Yachen Tang

Yachen Tang

Auto-generated profile, based on publicly available information

Yachen Tang

Univers, Santa Clara, CA, USA

Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 February 2021 in Carbon
Verified
Journal Article
Closed access
XZXiang Zhang
Xiang Zhang
YTYachen Tang
Yachen Tang
PHPingge He
Pingge He
ZZZhao Zhang
Zhao Zhang
TCTengfei Chen
Tengfei Chen

8 August 2024 in IEEE Open Access Journal of Power and Energy
Verified
Journal Article
Open access
ZDZhen Dai
Zhen Dai
SLShouyu Liang
Shouyu Liang
YTYachen Tang
Yachen Tang
JTJun Tan
Jun Tan
GLGuangyi Liu
Guangyi Liu
XLXuanang Li
Xuanang Li
The seamless integration of swift and precise topological analysis with state estimation is crucial for ensuring the dependability, stability, and efficiency of the power system. In response to this need, this paper introduced a novel approach to constructing a spatiotemporal “Power Grid One Graph” model using a graph database, enabling rapid topological analysis and state estimation. Initially, a spatiotemporal power grid model was created by merging grid topology with dynamically updated telemetry and telesignaling data. Subsequently, utilizing the graph model and entity mapping, the spatiotemporal node-breaker graph model was obtained and the corresponding bus-branch model was generated. Based on the node-breaker graph model, topological error identification was conducted, and a fast topological analysis optimization algorithm, considering component functionality, was applied to update the bus-branch graph model, facilitating graph-based state estimation. Finally, the proposed method was validated on a real power system, and its application, along with performance enhancements of the spatiotemporal power grid model considering topological changes, was investigated. The presented method provides both theoretical and practical support for the digital transformation of the power system and the advancement of the digital twin power grid.

19 September 2019 in IEEE Transactions on Smart Grid
Verified
Journal Article
Closed access
YTYachen Tang
Yachen Tang
SZShuaidong Zhao
Shuaidong Zhao
CTChee-Wooi Ten
Chee-Wooi Ten
KZKuilin Zhang
Kuilin Zhang
TLT Logenthiran
T Logenthiran
Over the past decades, ther have been an increased number of the Internet of Things (IoT) sensor deployment in electrical distribution networks. This paper proposes a statistical approach to establish the correlations between estimated occupancy within physical proximity and the associated loads. This study includes a sensitivity analysis of occupancy and how it influences load consumptions. First, a statistical distribution with a regression model is formed to correlate these two heterogeneous properties, in terms of occupancy and power (OP) consumption, to generate a time-dependent model. Since model deviations from an estimate based on regular pattern as a time-varying process with the curve fitting, the possible structural deviations in real-time demand are then considered to update the initial regression model using new real-time estimates and observations obtained every demanded time duration. The dynamic profile of human movements and their load characteristics are established with parametric adjustments and observed using the case studies.

13 October 2025 in IEEE Access
Verified
Journal Article
Open access
ZHZhiwei Huang
Zhiwei Huang
YTYachen Tang
Yachen Tang
CTChee-Wooi Ten
Chee-Wooi Ten
GLGuangyi Liu
Guangyi Liu
This paper addresses security challenges in modern electricity distribution systems, where supervisory control and data acquisition (SCADA) and advanced metering infrastructure (AMI) networks are increasingly exposed to cyber threats and fraudulent activities, leading to metering discrepancies. As smart grids become more interconnected, identifying compromised meters and devices at scale requires a robust inference framework. While alarm systems provide real-time detection of anomalies such as abrupt energy consumption changes, data losses, and security breaches, the lack of seamless integration between SCADA and AMI alarm data limits their effectiveness. To enhance grid security and resilience, this paper presents a data-driven approach for evaluating metering network trustworthiness by analyzing measurement variations from feeder remote terminal units (FRTUs) and IP-based energy meters (EMs) across primary and secondary distribution networks. The proposed probabilistic trust model leverages historical alarm data and event logs, demonstrating its ability to detect discrepancies in a simulated environment. The inferred trust scores are then used to re-weight and reconcile conflicting measurements, allowing the system to attenuate the influence of untrusted data during anomaly detection and state estimation. By correlating alarm patterns with metering anomalies, this approach strengthens cyber-physical security, enhances operational transparency, and supports the transition to more secure and intelligent distribution networks.

24 August 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
JZJin Zhang
Jin Zhang
GHGang Huang
Gang Huang
MCMingjian Cui
Mingjian Cui
SLShi Liang
Shi Liang
YTYachen Tang
Yachen Tang
XLXiaohu Li
Xiaohu Li
Accurate photovoltaic (PV) power forecasts support grid operation and economic scheduling as PV penetration rises, particularly in China. Transformer, PatchTST, and Mamba based forecasting models usually use a fixed-length window of recent observations and have limited access to similar historical contexts. Traditional similar day matching can retrieve numerically close references from meteorologically different periods, such as low irradiance in the early morning and cloud occlusion at noon. We propose a Meteorological Pattern Matching Network (MPM-Net) to address this limitation. The model represents weather with the clear sky index, diffuse fraction, and solar variability index, then constructs a meteorological pattern set from historical data. It retrieves power sequences from periods whose weather states match the forecast day and uses them as an additional matched window. A power boundary correction keeps the forecast within physical limits. Experiments on a real-worlddataset of 228 PV units from northwestern China show that MPM-Net remains competitive over short horizons, with higher accuracy at longer horizons and under complex weather conditions.

2 August 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
JZJin Zhang
Jin Zhang
GHGang Huang
Gang Huang
MCMingjian Cui
Mingjian Cui
SLShi Liang
Shi Liang
YTYachen Tang
Yachen Tang
XLXiaohu Li
Xiaohu Li
As renewable energy penetration rises in modern power systems, particularly in China, accurate photovoltaic (PV) power forecasting has become important for grid stability and economic scheduling. Existing deep forecasting architectures based on Transformers, PatchTST, and Mamba usually rely on a current window of fixed length containing recent observations and thus have limited access to globally similar historical contexts. Meanwhile, traditional similar day matching methods may retrieve references that are numerically close but meteorologically different, such as low irradiance in the early morning and cloud occlusion at noon. To address these limitations, this paper proposes a Meteorological Pattern Matching Network (MPM-Net). A weather state representation combining the clear sky index, the diffuse fraction, and a solar variability index is developed to construct a representative meteorological pattern set from historical data. The matching mechanism retrieves historical patterns similar to the weather state of the forecast day and introduces their power sequences as an additional matched window, extending the effective temporal context beyond the current window of fixed length.A lightweight power boundary correction is further applied to keep forecasts within physical limits. Experiments on a real world dataset of 228 PV units from northwestern China show that MPM-Net achieves competitive performance over short horizons and stronger accuracy at longer horizons, especially under complex weather conditions.