{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,30]],"date-time":"2026-07-30T15:22:30Z","timestamp":1785424950562,"version":"3.56.0"},"reference-count":38,"publisher":"MDPI AG","issue":"10","license":[{"start":{"date-parts":[[2026,5,11]],"date-time":"2026-05-11T00:00:00Z","timestamp":1778457600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["52508137"],"award-info":[{"award-number":["52508137"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"award":["52508137"],"award-info":[{"award-number":["52508137"]}],"id":[{"id":"https:\/\/ror.org\/01h0zpd94","id-type":"ROR","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Sensors"],"abstract":"<jats:p>The operating status of chillers in data centers critically affects temperature stability and energy efficiency, imposing stringent requirements on the accuracy, robustness, and adaptability of sensor fault detection systems. Existing Physics-Informed Autoencoders (PIAE) for chiller fault detection face challenges such as gradient conflicts and limited generalization, especially under complex fault scenarios like sensor drift, stuck faults, and precision attenuation. To address these limitations, we propose a Multi-Gradient Descent Algorithm-integrated PIAE (MGDA-PIAE), which embeds the chiller thermal balance equation as a hard constraint and dynamically determines Pareto-optimal weights between data reconstruction and physical consistency. By unifying feature learning with physical law constraints, the model achieves improved generalization and stable performance across diverse operational conditions. Validation using operational data from a data center in Guangzhou demonstrates that MGDA-PIAE significantly outperforms conventional AE and PIAE models: the average recall increases by approximately 20% and the F1-score by 10%, while for flow sensor faults, the F1-score improves by over 80% compared with AE and 20% compared with PIAE. Further tests on multiple sensor types confirm that the model maintains high-precision detection with low false negatives under varying operating modes. By adjusting a single physical threshold, MGDA-PIAE can flexibly meet different fault detection requirements, providing a practical and reliable solution for maintaining the efficient, stable, and safe operation of data center refrigeration systems.<\/jats:p>","DOI":"10.3390\/s26103025","type":"journal-article","created":{"date-parts":[[2026,5,11]],"date-time":"2026-05-11T15:39:32Z","timestamp":1778513972000},"page":"3025","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["A Multi-Gradient-Descent-Integrated Physics-Informed Autoencoder for Sensor Fault Detection in Data Center Chillers"],"prefix":"10.3390","volume":"26","author":[{"given":"Xinyue","family":"Shen","sequence":"first","affiliation":[{"name":"School of Urban Construction, Wuhan University of Science and Technology, Wuhan 430065, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Pan","family":"Li","sequence":"additional","affiliation":[{"name":"School of Urban Construction, Wuhan University of Science and Technology, Wuhan 430065, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Chen","family":"Xu","sequence":"additional","affiliation":[{"name":"School of Urban Construction, Wuhan University of Science and Technology, Wuhan 430065, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Chengliang","family":"Xu","sequence":"additional","affiliation":[{"name":"School of Urban Construction, Wuhan University of Science and Technology, Wuhan 430065, China"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2026,5,11]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","first-page":"121830","DOI":"10.1016\/j.apenergy.2023.121830","article-title":"A Thermodynamic-Law-Integrated Deep Learning Method for High-Dimensional Sensor Fault Detection in Diverse Complex HVAC Systems","volume":"351","author":"Ren","year":"2023","journal-title":"Appl. Energy"},{"key":"ref_2","doi-asserted-by":"crossref","first-page":"113761","DOI":"10.1016\/j.jobe.2025.113761","article-title":"A Review of State-of-the-Art Solutions to Common Sensor Problems in HVAC Systems for Enhancement of Energy Efficiency and Reliability","volume":"112","author":"Sun","year":"2025","journal-title":"J. Build. Eng."},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"112365","DOI":"10.1016\/j.buildenv.2024.112365","article-title":"Sensor Fault Diagnosis and Calibration Techniques in Building Energy Systems: A Review and Future Outlook","volume":"269","author":"Hu","year":"2025","journal-title":"Build. Environ."},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"115360","DOI":"10.1016\/j.enbuild.2025.115360","article-title":"Study on Enhanced Fault Diagnosis of Chiller Units in HVAC Systems under the Imbalanced Data Environment Using GA-Optimized LightGBM","volume":"330","author":"Li","year":"2025","journal-title":"Energy Build."},{"key":"ref_5","doi-asserted-by":"crossref","first-page":"113419","DOI":"10.1016\/j.enbuild.2023.113419","article-title":"Abnormal Energy Consumption Detection Using Ensemble Model for Water Chilling Unit on HVAC System","volume":"297","author":"Cheng","year":"2023","journal-title":"Energy Build."},{"key":"ref_6","doi-asserted-by":"crossref","first-page":"114474","DOI":"10.1016\/j.enbuild.2024.114474","article-title":"A New Integrated Framework to Fault Detection and Diagnosis of Air Handling Unit: Emphasizing the Impact of Symptoms","volume":"319","author":"Cha","year":"2024","journal-title":"Energy Build."},{"key":"ref_7","doi-asserted-by":"crossref","first-page":"123245","DOI":"10.1016\/j.applthermaleng.2024.123245","article-title":"A Two-Stage Fault Diagnosis Strategy for Air Handling Units via a Backpropagation Multidimensional Taylor Network Fitter and a Novel Statistical Process Control","volume":"248","author":"Yan","year":"2024","journal-title":"Appl. Therm. Eng."},{"key":"ref_8","doi-asserted-by":"crossref","first-page":"113367","DOI":"10.1016\/j.enbuild.2023.113367","article-title":"An Efficient Sensor and Thermal Coupling Fault Diagnosis Methodology for Building Energy Systems","volume":"296","author":"Liu","year":"2023","journal-title":"Energy Build."},{"key":"ref_9","doi-asserted-by":"crossref","first-page":"108837","DOI":"10.1016\/j.jobe.2024.108837","article-title":"Experimental Studies on Performance Analysis and Cross-Equipment Parameter Comparison of Variable Refrigerant Flow Systems under Common Faults","volume":"86","author":"Zhang","year":"2024","journal-title":"J. Build. Eng."},{"key":"ref_10","doi-asserted-by":"crossref","first-page":"123734","DOI":"10.1016\/j.applthermaleng.2024.123734","article-title":"A Physics-Constrained Deep Learning Framework for Dynamic Modeling of Vapor Compression Systems","volume":"254","author":"Ma","year":"2024","journal-title":"Appl. Therm. Eng."},{"key":"ref_11","doi-asserted-by":"crossref","first-page":"114210","DOI":"10.1016\/j.jobe.2025.114210","article-title":"Prior-Informed Transductive Inference Framework: A Chiller Multi-Source Faulty Sensor Detection and Diagnosis under Information Incompleteness Conditions","volume":"114","author":"Cao","year":"2025","journal-title":"J. Build. Eng."},{"key":"ref_12","doi-asserted-by":"crossref","first-page":"111124","DOI":"10.1016\/j.buildenv.2023.111124","article-title":"Sensor Fault Diagnosis and Correction for Data Center Cooling System Using Hybrid Multi-Label Random Forest and Bayesian Inference","volume":"249","author":"Wang","year":"2024","journal-title":"Build. Environ."},{"key":"ref_13","doi-asserted-by":"crossref","first-page":"113609","DOI":"10.1016\/j.jobe.2025.113609","article-title":"Evaluation of HVAC & Refrigeration System Fault Behaviors and Impacts: A Systematic Review","volume":"112","author":"Chen","year":"2025","journal-title":"J. Build. Eng."},{"key":"ref_14","doi-asserted-by":"crossref","first-page":"109620","DOI":"10.1016\/j.buildenv.2022.109620","article-title":"Unsupervised Outlier Detection Using Neural Network-Based Mixtures of Probabilistic Principal Component Analyzers for Building Chiller Fault Diagnosis","volume":"225","author":"Tra","year":"2022","journal-title":"Build. Environ."