ELECTRIC TRANSMISSION OF ENERGY
1 Wide-area monitoring and analysis systems 1 11 Introduction 1 12 Wide-area monitoring systems: a conceptual overview 1 121 WAMS architectures 2 122 Measurement and sensing technologies 4 123 Design of WAMS architecture 4 13 Distributed and semi-distributed monitoring frameworks 5 14 Data collection and management 8 15 Synchrophasor networks 9 151 Network graphs 9 152 Spectral graph analysis 11 153 Connected dominant sets 14 154 Geometric approaches for signal processing 14 16 Challenges of future intelligent monitoring and analysis systems 14 References 18 2 Wide-area monitoring system architectures 25 21 Introduction 25 22 WAMS architectures 25 221 Centralized WAMS architectures 28 222 Hierarchical WAMS architectures 29 223 Hybrid WAMS architectures 30 224 Distributed WAMS architectures 31 23 Issues in data fusion 32 231 Data 32 232 Intelligent synchrophasor data fusion 32 233 Power system data fusion strategies 34 234 General framework for data assimilation 40 235 Fusion or integration of multivariate PMU data 43 24 Relationship between multiblock and single-block models 47 References 48 3 Spatio-temporal modeling of power system dynamic processes 53 31 Introduction 53 32 Visualization of large space-time measurement data 54 33 Spatio-temporal modeling of multivariate processes 55 331 POD analysis 55 332 Singular value decomposition-based POD 59 333 Departure from the mean value 62 34 Spatio-temporal interpolation methods 63 341 Background 63 342 Similarity measures 64 343 Spatial structures 65 344 Derivation of weights 66 345 Practical issues 67 35 Dimensionality reduction 67 351 Projection-based reduced-order modeling 67 352 Proximity (similarity) measures 68 353 Nonlinear spectral dimensionality reduction 70 36 Motivational example 77 361 Small-signal response 77 362 Large system response 78 363 Statistical analysis 78 37 Sensor placement 81 371 Problem formulation 82 372 Constrained sensor placement 83 373 Other approximations 87 374 Compressed sensing 87 375 Sensor placement for signal (state) or process reconstruction 88 References 89 4 Advanced data processing and feature extraction 95 41 Introduction 95 42 Power oscillation monitoring 96 43 Time–frequency representations 97 431 Hilbert–Huang analysis 97 432 Wavelet analysis 105 44 Multivariate multiscale analysis 116 441 Multisignal Prony analysis 116 442 Koopman analysis and its variants 116 443 Dynamic mode decomposition 119 444 Multichannel Prony analysis 120 445 Deep autoencoders 120 45 Response under ambient stimulus 121 451 Formulation of the model 121 452 Modal response 122 453 Ensemble system response 123 46 Application to measured data 124 461 HHT analysis 125 462 Wavelet analysis 128 References 130 5 Multisensor multitemporal data fusion 135 51 Introduction 135 52 Data fusion principles 135 53 Data preprocessing and transformation 138 531 Bandpass filtering and denoising 138 532 Local-level fusion 139 54 Feature extraction and feature selection 140 541 Feature extraction 140 542 Data compression 141 543 Individual scales 144 544 Filtering and multiscale monitoring 144 55 Multisensor fusion methodologies for system monitoring 146 551 Single-scale analysis 146 552 Nonlinear PCA using autoassociative neural networks 147 553 Multiblock POD (PCA) analysis 147 554 Nonlinear PCA 153 555 Blind source separation 154 56 Other approaches to multisensor data fusion 159 57 Illustration 160 571 Multisensor (multiview) data fusion 163 References 164 6 Monitoring the status of the system 167 61 Introduction 167 62 Power system health monitoring 168 63 Disturbance and anomaly detection 169 64 Modal-based health monitoring methods 170 641 Filtering and data conditioning 170 642 Entropy and energy 174 643 Entropy-based detection of system changes 178 65 Wide-area inter-area oscillation monitoring 179 651 Case A 180 652 Case B 180 66 High-dimensional pattern recognition-based monitoring 183 661 Sparse diffusion implementation 183 662 Data clustering 185 663 Numerical example 186 664 Hybrid schemes 186 67 Voltage and reactive power monitoring 187 671 Measured data 188 672 Statistical approach to voltage monitoring 189 673 Complex POD/PCA analysis 192 References 193 7 Wide-area monitoring of power systems with high penetration of distributed generation 197 71 Introduction 197 72 