The operators of a CHP plant at a large petrochemcial facility in Portugal were struggling to run the plant in the most economic way because of the requirement for high operational flexibility, liquidity of the cost and constantly varying site demands. João Coelho and Pascal Stijns explain how this issue has been successfully resolved.

The Sines petrochemical complex is Repsol’s largest chemical facility in Portugal Credit: Repsol


Repsol is an integrated global energy company with a presence in more than 30 countries. It operates in the upstream areas of exploration and production of hydrocarbons, as well as downstream refining and the production of chemicals, and new energy.

Repsol’s largest chemical facility in Portugal is a petrochemical complex in Sines, which manufactures polymers. The heat and power requirements for this large chemical complex are provided by a cogeneration facility. The CHP plant consists of three high-pressure boilers and one medium-pressure auxiliary boiler, and has a maximum steam production capacity of 600 tonnes, which is used to meet the site’s electricity and heat requirements.

The boilers are able to combust six different fuels of varying quality, availability and cost simultaneously. The steam is reduced and distributed via five steam headers within the site via a 35 MW back-pressure turbine or pressure reducing stations. A 24 MW condensing turbine is also available to produce extra electrical power when needed. The site has two different electrical contracts, of which the condensing turbine contract is the most complex.

Figure 1. A representative screenshot

However, due to the requirement for operational flexibility, liquidity of cost and constantly varying site demands it was proving almost impossible for operations to make the right economical decisions to achieve the optimal utility production cost.

Thus, the Sines facility, working with Honeywell, decided to develop and install an on-line and real-time thermodynamic and economic model that could determine the optimal production settings, and thereby enable operations to run the CHP plant in the most economical way.

The optimization model

The plant’s Honeywell Experion PKS distributed control system (DCS) enabled the use of various Microsoft standard tools such as Task Scheduler, Excel and Visual Basic, which helped to simplify the optimization application. Furthermore, three MicroSoft Excel add-ins are employed: the Honeywell Water and Steam Physical Properties, the FrontSys Premium Solver and the Microsoft Excel Data Exchange. Figure 1 shows a sample optimization window from the application.

Data are read from the DCS into an Excel workbook via Honeywell’s Medex OPC-based add-in. On-line values, pricing information, physical properties, etc, are linked to the model. The Solver add-in executes and inputs the results into the model tab. From there the values are written to defined SCADA points in the DCS for further display, historization, reporting and alarming via OPC.

A copy of the workbook, without the input and output sheets, can be used for off-line optimization too, enabling the user to run various multi periods (i.e. hours, days, weeks, months, years) and analyse ‘what if’ scenarios.

The Solver add-in enables the use of various solving techniques, ranging from Mixed Integer Linear Programming (MILP) to Mixed Integer Quadratic Constraint Programming (MIQCP), as well as the more commonly used, Mixed Integer Non Linear Programming (MINLP).

The objective function of the model represents the sum of the variable and fixed costs, including depreciation, personnel, insurance and fixed charges. The user is able to view the impact of various optimization modes, constraints and loading, including switching devices on or off. The model also provides the operating cost of the CHP in actual mode and optimum mode in a real-time environment.

Positive outcome

Repsol Sines found some surprising results. Running the station with a 1 MW electrical feedwater pump instead of a turbo pump delivered an astonishing saving of more than 9%. Even more surprising was the 13% saving by running the station with two feedwater lines and a turbo pump, compared to running with a single feedwater line and one electrical pump.

Comparing several different operational scenarios before and after the optimizer also proved to be an eye opener, as can be seen in Figure 2. The vertical axis represents the cost (%) relative to the way the CHP plant was operated in the past at low loads.

The difference in Scenarios 1 (first left, red) and 2 (second left, red) is small in terms of cost , although a 2% difference does represent a considerable amount of money on a yearly basis. However, these were rejected by Repsol due to environmental considerations.

Scenarios 3, 4 and 5 (in blue) all comply to Repsol’s sustainability targets (i.e. no flaring and steam venting) and clearly show significant differences in operating cost – close to 22% between operating the power station with only the condensing set at a minimum and one boiler (scenario 3) and scenario 5, i.e. running the station as it used to run.

Figure 2. Model highlights significant operating cost differences

In addition, users are able to justify improvements, including various efficiency improvement projects by changing the layout of the plant and comparing it with previous scenarios over a certain operational time period (i.e. one year).

Because Honeywell had previously provided the control application and equipment for the Repsol Sines utilites, the first off-line model was ready within a couple of months. The on-line model took longer since inputs, such as process measurements, had to be validated, selected and prepared as is the usual case when doing model based power station economical optimization. However, the pay-back period for the whole application, which proved to be couple of months, easily justifying its implementation.

Without both on-line and off-line optimization applications based on open systems, fundamental engineering knowledge and experience, such critical savings would have gone unnoticed. In these harsh economical environments nowadays such savings can mean the difference between profits or lost.

Ing. João Coelho is plant manager utilities at Repsol’s Sines facility in Portugal. Ing. Pascal Stijns is a power & energy consultant at Honeywell Europe, based in Belgium.;

More COSPP Articles
Past COSPP Issues