440 likes | 454 Views
2 nd Workshop on Domain Driven Data Mining, Session I : S2208 Dec. 15, 2008 Palazzo dei Congressi, Pisa, Italy. Identification of Causal Variables for Building Energy Fault Detection by Semi-supervised LDA & Decision Boundary Analysis.
E N D
2nd Workshop on Domain Driven Data Mining, Session I: S2208 Dec. 15, 2008 Palazzo dei Congressi, Pisa, Italy Identification of Causal Variables for Building Energy Fault Detection by Semi-supervised LDA & Decision Boundary Analysis Keigo Yoshida, Minoru Inui, Takehisa Yairi, Kazuo Machida (Dept. of Aeronautics & Astronautics, the Univ. of Tokyo) Masaki Shioya, and Yoshio Masukawa (Kajima Corp.)
Main Point of the Presentation We propose … A Supportive Method forAnomaly Cause Identification by Combining Traditional Data Analysis and Domain Knowledge Applied to Real Building Energy Management System (BEMS) Root cause of energy wastes was found successfully
Outline • Introduction • Theories • Experiments for Real Data • Conclusions
I/F BEMS Introduction: What is BEMS ? • Building Energy Management Systems • Collect/Monitor Sensor Data in BLDG (temperature, heat consumption etc…) • Energy-efficient Control • Discover Energy Faults (wastes)
Introduction: Problem of BEMS • Hard to identify root causes of Energy Faults (EF) • Complex Relation between Equipments • Data Deluge from Numerous Sensors (approx. 2000 sensors, 20000 points for 20-story) • Current EF Detection: Heuristics Based on Expert’s Empirical Knowledge, usually fuzzy “IF-THEN” rules. “Heuristic Diagnostics is Incomplete” • Fuzziness False Negative Error • Detection-Only Cannot Improve Systems
Data-Driven Knowledge-Based Modeling-Based Experts Data Source Easy Hard Interpretation Expensive Low Modeling Cost Poor High Versatility Early Fault Diagnosis Methods Performance • Feature Extraction • Neural Networks… • FTA/FMEA • Bayesian • Filtering • FDA… Expert System Fuzzy Logic Supervised Learning Unsupervised Learning / Data Mining Knowledge Acquisition Bottleneck Neglecting Useful Knowledge
Data-Driven Knowledge-Based Modeling-Based Experts Data Source Easy Hard Interpretation Expensive Low Modeling Cost Poor High Versatility Proposed Method Performance Proposal Domain Knowledge + Data Analysis Expert System Fuzzy Logic Supervised Learning Unsupervised Learning / Data Mining - Characteristics - Interpretation: exploit domain knowledge Cost: not so high, empirical knowledge only Versatility: easy to apply to various domains & problems Performance: better than heuristics
* Assumption * Incomplete heuristics surely represent abnormal phenomena Variable Identification Contribution to EF Variable # Conceptual Diagram Learning Boundary Experts Detection Rule e.g. Feedback Data Distribution Acquire Reliable Labels with Given Rule DBA Semi-supervised LDA
Outline • Introduction • Theories • Semi-Supervised Linear Discriminant Analysis • Decision Boundary Analysis • Experiments for Real Data • Conclusions
Semi-supervised LDA Learning Boundary Data Distribution Acquire Reliable Labels with Given Rule
Manifold Regularization [M. Belkin et al. 05] Labeled data only • Regularized Least Square Penalty Term (usually squared function norm) Squared loss for labeled data
Squared loss Penalty Term Additional term for intrinsic geometry : graph Laplacian Manifold Regularization [M. Belkin et al. 05] Labeled data only • Regularized Least Square • Laplacian RLS: Penalty Term (usually squared function norm) Squared loss for labeled data Use labeled & unlabeled data Assumption: Geometrically close ⇒ similar label
Regularizer Semi-Supervised Linear Discriminant Analysis (SS-LDA) • LDA seeks projection for small within-cov. & large between-cov. • Regularized Discriminant Analysis: [Friedman 89] • Semi-Supervised Discriminant Analysis (SS-LDA): Between-class Within-class
Learning Boundary Data Distribution Acquire Reliable Labels with Given Rule Semi-supervised LDA Decision Boundary Analysis
Learned Boundary Top view Cross-section view Class 2 Class 1 Normal vec. : disciminantly informative : discriminantly redundant Decision Boundary Analysis • Feature Extraction method proposed by Lee & Landgrabe C. Lee & D. A. Landgrabe. Feature Extraction Based on Decision Boundary, IEEE Trans. Pattern Anal. Mach. Intell. 15(4): 388-400, 1993 • Extract informative features from normal vectors on the boundary
Decision Boundary Feature Matrix • Define responsibility of each variables for discrimination • Linear: • Nonlinear:
Outline • Introduction • Theories • Experiments • Application to Energy Fault Analysis • Conclusions
Energy Fault Diagnosis Problem EF: Inverter overloaded Detection Rule 6h M.A. of Inverter output = 100 EF … but I don’t know the cause cold Inverter hot coil Air Handling Unit humidity
DATA cold & Inverter RULE hot coil Air Handling Unit humidity Energy Fault Diagnosis Problem EF: Inverter overloaded Detection Rule 6h M.A. of Inverter output = 100 EF … but I don’t know the cause Find out root cause of inverter overload