Wide-area monitoring of power systems integrated with geographically distributed generation sources 198 721 Critical issues in system monitoring 198 722 Distributed setup 200 723 Data fusion 200 724 Sensor placement and signal (state) reconstruction 201 725 Spatiotemporal correlation 202 73 Frequency monitoring 205 731 Large-scale frequency response 205 732 RoCoF estimation 207 733 Frequency response monitoring: the COG approach 217 734 Distributed monitoring 220 735 Oscillation detection analysis 220 74 Voltage and reactive power monitoring 222 741 Voltage monitoring 222 742 Dimensionality reduction and visualization 223 743 Islanding and coherency 226 744 Statistical quality control 227 75 Data correlation 230 751 Data correlation and distance measures 230 752 Oscillation detection analysis 233 76 Tensor representations for system monitoring 233 761 Background 234 762 Tensor decompositions 234 763 The CP decomposition 235 764 The Tucker decomposition 236 765 Tucker reconstruction 237 77 A case study involving simultaneous analysis of multiple datasets 238 771 Base case and modified case description 238 772 Time series representation 239 773 Tensor representation 240 774 Implementation and results 242 References 243 8 Distributed wide-area oscillation monitoring 251 81 Introduction 251 82 Hierarchically distributed, wide-area oscillation monitoring 252 83 Hierarchical-distributed WAMS architectures 254 831 Background and motivation 254 832 Partitioning of data blocks 255 833 Partitioning of sensor blocks 257 84 Single-block methods 259 841 Single-block PCA (SB-PCA) models 259 85 MB-PCA methods 261 851 MB-PCA models 261 852 Numerical implementation 263 86 Partial least squares regression 264 87 MB canonical correlation analysis 265 871 Two-block CCA 265 872 Multiset canonical correlation analysis 266 88 Tensor representations 267 881 Tensor-based multiblock representations 267 882 The Tucker decomposition 269 883 HO-SVD analysis 270 884 Tensor rank 271 885 Dynamic system reconstruction from tensor approximations 271 886 Three-way clustering 272 887 Sequentially truncated higher-order SVD (ST-HO-SVD) 272 888 Numerical considerations 273 889 Reconstruction error 273 89 Application to simulated data 274 891 The 14-machine Australian test power system 274 892 Application to a large-scale power system 280 810 Recent work 291 References 291 9 Near real-time analysis and monitoring 299 91 Introduction 299 92 Toward near real-time monitoring of system behavior 299 93 Data processing and conditioning 300 931 Wavelet denoising and filtering 300 932 EMD-based filtering 303 94 Anomaly detection from changes in system behavior 304 941 Event trigger 305 942 Event detection based on linear filtering 306 943 An illustration 308 95 Time-series approaches to detection of abnormal operation 308 951 Near real-time implementations 308 952 Near real-time implementation of the Hilbert transform (HT) 311 953 Local mean speed 317 954 Online applications of wavelet transform 320 955 Other approaches 320 96 Pattern recognition-based disturbance detection 320 97 Sliding window-based methods 322 971 Local HHT analysis 322 972 Numerical example 325 973 Sliding window-based Koopman mode analysis 327 98 Recursive processing methods 327 981 State-space model for linear regression 327 982 Adaptive tracking of system oscillatory modes 328 99 Data-driven prognosis 333 References 333 10 Interpretation and visualization of wide-area phasor measurement unit measurements 339 101 Introduction 339 102 Loss of generation oscillation event 339 1021 Operational context 340 1022 Recorded measurements 341 103 Analysis and visualization of recorded data 344 1031 Mode shape characterization 345 1032 Damping estimation 347 1033 Instantaneous parameters 347 1034 Multitemporal, multiscale analysis of measured data 355 1035 Performance evaluation 359 104 Pattern recognition analysis 361 1041 Diffusion map analysis 361 1042 Comparison with other approaches 363 105 POD/BSS analysis 365 106 Distributed monitoring 367 1061 Single-block representation 368 1062 Multiblock PCA of PMU data 369 1063 C-means analysis 371 107 Validation of power system model 372 1071 Frequency scanning analysis 373 1072 Single-machine infinite bus scan 373 1073 Large system performance 374 108 Evaluation of control performance 376 References 380