Management system, smart meter, server, operation method and management method

An operation method of smart meter, comprising:

• measuring electrical energy of at least one power line to obtain at least one batch of electricity data;
• detecting whether a loading event occurs;
• performing data compression on the electricity data obtained during an event period corresponding to the loading event to obtain compressed data when the loading event occurs;
• and
• uploading the compressed data to a remote server,
• wherein the step of performing data compression on the electricity data obtained during an event period comprises:
• analyzing the electricity data obtained during the event period to determine an electrical appliance type corresponding to the loading event;
• dynamically determining a compression rate of the data compression according to the electrical appliance type;
• and
• performing the data compression on the electricity data obtained during the event period according to the compression rate, so as to obtain the compressed data,
• wherein the step of determining the compression rate of the data compression comprises:
• providing a plurality of compression rate-identification accuracy models corresponding to different electrical appliance types;
• selecting a corresponding model a=fi(cp) from the compression rate-identification accuracy models according to the electrical appliance type, wherein a is an identification accuracy, and cp, is the compression rate;
• dynamically determines the compression rate from the corresponding model according to a threshold identification accuracy corresponding to the remote server;
• and
• resolving a following problem to dynamically determine compression rate cp:
• Adjust cp, max fi(cp)
• In subject to
• Rb+Rs+Rc≦b
• where
• Rb=gb(D, cp, Ub)
• Rs=gs(D, cp, Us)
• Rc=gc(D, cp, Uc)
• where D is an amount of data uploaded to the remote server, Ub is a bandwidth unit price, Rb is bandwidth cost, Us is a storage unit price, Rs is storage cost, Uc is a calculation unit price Rc is calculation cost and b is a cost amount, and the identification accuracy a is maximized by determining the compression rate cp under a limited cost amount b.

A smart meter, comprising:

• a power measuring unit, measuring electrical energy of at least one power line to obtain at least one batch of electricity data;
• a communication unit;
• and
• a processing unit, coupled to the power measuring unit and the communication unit, and detecting whether a loading event occurs, wherein, when the loading event occurs, the processing unit provides a plurality of compression rate-identification accuracy models corresponding to different electrical appliance types, analyzes the electricity data obtained during the event period to determine an electrical appliance type corresponding to the loading event, selects a corresponding model a=fi(cp) from the compression rate-identification accuracy models according to the electrical appliance type, wherein a is an identification accuracy, and cp is a compression rate, dynamically determines the compression rate cp from the corresponding model according to a threshold identification accuracy, performs a data compression on the electricity data obtained during the event period to obtain compressed data according to the determined compression rate cp, and the processing unit uploads the compressed data to a remote server through the communication unit,
• wherein the processing unit resolves a following optimization problem to dynamically determine compression rate cp by using an optimization algorithm:
• Adjust cp, max fi(cp)
• In subject to
• Rb+Rs+Rc≦b
• where
• Rb=gb(D, cp, Ub)
• Rs=gs(D, cp, Us)
• Rc=gc(D, cp, Uc)
• where D is an amount of data uploaded to the remote server, Ub is a bandwidth unit price, Rb is bandwidth cost, Us is a storage unit price, Rs is storage cost, Uc is a calculation unit price, Rc is calculation cost, and b is a cost amount, and the identification accuracy a is maximized by determining the compression rate cp under a limited cost amount b.