Introduction

The invention implements a novel formulation of a known algorithm based on Hidden Markov Models, with significant innovations which represent progress with respect to the state of the art and which ensure a more efficient disaggregation of the electrical load.

 

Technical features

The invention implements a novel formulation of a known algorithm based on Hidden Markov Models, with significant innovations which represent progress with respect to the state of the art. Different loads (e.g. domestic devices) are modeled using HMM (Hidden Markov Models, a more informative representation than FSM, Finite State Machines), while the electrical network aggregating the loads is modeled using FHMM (Factorial HMM), i.e. an aggregation of the individual HMMs modeled on the basis of active and reactive power consumption profiles, and in which each evolves independently in parallel to the others. The data obtained from the modeling of the electric loads and the network are used in an optimization problem to extract the disaggregated profiles of the active and reactive power for each device.

 

Possible Applications

• Electrical measurements (including devices);
• Optimization of electric load (energy efficiency);
• Demand Side Management.

Advantages

• Discrimination of different loads (devices) using the reactive component of power, even in the presence of similar active components;
• Accurate modeling of state transitions, using the combination of active and reactive power, i.e. the value pair (Pa, Pr);
• Disaggregation of reactive power profiles.