|The main airport in Milan, Italy, increased revenue from power generation by €1.5 million (US$1.97 million) in less than a year|
A 68 MW tri-generation power plant produces electricity, heat and chilled water for the main airport in Milan, Italy. Part of the electricity is sold to third parties through the national grid, while heat and chilled water are only used inside the airport. Most of the electricity produced is, of course, used to feed the energy needs of the airport and its facilities.
A large airport is an excellent recipient for the combined generation of electricity and heat, as it guarantees that the supply will be absorbed with a high level of continuity, day and night, all year round. For the Milan airport in particular, co-generation is even an important source of revenue, as unused electricity is sold to the market.
|Figure 1. Energy production control dashboard based on an energy model
Source: Inspiring Software International
The simple cogeneration system serving Terminal 1 began operating in October 1998. It had installed power of 32 MW of thermal capacity and 20 MW of electrical capacity. In 2000 followed an expansion of the thermal cooling plant in view of an expected increase in the airport’s thermal cooling load. The project provided for a 25% capacity increase in the production of chilled water, doubling of the heated water storage capacity, and expansion of the heated and chilled water distribution network. In view of the expected growth of the airport, from 2001-2003, construction of a new cogeneration plant (combined-cycle with heat recovery) began, with an installed capacity of 30 MWth and 30 MWe.
At present the plant has a next-generation transformer room, the most powerful absorption refrigeration system for the production of chilled water in Europe, and a roadmap of investments lined up to increase its efficiency and sustainability. For example, two gas turbines were recently replaced with more efficient ones.
This determination to reduce waste and increase the ability to manage all of the plant’s systems on an automated basis led the company to seek new smart and sophisticated control room technologies. Energy management software was one of them.
The cogeneration plant
The plant’s current configuration includes:
- Combined-cycle 1: one 25 MWe gas turbine (TGC) and one 5 MWe counter pressure steam turbine (TV4);
- Combined-cycle 2: one 30 MWe gas turbine (TGD) and one 5 MWe condensation steam turbine (TV5);
- One 10 MWe gas turbine (TGA).
Depending on energy requirements, TGD exhaust gases can be conveyed either to two simple recovery boilers of 16 MWth each (if heat demand is higher), or to a steam generator (GVR2) for the generation of additional electricity through the TV5 10 MW condensation steam turbine (if electricity demand is higher).
|Figure 2. Live diagnostic measurement working mechanism
Source: Inspiring Software International
A superheated water production unit inside the GVR2 provides an additional 3 MW of thermal energy. Another steam generator is coupled to the TGC, for thermal power of 30 MWth and combined-cycle production of 30 MW. Exhaust gases from the TGA are conveyed to the recovery boiler (REC A) for thermal production of 16 MWth.
The thermal section is completed by a 22 MWth natural gas-fired ancillary conventional boiler (CB50). The plant’s total thermal power is therefore 87 MW, and its electrical power is 80 MW.
The plant’s automation and supervision system is an essential factor in achieving the operational flexibility required. It has adopted a system with a distributed control system (DCS) architecture. The advanced technology of the supervision and control system allows a high level of plant automation which eliminates the need for manual interventions during ordinary operation and enables plant operation with a very small internal workforce.
Furthermore, the plant is equipped with specialised energy management software (EMS) designed to increase system efficiency.
As in every power plant, the number of sensors and data required in the control room is incredibly high, reaching approximately 2000 signals which are constantly monitored by the DCS system. The control room technicians constantly control alarms and define setup parameters for all the equipment.
Based on the next-day trading energy prices, they decide which combined cycle is better to use each day. The turbines are always set to the nominal load, as this usually means the highest performance. During the day the technicians follow the scheduled plan and manage the electrical and thermal request, selling the planned capacity to the grid.
|Figure 3. Energy production optimisation working mechanism
Source: Inspiring Software International
Even if the DCS was already acquiring a wide range of measurements, the operators wanted a detailed monitoring of the plant’s efficiency, and of the efficiency of each component. With a normal DCS system, it is possible to check whether the system parameters are in a common working range, but it is almost impossible to understand whether the plant is producing energy with the correct efficiency. This becomes a problem if revenues from the energy sold to the grid are decreasing. To increase those revenues it was important to identify system inefficiency as soon as it happened.
Before implementing an energy management software system the operation manager was deciding plant plan based on experience, and trying to find the best plant asset based on the electricity price to the grid.