},{"key":"ref_15","doi-asserted-by":"crossref","first-page":"111683","DOI":"10.1016\/j.buildenv.2024.111683","article-title":"A Novel Evaluation Method of Measurement Sensitivities on Common Faults in VAV HVAC Systems","volume":"261","author":"Chen","year":"2024","journal-title":"Build. Environ."},{"key":"ref_16","doi-asserted-by":"crossref","first-page":"110573","DOI":"10.1016\/j.jobe.2024.110573","article-title":"Multi-Sensor Fault Detection and Correction for Automated IAQ Monitoring in Smart Buildings through Attention-Aware Autoencoders with Spatial Prediction Module","volume":"96","author":"Tariq","year":"2024","journal-title":"J. Build. Eng."},{"key":"ref_17","doi-asserted-by":"crossref","first-page":"114104","DOI":"10.1016\/j.buildenv.2025.114104","article-title":"Agentic Fault Detection and Diagnosis for Building HVAC Systems Using AI Agent, Ontology, and Diagnosis Rules","volume":"289","author":"Li","year":"2026","journal-title":"Build. Environ."},{"key":"ref_18","doi-asserted-by":"crossref","first-page":"116580","DOI":"10.1016\/j.enbuild.2025.116580","article-title":"Explainable Fault Detection and Diagnosis in HVAC Systems Using a Koopman Operator with Normal Data","volume":"350","author":"Park","year":"2026","journal-title":"Energy Build."},{"key":"ref_19","doi-asserted-by":"crossref","first-page":"114235","DOI":"10.1016\/j.enbuild.2024.114235","article-title":"A Machine Learning Framework for Detection and Severity Estimation of Faults for Chillers and Air Handling Units in HVAC Systems","volume":"313","author":"Bam","year":"2024","journal-title":"Energy Build."},{"key":"ref_20","doi-asserted-by":"crossref","first-page":"110574","DOI":"10.1016\/j.jobe.2024.110574","article-title":"Multi-Condition Incipient Fault Detection for Chillers Based on Local Anomaly Kernel Entropy Component Analysis","volume":"96","author":"Lu","year":"2024","journal-title":"J. Build. Eng."},{"key":"ref_21","doi-asserted-by":"crossref","first-page":"120862","DOI":"10.1016\/j.apenergy.2023.120862","article-title":"An Enhanced Principal Component Analysis Method with Savitzky-Golay Filter and Clustering Algorithm for Sensor Fault Detection and Diagnosis","volume":"337","author":"Wen","year":"2023","journal-title":"Appl. Energy"},{"key":"ref_22","doi-asserted-by":"crossref","first-page":"116462","DOI":"10.1016\/j.enbuild.2025.116462","article-title":"A Fault Detection and Diagnosis Method Based on DWVMD-PCA-LightGBM for Multi-Sensor Faults at VAV Terminals in Office Areas of Industrial Buildings","volume":"348","author":"Xu","year":"2025","journal-title":"Energy Build."},{"key":"ref_23","doi-asserted-by":"crossref","first-page":"108230","DOI":"10.1016\/j.jobe.2023.108230","article-title":"Fault Detection, Diagnosis and Calibration of Heating, Ventilation and Air Conditioning Sensors by Combining Principal Component Analysis and Improved Bayesian Inference","volume":"82","author":"Li","year":"2024","journal-title":"J. Build. Eng."},{"key":"ref_24","doi-asserted-by":"crossref","first-page":"104243","DOI":"10.1016\/j.jobe.2022.104243","article-title":"Fault Detection and Diagnosis Using Tree-Based Ensemble Learning Methods and Multivariate Control Charts for Centrifugal Chillers","volume":"51","author":"Yao","year":"2022","journal-title":"J. Build. Eng."},{"key":"ref_25","doi-asserted-by":"crossref","first-page":"127557","DOI":"10.1016\/j.apenergy.2026.127557","article-title":"Contrastive Self-Supervised Learning for Lightweight and Automated Fault Detection and Diagnosis in HVAC Systems","volume":"410","author":"Gao","year":"2026","journal-title":"Appl. Energy"},{"key":"ref_26","doi-asserted-by":"crossref","first-page":"111893","DOI":"10.1016\/j.enbuild.2022.111893","article-title":"Outlier Detection via Multiclass Deep Autoencoding Gaussian Mixture Model for Building Chiller Diagnosis","volume":"259","author":"Tra","year":"2022","journal-title":"Energy Build."},{"key":"ref_27","doi-asserted-by":"crossref","first-page":"113251","DOI":"10.1016\/j.jobe.2025.113251","article-title":"Fault Detection and Diagnosis and Fault-Tolerant Control for Air-Handling Unit via Agent-Based Hardware-in-the-Loop Simulation","volume":"111","author":"Kim","year":"2025","journal-title":"J. Build. Eng."