Process Integration of a Dynamic Industrial System Raquel Durana Moita1, Henrique A. Matos1,*, Cristina Fernandes1, Clemente Pedro Nunes1, Jorge Prior2 and Diogo Almeida Santos2

1 Instituto Superior Técnico –DEQ, Av. Rovisco Pais, 1049-001 Lisboa, Portugal 2 Quimigal, Quinta da Indústria, 3864-755 Estarreja, Portugal

Abstract The aim of this study is to obtain a dynamic modelling and simulation of a Portuguese industrial integrated system composed of three different processes. This work includes the analysis of the optimal integration of the different units, and the study of the effect of some operational and atmospheric conditions on the system to maximize its global thermal efficiency. The cogeneration system was modelled and analysed using the GateCycle 5.34.0.r software. It was concluded that the electric and the thermal power obtained strongly depend on air and economizer cooling water temperatures. The whole integrated process (cogeneration, plate exchangers and salt production unit) is simulated and exploited through gPROMS 2.1.1. The best start-up conditions were established. The minimum number of ponds required strongly depends on atmospheric conditions, but it can never be less than three in order to obey the operational defined temperature intervals. The scheduling of the evaporation ponds to be put into operation is also investigated in order to enhance the salt production and to optimise the salt harvesting. The simulation indicates that it is better to have the minimum number of ponds working (higher temperatures inside the ponds) and to have a “turbo” pond that receives a larger quantity of heated brine than the others. It was also studied the effect of different atmospheric conditions, and the number of ponds in service required to overcome the more adverse atmospheric conditions. The global process efficiency (thermal and electric power over natural gas consumption) is approximately 92%. However, the effective global thermal efficiency of the whole integrated site (accounting for the existing losses into the open air of the evaporation process) is in the range of 75- 80%, depending on the atmospheric and operational conditions considered. Keywords: Dynamic modelling, optimisation, process integration, industrial case study 1. Introduction The main goal of this study is to identify the best operational conditions of an industrial integrated system of three processes, in order to optimise the heat and raw materials usage that permits to maximise its global and energy efficiency and to minimize the

* Author to whom correspondence should be addressed : henrimatos@ist.utl.pt

environmental impacts. This industrial process is located at Carriço, Pombal (Portugal) and includes a natural gas storage that was built in salt caverns from Transgás, a gas turbine cogeneration system from Galp Power and a salt recrystallization process from Renoeste. The system of the three processes considered as separated units is neither efficient nor feasible. The integration of these independent units, as shown in figure 1, improves the global process efficiency. The cogeneration system provides the electrical power to export to the regional net and to satisfy the process needs, and its associated thermal energy is used in the salt recrystallization process improving the global energy efficiency. The leaching programme to construct the caverns for future Natural Gas storage generates the brine needed to feed the salt recrystallisation ponds, minimizing the environmental damage.

Mineral Salt

Caverns

Water

Fresh Brine

Cavern Construction Power Generation Salt Crystallization

PondsSalt

Solar Energy

QuimigalNatural Gas

Gas Turbine

Air

Electric Energy

Exhaust Gases (~ 500 ºC)

Plate ExchangersHeated Brine

Cold Brine

EconomizerExit Gases

Cold WaterHeated Water

(~ 90 ºC)

Ocean

Mineral Salt

Caverns

Mineral Salt

Caverns

Water

Fresh Brine

Cavern Construction Power Generation Salt CrystallizationCavern Construction Power Generation Salt Crystallization

PondsPondsSalt

Solar Energy

QuimigalNatural Gas

Gas TurbineGas Turbine

Air

Electric Energy

Exhaust Gases (~ 500 ºC)

Plate ExchangersPlate ExchangersHeated Brine

Cold Brine

EconomizerEconomizerExit Gases

Cold WaterHeated Water

(~ 90 ºC)

Ocean

Figure 1. The three process integrated system.

Those caverns will allow a strategic storage for the Portuguese natural gas supply network, providing a buffer for eventual future fluctuations on the national consumption. The pure salt produced by Renoeste is the main raw material to Quimigal’s chemical industrial site at Estarreja (Aveiro).

2. Framework A global model of the whole system was developed to determine the best conditions to maximize the global energy efficiency of the process integration between the Trangás cogeneration unit and the salt (NaCl) production process by Renoeste. The framework developed includes mainly the following procedure and tools: i. Modelling and simulation of the cogeneration system through GateCycle 5.34.0.r of

the GE Enter Software; ii. Determination of a thermal power correlation, through TableCurve 3D from SPSS

Inc., using data from cogeneration system; iii. Creation of a dynamic distributed model of the whole integrated system

(cogeneration, plate heat exchangers and salt production process) via gPROMS 2.1.1 of the PSE, Ltd.