During system design, it was proposed to use software to simulate plant behaviour when operational conditions vary, in a way that allows identification of the production plan that makes the highest profits possible with defined operation conditions and resource costs. The project’s target was to enable the plant operators to control its efficiency in a real-time way. There were no profitability targets at this stage.
The software is focused on two typical problems for this type of plant:
1. Diagnosis of system malfunction and waste detection, i.e., identifying anomalies in system behaviour and finding the root cause.
This problem was solved with a Live Diagnostic software application based on statistical deviation control of the energy models of the plant’s components. An energy model is a transfer function that allows prediction of each system component’s expected output values based on working conditions and operational input. Due to the complexity of the system, both physical laws and statistical regression models are analyzed. Process control techniques are used to compare the actual system status with the theoretical reference status obtained by the modelling process. Any deviation from these two patterns identifies an ‘out of control’ status. Additionally, a system of diagnostic reports has been structured to allow identification of the exact cause of the deviations. This resulted in a 3% reduction in annual overall equipment maintenance costs.
The Live Diagnostic Measurement uses asset parameters monitored by the DCS system to calculate efficiency and performance value for the system. This allows the user to instantly check working parameters such as electrical and thermal power produced, consumed by the airport and sold to the grid, but also to constantly check system and component yields. This creates an efficiency monitoring system, but the analysis has been made easier by creating an energy model for the main components. In this way it is not only possible to check plant performance, but there is a theoretical value that defines how much the system should produce at each moment based on operating conditions and system set-points.
2. Daily production plan design, i.e., finding the production plan that guarantees the highest possible profit, considering the thermal and electrical needs of the airport and offering the possibility of selling the extra electrical capacity to the grid.
For this problem energy models are used to predict plant production and system revenues at the projected operational requests and economic conditions. The Energy Management Software optimises the hourly system setup parameters that maximise profits (the difference between revenue from energy trading and fuel costs), respecting all operational constraints. Next-Day hourly electricity trading prices are decided by GME (the Italian Energy Trading Institution). Prices are input into the software in the same way as any other operating condition. This provides a daily plant optimiser, rather than a parameter optimiser. An essential part of this plant optimiser is the hybrid optimisation algorithm, which decides the hourly setup of the plant’s seven pieces of equipment (two gas turbines, one steam turbine, two recovery boilers, one auxiliary boiler, and one four-tank heat accumulator system) by considering 16,384 possible operation scenarios per hour. Finally, thanks to research and methodologies for reducing situational complexity, the software achieved the target of ten minutes for calculation time with a standard PC. Therefore, the application effectively allows the plant operator to select and verify the most convenient and profitable daily production plan in a significantly reduced amount of time. As a result, the Milan airport achieved €1.5 million in additional profit, or an improvement of 10%.
To build this solution, the software house used the existing DCS to acquire the main operating values of the plant. The remaining parameters related to the gas turbine were acquired by directly connecting to the equipment control system.
Production models for the equipment were created based on datasheet information as well as physical and statistical models. After a pre-validation, the system was tested to verify its correspondence to the customer’s needs and the accuracy of its forecasted values.
The remote-management capability of the system makes it possible to also control the system status from an external location, after secure validation access. This was used in the beginning to check the system daily for possible errors, and then to offer remote support for plant efficiency management and help operators during the optimisation process.
The system is now updated to follow equipment changes or big changes in plant performance. A periodic reporting service allows the customer to have a guide in energy-based maintenance analysis.
Results and future actions
The first result is a system that can monitor and control plant efficiency in real time. Visualisation of the values is in the same place as the pre-existing SCADA monitors.
As an economic result, €1.2 million in higher profits was calculated as the difference before and after system installation. This result is mainly due to the daily production plan optimisation, as the system can combine the operator’s skill with a simulation system that calculates profits for every possible scenario. Therefore, the system also helped to increase system knowledge, making the effects of every change in system management parameters easy to understand.
Furthermore, Live Diagnostics Monitoring and periodic reporting helped in defining loss in system performance, and in defining improvement after equipment changes. For example, it was possible to estimate the advantages due to the renewal of the compressor part of the gas turbine. Thanks to the system, it was also possible to identify an initial loss in performance for one of the steam generators, anticipating maintenance action to the first cycle stop and avoiding potential overconsumption of the system, estimated at €10,000 per month.
Because of the success of this project, the company is commissioning the same system at the 24 MW cogeneration plant operating at Milan’s other airport.
Aleksandra Peneva is Technical Sales Manager at Inspiring Software International.