NN = 5, Energy Fault Diagnosis - Settings • Air-conditioning time-series sensor data for 1 unit • instances: 744 • Labeled sample:10for each (3% of all) (based on probability proportional to distance from boundary) • Hyper-parameters: • 13 attributes, all continuous
0 20 40 60 80 100 Contribution Score [%] Results (100 times ave.) Inverter <LDA> Inverter (96%) Trivial
0 20 40 60 80 100 Contribution Score [%] Results (100 times ave.) SA Temp. Cooling water <LDA> Inverter (96%) <SSLDA> Cool water (75%) SA temp. (12%)
0 20 40 60 80 100 Contribution Score [%] Results (100 times ave.) Not Distinctive ! <LDA> Inverter (96%) <SSLDA> Cool water (75%) SA temp. (12%) <KDA> Cool water (19%) MA. Pressure (15%) Inverter (15%) …
0 20 40 60 80 100 Contribution Score [%] Results (100 times ave.) [1] SA Temp. [2] SA Setting Inverter [3] Cooling water <LDA> Inverter (96%) <SSLDA> Cool water (75%) SA temp. (12%) <KDA> Cool water (19%) MA. Pressure (15%) Inverter (15%) <SSKDA> Inverter (33%) SA temp (19%) Cool Water (17%) SA setting (13%) …
Energy Fault Diagnosis: Examine Row Data • Cooling water valve Opening [3] valve opens completely, but this is result of EF, not cause
deviation of SA temp. Energy Fault Diagnosis: Examine Row Data • Cooling water valve Opening valve opens completely, but this is result of EF, not cause • SSLDA/SSKDA show SA temp. [1] & setting [2] responsible • To reduce this deviation… • Operate inverter at peak power • Open cooling water valve
Outline • Introduction • Theories • Experiments for Real Data • Conclusions
Conclusions • Introduce identification method of causal variables by combining semi-supervised LDA & DBA • Labels are acquired from imperfect domain-specific rule • SS-LDA/SS-KDA: reflect domain knowledge & avoid over-fitting • DBA: extract informative features from normal direction of boundary • Apply to energy fault cause diagnosis • Succeeded in extracting some responsible features beginning with fuzzy heuristics based on domain knowledge
Room for improvements • Consider temporal continuity • Time-series is not i.i.d. • Find True Cause from Correlating Variables
Minor improvements • Optimize Hyper-parameters • AIC, BIC, … • Cross Validation • Regularization Term • L1-norm will give sparse solution • Comparison to other discrimination methods • SVM • Laplacian SVM… etc.
Extension to Multiple Energy Faults • In real systems, various faults take place • Fault cause varies among phenomena • Need to separate phenomena and diagnose respectively <Our Approach> 1. Extract points detected by existing heuristics 2. Reduce dimensionality and visualize data in low-dim. space 3. Clustering data and give them labels 4. Identify variables discriminating that cluster from normal data
Experimental Condition & Results • Air-conditioning sensor data, 13 attributes, same heuristics • 748 instances, operating time only (hourly data for 2 months) • 137 points are detected by heuristics • Reduce dimensionality by isomap [J.B. Tenenbaum 00] (kNN = 5) • Contribution score is given by SS-KDA (kNN = 5, ) <2D representation> 2 major cluster, 4 anomalies
Room air Temp. superficial Contribution score for red points Experimental Condition & Results • Air-conditioning sensor data, 13 attributes, same heuristics • 748 instances, operating time only (hourly data for 2 months) • 137 points are detected by heuristics • Reduce dimensionality by isomap [J.B. Tenenbaum 00] (kNN = 5) • Contribution score is given by SS-KDA (kNN = 5, ) <2D representation> Deviation of Room Air Temp. around detected points Detected, this is EF 2 major cluster, 4 anomalies
Properly Controlled System Deviation Data Distribution
Linearly Separable for Cooling Water Valve [3] Cooling Water Valve [%] Data Distribution
: Distance from boundary of point Probabilistic Labeling • Points distant from boundary are reliable as class labels • Keep robustness against outliers Points are stochastically given labels based on reliability Rule outlier Unreliable
Feature space Input space Estimate DBFM • Linear Case: • Nonlinear Case Difficult to acquire points on boundary & calculate gradient vector Disciminant function is linear in feature space Kernelized SSLDA (SS-KDA)
Feature space DBFM for Nonlinear Distribution (1) 1. Generate points on boundary in feature space 2. Gradient vector at corresponding point for Gaussian kernel But to find pre-image is generally difficult… By kernel trick, pre-image problem is avoidable Input space
DBFM for Nonlinear Distribution (2) Finally we have gradient vectors on boundary for each point 3. Construct estimated DBFM • Define responsibility of each variables for discrimination Max. eigenvalue
質問されそうなこと • リアルタイム性は? • 事後処理を想定 • 他の手法と比較したか?なぜLDAか? • SVMでも適用できるので試したい • なぜこういう結果になったのか • 原因変数のデータを見ると線形判別は難しい