Using different atmospheric and operational conditions as state variable it is possible to obtain as output of the model, among other variables, the pond level, the brine temperature and concentration profiles, the salt production and the process thermal efficiency.

3. Analysis of the Cogeneration System Behaviour The cogeneration system installed at the industrial site consists of a Rolls-Royce natural gas turbine (RB211T DLE) and its associated economizer. The economizer is subdivided into 4 internal set of tubes, each having a superficial area of 2302.5 m2, in which circulates the water that will be heated with the exhaust turbine gases (leaving the turbine at approximately 500ºC). Through modelling in GateCycle 5.34.0.r, of the GE Enter Software, it was possible to study the influence of some variables, such as air and natural gas temperatures or the water circulation conditions, on the gas turbine and economizer performance. The main conclusion of this analysis was that air temperature strongly influences the cogeneration system response: an increase of 10ºC in Tair causes a reduction of almost 10% in the electric power and around 3% in the thermal power (TP). The thermal power available is also very dependent on the temperature of the cold water entering in the economizer (T1): an increase of 10ºC in T1 leads to a reduction of around 2% in the TP. Furthermore, its maximum value is reached at Tair around 15ºC, with the smallest working T1 value. It was also concluded that air humidity does not have a significant effect on the system and the pre-heating of the natural gas slightly increases the turbine efficiency in the temperature range used at the site (~35 ºC). The advisable operational temperature intervals of the circulation water in the economizer at its entrance and exit are, respectively: T1 = 65 ± 5 ºC and T2 = 90 ± 5 ºC. 4. Dynamic Modelling and Simulation of the Integrated System The integrated system includes the cogeneration system, the four plate heat exchangers and the salt production unit. This unit involves mainly the recrystallization ponds, a feed tank and a collecting channel, as illustrated in figure 2.

Economizer

Cold Brine, Ttank

Cold Water, T1Heated Water, T2

Exhaust Gases, Tg Exit Gases, Tgs

Plate Exchanger

Heated Brine, TEntP

Pond 1

Cogeneration System

Plate Exchanger

Plate Exchanger

Plate Exchanger

(RB211T DLE )

Air, Tair

Nat. Gas, Tng

Pond 6

…

Cha

nnel

Tank Fresh Treated Brine, TF

Purge, Tpurge

TExitP

Tchannel

Economizer

Cold Brine, Ttank

Cold Water, T1Heated Water, T2

Exhaust Gases, Tg Exit Gases, Tgs

Plate Exchanger

Heated Brine, TEntP

Pond 1

Cogeneration System

Plate Exchanger

Plate Exchanger

Plate Exchanger

(RB211T DLE )(RB211T DLE )

Air, Tair

Nat. Gas, Tng

Pond 6

…

Cha

nnel

Tank Fresh Treated Brine, TF

Purge, Tpurge

TExitP

Tchannel

Figure 2. Integrated process: cogeneration, plate exchangers and salt production unit.

The plate heat exchangers (PHE) set was included in the model using its design equation and its heat balance. The model of the integrated process was built through algebraic and differential equations taking into account the phase equilibrium thermodynamics: solid, liquid and gas equilibrium. The mass and heat balances are axially distributed providing the expected profile inside the ponds.

The water evaporation rate is a function of the salt concentration, the temperature in the brine solution and of the atmospheric conditions (air temperature, humidity and wind

velocity), corresponding to the mixed laminar-turbulent flow (Sartori, 2000; 1991). The contribution of the forced convection loss also corresponds to the same flow regime, and depends on the atmospheric conditions. The radiation energy loss is affected by the air and brine temperatures values (Incropera and DeWitt, 2001). It was also included the thermal power correlation obtained through the simulated values of the cogeneration system, which was reduced by 3% due to general energy losses in the pathway. The solar energy contribution absorbed through the brine in the recrystallization ponds is the diffuse part and a fraction of the direct solar energy, which is dependent of the brine ponds level. The dynamic model was implemented in gPROMS 2.1.1, of the Process System Enterprise, Ltd, according to the structure illustrated in figure 3. The total number of model variables depends on the number of discretisation intervals used in the axial distributed domain. For 20 intervals its value is approximately 5 700, while for 60 intervals is around 12 000.