},{"key":"ref_28","doi-asserted-by":"crossref","first-page":"113572","DOI":"10.1016\/j.enbuild.2023.113572","article-title":"Diagnosis of Single and Multiple-Source Faults of Chiller Sensors Using EWEEMD-ICKNN by Time Sequence Denoising and Non-Gaussian Distribution Feature Extraction","volume":"298","author":"Zhang","year":"2023","journal-title":"Energy Build."},{"key":"ref_29","doi-asserted-by":"crossref","first-page":"111515","DOI":"10.1016\/j.asoc.2024.111515","article-title":"An Adaptive Fault Detection Model Based on Variational Auto-Encoders and Unsupervised Transfer Learning","volume":"157","author":"Shang","year":"2024","journal-title":"J. Appl. Soft Comput."},{"key":"ref_30","doi-asserted-by":"crossref","first-page":"114009","DOI":"10.1016\/j.enbuild.2024.114009","article-title":"An Online Fault Detection and Diagnosis Method of Sensors in District Heating Substations Based on Long Short-Term Memory Network and Adaptive Threshold Selection Algorithm","volume":"308","author":"Xue","year":"2024","journal-title":"J. Energy Build."},{"key":"ref_31","doi-asserted-by":"crossref","first-page":"112876","DOI":"10.1016\/j.enbuild.2023.112876","article-title":"A Timeseries Supervised Learning Framework for Fault Prediction in Chiller Systems","volume":"285","author":"Jiang","year":"2023","journal-title":"Energy Build."},{"key":"ref_32","doi-asserted-by":"crossref","first-page":"112646","DOI":"10.1016\/j.jobe.2025.112646","article-title":"Augmentation Framework for HVAC Fault Diagnosis Based on Denoising Diffusion Models","volume":"106","author":"Zhang","year":"2025","journal-title":"J. Build. Eng."},{"key":"ref_33","doi-asserted-by":"crossref","first-page":"113795","DOI":"10.1016\/j.jobe.2025.113795","article-title":"A Concurrent Fault Detection Method for Chillers Based on Feature Decoupling Multi-Label Resnet-FCNN Hybrid Model","volume":"112","author":"Li","year":"2025","journal-title":"J. Build. Eng."},{"key":"ref_34","doi-asserted-by":"crossref","first-page":"108904","DOI":"10.1016\/j.buildenv.2022.108904","article-title":"Diagnosis for Multiple Faults of Chiller Using ELM-KNN Model Enhanced by Multi-Label Learning and Specific Feature Combinations","volume":"214","author":"Li","year":"2022","journal-title":"Build. Environ."},{"key":"ref_35","doi-asserted-by":"crossref","first-page":"111420","DOI":"10.1016\/j.ymssp.2024.111420","article-title":"Physics-Informed Deep Autoencoder for Fault Detection in New-Design Systems","volume":"215","author":"Lai","year":"2024","journal-title":"Mech. Syst. Signal Process."},{"key":"ref_36","doi-asserted-by":"crossref","first-page":"686","DOI":"10.1016\/j.jcp.2018.10.045","article-title":"Physics-Informed Neural Networks: A Deep Learning Framework for Solving Forward and Inverse Problems Involving Nonlinear Partial Differential Equations","volume":"378","author":"Raissi","year":"2019","journal-title":"J. Comput. Phys."},{"key":"ref_37","doi-asserted-by":"crossref","first-page":"313","DOI":"10.1016\/j.crma.2012.03.014","article-title":"Multiple-Gradient Descent Algorithm (MGDA) for Multiobjective Optimization","volume":"350","year":"2012","journal-title":"Comptes Rendus Math."},{"key":"ref_38","unstructured":"Sener, O., and Koltun, V. (2018, January 2\u20138). Multi-Task Learning as Multi-Objective Optimization. Proceedings of the Neural Information Processing Systems, Montr\u00e9al, QC, Canada."}],"container-title":["Sensors"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/1424-8220\/26\/10\/3025\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,16]],"date-time":"2026-05-16T04:11:28Z","timestamp":1778904688000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/1424-8220\/26\/10\/3025"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,5,11]]},"references-count":38,"journal-issue":{"issue":"10","published-online":{"date-parts":[[2026,5]]}},"alternative-id":["s26103025"],"URL":"https:\/\/doi.org\/10.3390\/s26103025","relation":{},"ISSN":["1424-8220"],"issn-type":[{"value":"1424-8220","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,5,11]]}}}