Ga l a x y C l oud So f t wa r e a s a Se r v i c e f o r Ene r g y Ma n agemen t 17
from the data. Through Gbots’ supervised and
unsupervised learning, including
case-based reasoning and neural network
techniques, it is possible to identify the early stages
of degradation and signs of malfunction. The Gbots
provide continuous feedback to the facility maintainer
-
or the maintenance Gbots if that role has been
delegated to them. The intelligence and knowledge
acquired allows the GCCC to forestall
potential problems through undertaking appropriate
maintenance, making system
failures extremely rare.
Alerts relating to degrading or faulty equipment are
forwarded to the GCCC. Galaxy chooses appropriate
Gbots to attempt a diagnosis based on the type of
alert. They capture
data and compare it with
the benchmarks established from past experience.
The Galaxy platform commands the Gbots to initiate
actions, such as recording data on equipment
degradation, generating alerts, or turning off a piece of
equipment, based on expertise developed through the
Gbot’s observations and their acquired knowledge.
Corrective
action will be taken automatically using heuristic
approaches, such
as iterative measurement and adjustment of process
parameters, until the problem is rectifed. If the local
Gbots fail to rectify the fault, the GCCC will call in the
central support team to resolve the
problem, which may mean human experts at
the GCCC fxing the problem remotely. If remote
techniques fail, a technician will be
called out.
Pacifc Controls is continuously improving the ways
in which Gbots can be used for remote intelligent
diagnostics.
between a building and its surroundings. The ability
of occupants
to make their own choices and control their
environment is critical to their comfort and
satisfaction but open windows and altered
thermostats can wreck energy management
strategies. Dynamically steering individual’s
attempts to manage their comfort and control their
building can save up to 20% of current energy
consumption. Doing this by allowing greater
interaction between the user and the control
systems through deploying Gbots will increase
comfort and improve performance.
The graphic below illustrates the problem and the
role of Gbots in solving it.
Gbots have proved to be a useful tool to allow
integration and optimization of building systems
from the cloud. In current building automation and
control systems, supervisory control strategies
are developed from heuristics or past operator
experience, making use of Artifcial
Intelligence (AI) and Artifcial Neural
Networks (ANN). The system uses
this experience
to maintain comfort levels, based
on settings that worked in the past
in similar weather or load conditions
without the occupants complaining.
Although this approach does meet
the basic needs of the building’s
occupants, it is almost impossible
to maintain high levels of energy
effciency as conditions change. This problem
can be addressed by Gbots, which can be
programmed to maintain the desired comfort levels
at all times and alert operators immediately to any
problems.
4.3 Gbots carry out Remote
Intelligent Diagnosis
The GCCC on the Galaxy platform is capable of
predicting
imminent component failures using the wealth of
stored data
available from the sensors being monitored. Data
about energy consuming equipment is
downloaded frequently and analyzed immediately
using intelligent
diagnostic techniques. Gbots are intelligent in
extracting knowledge and useful information