Written by Jan Weustink, Stefan Niebler, Tarun Dhall and Andy Sleeman
09:00 AM: Felix
grabs a coffee in his office at home. He is the maintenance engineer for a
fleet of dozens of “Power to X plants”, which generate e.g. hydrogen from
solar/wind energy offshore and a few more localized Combined Cycle power
plants. These plants are distributed all around the world, but Felix knows
their setup and behavior by heart, from working with their “Digital Twins”. In
another part of the world, Siemens Mindsphere hosts a virtual copy of each site
and each power plant, each environment. It accurately simulates the site
conditions and power plant behavior. Each virtual power plant is being
continuously updated and optimized using data gathered from the real power
plant. A “Digital Twin” continuously simulates the power plant and predicts the
best settings for the control algorithms, automatically adjusting these in
simulation of the actual control software which is also running in the cloud
services.
09:15 AM: Felix
had a good rest last night, without any disturbances and now the first site
analysis reports pop up at his tablet. The Omnivise suite provides him with
prioritized reports classified as “low priority” and “medium priority”, along
with all the relevant data, such as generated load, environmental data,
forecast for the day, but also with the findings of the previous days. The
image analysis system detected a deviation in the temperature profile of a
steam pipe insulation and also an increased temperature at one gearbox. These
images, both full vision and thermal scans, plus sound files were recorded by
flying drones and robots. These autonomous drones and robots frequently
navigate through the power plant, on predefined routes, looking for anything
out of the ordinary. Right after being streamed to the Mindsphere, the images
are analyzed, together with the data collected from the site control system. In
this instance, the site solution is a D3000 and part of a wider DCS. Felix is
now free to work on the problems remotely. He sends a request to visit the
cause of the first report to the site drone. Felix puts on his VR-goggles and
is then able to have a look at the findings himself, with live images. The laser
scanners of the drone provide photo enhanced 3D images, so Felix can even
change the angle of sight without moving the drone. He sees that the insulation
needs repairing, so dispatches a robot to the scene. He can now carry out the
repair work by remotely controlling the robot. Felix can perform mechanical
fixes, use manual tools, operate hand valves, all using his own hands, because
the arms and hands of the robot are controlled by his joysticks or augmented
gloves. These interfaces have been used
already for many years by astronauts and surgeons. In rare cases, Felix orders
spare parts and once they arrive, he sends local engineers to site. These local
engineers are equipped with augmented reality goggles, so Felix can easily
guide them to fix the issues. Felix can even assist them using one of the drones
or robots. He can hold the torch, offer support with heavy tools and give a
high five if the work was done well.
10:15 AM: One
motor shows higher temperatures than expected and some vibration, which cannot
be clearly identified by the Mindsphere analysis software. In more complicated
cases, Felix can do a detailed device analysis. This motor was damaged
thousands of times in the Digital Twin simulation and each time the sensor
signals were recorded to a database.
This data has been used for the damage predictions and FMEA analysis,
but Felix wants to see more. He picks this motor in his augmented reality
application and puts it on his desk. There he can see the simulation model fed
with live data from the real plant. He can disassemble the motor, while the
shaft is moving and see the temperature profile, which is extrapolated from the
measurement data. Superman’s X-ray vision has become a reality!
10:43 AM: The
fleet management system forecasts cloudy weather and medium winds in the
evening in the districts around one of the combined cycle power plants, so it
automatically triggers the prewarming process of the power plant. With the Omnivise
portfolio, the P3000 plant optimization suite improved the load operation and
the load range and with Digital Lifecycle Services (DLS), the alarms and manual
interactions are further reduced. The whole power plant fleet is operated from
the headquarters and no operators have been needed at site for many years. Local
human interaction is only required for safety relevant tasks. The fleet
monitoring system calculates which power plant should be started, which one
should be in hot or in cold standby and triggers the required operation mode
directly. Classical DCS systems have become obsolete and only a single server
at site controls and sends commands to the EDGE-Devices, via ProfiNet.
EDGE-Devices are smart devices like sensors or drives, which can run more
software like control loops and they can connect directly and operate mainly
independently from DCS systems. All subloop controls are operated by the
EDGE-devices themselves, with only superior step sequences or unit control
tasks being carried out by this server.
Since no local control room is required
anymore, the application server is hosted in the cloud, which enables an easy
worldwide plant operation. This makes it easy for Felix to roll out new
features or updates to several sites in one operation.
Since Blockchain transactions has become more mature, the plants are operated
and billed in real time and several power plant owners do not operate the power
plants by themselves anymore. They pay SIEMENS for the operation of the plant
and for all maintenance. The I&C and the DCS systems are no longer
purchased by customers. All services are leased, and payments only occur when
the load requests are fulfilled successfully. The power plant owner buys a
basic package and the various features and optional enhanced services can be
added on top, when they are required. Whenever power plant startups or load
ramps are successful, payments are automatically collected. When the plant is
not in operation, our customers will not have running costs, only standby
costs.
With the lean DCS and DCS in the cloud
approach, which is accompanied by the Omnivise Portfolio and Mindsphere, this
service solution can also be applied to a Non-SIEMENS Fleet and it can be
tailored to the needs of our customers to optimize their assets.
01:35 PM: Felix receives a notification on his watch from the machine learning AI, Siemens Autonomous Machine Learning (SAML), that there is an enhanced algorithm review required and to see his tablet for more information. SAML has been monitoring the start characteristics of multiple power plants and found that the hydraulic systems are far less efficient than the simulation models have predicted. SAML is suggesting some minor changes to the Start Sequence that will shorten start up time and save energy. The Start Sequence is at a security level which cannot be automatically modified by the AI system. SAML has raised a low priority request for Felix to simulate, and then live test, the new enhancements. Felix knows that he must carry out the live tests when demand on the system is at its lowest, so picks up his tablet and reviews the message from SAML. A suitable time slot for carrying out the tests at 4:00pm this evening has already been suggested by the Load Prediction AI. Felix can relax so decides it’s time for a cup of tea.
More about these topics and much more can be found here:
GP SCD GTE C&I: Mid & Long Term Innovation, Simulators and Digital Twins
