1: \begin{abstract}

2: %\boldmath

3: The smart grid vision is to build an intelligent power network with an

4: unprecedented level of situational awareness and controllability over its services

5: and infrastructure. This paper advocates statistical inference methods to 

6: robustify power monitoring tasks against the outlier effects owing to faulty readings

7: and malicious attacks, as well as against missing data due to privacy concerns and 

8: communication errors. In this context, a novel 

9: {\em load cleansing and imputation} scheme is developed 

10: leveraging the low intrinsic-dimensionality of spatiotemporal load profiles and

11: the sparse nature of ``bad data.'' A robust estimator based on

12: principal components pursuit (PCP) is adopted, which effects a twofold sparsity-promoting 

13: regularization through an $\ell_1$-norm of the outliers, and the nuclear

14: norm of the nominal load profiles.

15: Upon recasting the non-separable nuclear norm into a form amenable

16: to decentralized optimization, a \textit{distributed} (D-) PCP algorithm is developed to carry out

17: the imputation and cleansing tasks using networked devices

18: comprising the so-termed advanced metering infrastructure. If D-PCP converges 

19: and a qualification inequality is satisfied, the novel distributed estimator provably

20: attains the performance of its centralized PCP counterpart, which has access to 

21: all networkwide data. Computer simulations and tests with real load curve data 

22: corroborate the convergence and effectiveness of the novel D-PCP

23: algorithm.

24: \end{abstract}

25: