analysis 89 Bibliography 89
• Edition: 1st
• Publisher Details: London : Wiley-ISTE
• Publication Date: 2015
• Extent: 1 online resource
• Subjects: 696.028563; Buildings -- Energy consumption; Artificial intelligence
• Languages: English
• ISBNs: 9781118577592
• Notes: Note: Description based on CIP data; resource not viewed.
• Access Rights: Legal Deposit; Only available on premises controlled by the deposit library and to one user at any one time; The Legal Deposit Libraries (Non-Print Works) Regulations (UK).
• Access Usage: Restricted: Printing from this resource is governed by The Legal Deposit Libraries (Non-Print Works) Regulations (UK) and UK copyright law currently in force.