By: G-D. Wolff, S. Lauxtermann & R. Kumar, ABB
Desalination plants play an essential role in power and water production in the Middle East to meet ever increasing and dynamic demands. In particular, the number of hybrid desalination plants is on the increase, largely due to their flexibility in meeting different levels and combinations of production. But these plants have a complex system structure, especially considering that a hybrid plant uses at least two desalination processes. Nevertheless this structure is a hotbed of optimization possibilities.
One such hybrid desalination plant is located 20 km north of the city of Fujairah on the coast of the Gulf of Oman. Stringent cost pressure as a result of privatization meant Fujairah’s operators needed to find optimization initiatives to reduce production costs. Subsequently, Fujairah Water and Power Plant (FWPP) installed different optimization packages from ABB.
Fujairah Water and Power plant has installed ABB’s optimization packages
Power generation at the plant comes from four GE 106 MW PG9171E gas turbines (GTs) with associated heat recovery steam generators (HRSGs) and two Siemens NG90/90 119 MW steam turbines. Water production is realized with five multi-stage flash (MSF) distillers, each with a capacity of 12.5 million gallons per day (MIGD), and one two-stage reverse osmosis (RO) plant with a 37.5 MIGD capacity. The plant’s gross power capacity is 660 MW, and water production is 100 MIGD at 46 °C ambient temperature. The HRSGs feed high-pressure steam to a common header.
Apart from capital costs, fuel is by far the biggest expenditure in power and desalination plants (90 per cent), so the key to increased savings lies in optimizing fuel use. Besides optimizing fuel consumption, these tools also allow for maintenance and workflow improvements in the overall work process.
The optimization solution undertaken at Fujairah comprises:
- Load scheduling
- Hybrid optimization
- MSF optimization
- Process optimization
Loads must be provided by the water and power plants and vary during the day. This is especially true in regions in which the humidity and temperature vary, and it is those that affect the load. Figure 1 shows the example of a daily electrical load curve from the FWPP. The plant operates at 50 to 80 per cent of its net capacity. A 20 per cent change, however, is considered normal. A load change of more than 150 MW is equivalent to about 1.5 times the maximum capacity of one GT.
Figure 1: A daily electrical load curve from the Fujairah Water and Power plant
The daily water demand is prescribed to the plant. Additional flexibility for plant production exists if the storage capacity of the potable water tanks is taken into account.
The aim of a load scheduling system is to find the optimum combination of plant components to satisfy particular electrical and water export requirements. In other words, on the basis of the demands made by load dispatchers, the plant structure, fuel prices, variable maintenance costs, chemical costs, ambient conditions and operation modes, entire plant and individual unit performance calculations are carried out.
The result is a proposed cost optimal plant operation based on a particular combination of GT, ST, RO and MSF units. In fact, the main benefits of optimization are realized by finding the best combination of GT and ST production, the best combination of ST production and bypass steam flow, the best combination of MSF and RO production and the use of the water storage possibilities.
Suitable combinations enable plant optimization to remain valid from one to several days. The optimization software package comprises four sections; a graphical user interface (GUI); a kernel, which coordinates the GUI, optimizer and database; an optimizer (CPLEX), which is tasked with finding the global minimum; and an Oracle database to store all configurations and results.
In the Fujairah plant, using specific test loads and test conditions, an average of 2.7 per cent of the fuel costs can be saved by implementing this optimization software. The highest savings are realized in low-load operation modes, where a potential of six per cent or more could be achieved.
Potable water from the plant has to be of a defined quality. But water quality from the processes differ. Water from the MSFs is almost demineralized, yet water from the ROs is of a high quality, so the mixed potable water from the two processes still has to be mineralized by the potable water plant to reach the optimum mineralization criteria stipulated by the health authorities.
The RO plant comprises two lines, each with two passes. The salinity of the first-pass output is about 500 parts per million (ppm) compared with the 15 ppm of the second pass. Some water bypasses the second pass and is mixed at its output. The water from the RO plant before optimization achieved a salinity of 80-100 ppm.
The aim of optimization is to find the minimum number of second-path racks needed to meet the guaranteed water quality of the total plant. This in turn reduces electrical consumption for second-path rack pumps by 0.5 MW per pump. It also cuts maintenance costs for second-path racks and chemical costs in the potable water plant.
Also, reducing the number of running racks in the second pass raises RO plant water production because each second pass rejects about 10 per cent of the water. Hybrid optimization in the Fujairah plant saves the equivalent of 0.6 per cent of the total fuel costs.
The main operational costs in an MSF plant are incurred in energy input by steam, chemical additives and the electricity consumed by plant equipment. The job of the MSF optimizer is to minimize the sum of these costs by calculating other set-point values that will keep water production constant. Figure 2 shows a typical cost curve with varying top brine temperature (TBT) values.
Figure 2: A typical cost curve with varying top brine temperature values
The steam costs per cubic metre of distillate fall with higher TBT because the performance ratio rises if water production is to be kept constant. The chemical costs (e.g. anti-scaling) per cubic metre of distillate increase with a higher TBT because of higher scaling at higher temperatures. The optimizer calculates optimized values for TBT, brine recycle flow, seawater to reject temperature, seawater to reject flow, and make up flow because at a given load, different combinations of them can be used.
Additionally, a process simulation package is used, which is capable of modeling MSF units down to the stage level. A model of the MSF line is configured by combining calculated optimized set points for the above mentioned parameters with other components, such as a brine heater and pumps. The set points determined by the optimization package are then used by the operator for MSF control. Using the MSF optimization tool under various operation conditions has saved up to 1.78 per cent of total MSF production costs.
It is essential to monitor the performance of the different plant areas to detect any unusual drop in efficiency.
The FWPP carries out performance calculations for all major equipment, e.g. GTs, HRSGs, STs, feed water pumps, MSF desalination plant, seawater intake pumps, RO plant and HP pumps. Operators and engineers use the data to analyze the performance of the entire plant as well as individual areas.
Comparisons between actual and expected performance in connection with supervisory identifiers make analysis simple and effective. In one case, use of the performance monitoring and optimization system detected a low HRSG performance.
An investigation showed that the problem lay with the FD-Fan. To be more precise, the FD-Fan is started only when the average flue gas temperature behind the duct burners exceeds a control set point, set at 800°C in the FWPP. But design criteria required a value of 840°C, and the 40°C difference translated into a 1.7 per cent drop in boiler efficiency during FD-Fan operation. Also, the fans never stopped once they were started because the lower set-point temperature of 700°C was deemed too low (the temperature rarely fell below this value). The operation practice was changed after the implementation of the software package, and the result since then has been increased efficiency.
Work process optimization
In addition to process and operation optimization, areas of the work process can also be optimized and improved. These include the automatic creation of logs and reports, which saves working time and avoids manual entry errors, and the automatic exchange of data with other systems, for example a computerized maintenance management system.
In Fujairah, more than 100 logs and reports have been automated, resulting in an estimated daily saving of about 18 hours of work. Easy-to-use tools to configure different types of reports are developed by ABB using Microsoft Excel software in the system.
significant savings achieved
The proven installation of a performance monitoring and optimization system in the Fujairah water and power plant sums up the effectiveness of modern optimization techniques in power plants. In fact, most of the optimization techniques described can also be used in power and desalination plants with a non-hybrid structure.
Altogether, more than four per cent of the total fuel consumption is saved in Fujairah, with additional savings attributed to work process optimization.
This article was first published in the September 2007 issue of Middle East Energy (p.11-12).