Model Recrystallization Unit

Model Channel

Model TankModel Heating [Q, X, T, ρ ] tank

Atmospheric Conditions

Model Pond i, i=1,..,6

Model FlowSeparation

AtmosphericConditions

[Q, X, T, ρ ] HE

TP=f (T1, Tair)from GateCycle

[Q, X, T, ρ ] channel

Purge

[Q, X, T, ρ ] purge

Fresh Brine

[Q, X, T, ρ ] F

[Q, X, T, ρ ] EntP i…

[Q, X, T, ρ ] ExitP i

…

Model Recrystallization Unit

Model Channel

Model TankModel Heating [Q, X, T, ρ ] tank

Atmospheric Conditions

Model Pond i, i=1,..,6

Model FlowSeparation

AtmosphericConditions

[Q, X, T, ρ ] HE

TP=f (T1, Tair)from GateCycle

[Q, X, T, ρ ] channel

Purge

[Q, X, T, ρ ] purge

Fresh Brine

[Q, X, T, ρ ] F

[Q, X, T, ρ ] EntP i…

[Q, X, T, ρ ] ExitP i

…

Figure 3. Structure of the dynamic model of the integrated process in gPROMS 2.1.1.

Among all the model variables, it is important to identify the decision and the state variables. The decision variables are the ones that allow verifying if the system response is within its region of working feasibility. In this case, to respect the operational intervals of the temperature of the circulation water, the model decision variables would be either the heated water temperature T2 (with T2 = 90 ± 5ºC) or the brine temperature in the colleting channel (since it can be correlated with T1 and T2). The state variables are independent variables with a considerable influence on the system behaviour, and should be carefully analysed. So, for the integrated process the main state variables are: the number of ponds in service, the atmospheric conditions, the fresh brine flow rate, the initial brine conditions, the number of plate exchangers working and the flow rate of the brine pumped into the plate heat exchangers. 4.1. Analysis of scenarios in some special conditions Since it was necessary to have an operational plan to start-up as fast as possible, several studies were made to determine the best start-up conditions: minimum number

of ponds and brine level. From the solution of the model, it was concluded that the atmospheric conditions have a very strong influence on the system response since they modify significantly the water evaporation rate. Among several possible atmospheric conditions, it was concluded that it is not possible to start the reception of thermal energy from the cogeneration system with less than 3 ponds in service, to respect the operational defined temperature intervals. The brine level only influences the system response in the beginning, since it led to the same steady state conditions. The flow rate of fresh brine allowed is also very dependent on the evaporation rate. To address the main objective of this work several possible scenarios scheduling of working ponds have been analysed. Thus, four different scenarios were evaluated. In the first scenario, there are four ponds in service receiving the thermal power through an identical flow rate of heated brine. In the second scenario there are also four ponds working, however the first pond is a “turbo” because it receives a larger quantity of heated brine. The third scenario considers only three working ponds, with one of them as a “turbo” pond. In the last scenario there are also three simultaneously working ponds, the first one is a “turbo”, however it involves four ponds. This means that the total time interval was subdivided into three time periods, in the first time period ponds 1, 2 and 3 are working, in the second period ponds 1, 2 and 4, and in the third ponds 1, 3 and 4. The average atmospheric conditions considered were: Tair = 12ºC, Vwind = 4.5 m/s, Humidity = 82% and Solar energy = 2.3 MW. For all scenarios, the temperatures of the water circulating in the economizer are within their defined operational intervals. It was observed that more water evaporates in the scenario where three ponds are working rather than four. Having a pond that receives a larger quantity of heated brine than the others is even a more favourable scenario. The process efficiency is not increasing changing the ponds working, as it was done in the fourth scenario. Amongst all scenarios, the more efficient scheduling procedure was the third one since it enhances and accelerates the salt production (3.5 times more). Figure 4 shows the variation of the temperature of the water circulating in the economizer (T1 and T2) for two different atmospheric cases.

Figure 4. Variation of the circulation water temperatures for two atmospheric cases, with the number of ponds indicated ( - feasible working region).

In the first case study Tair = 20.7 ºC, Humidity = 41.7 % and Vwind = 1.5 m/s. The second case study has the same values for the air temperature and humidity, but a wind velocity of 2.0 m/s. In both cases, four ponds are in service (receiving heated brine) while the fifth is filled up, then five ponds are in service while the sixth is filled up, and finally all the six ponds are working. From this figure it is possible to analyse the strong influence of the wind velocity on the system and to conclude that for the most favourable atmospheric conditions (Vwind = 2 m/s) it is only required the use of four ponds. However, when the wind velocity is smaller, and therefore the water evaporation rate is reduced, it becomes necessary to use at least five ponds, to obey the operational defined temperature intervals. The global efficiency (thermal and electrical power over natural gas consumption) of the integrated process is approximately 92%. However, the simulations showed that the effective global efficiency (EGE) is in the range of 75 - 80%. There is a significant difference in these two efficiency values, since the last one takes into account the existing energy losses due to radiation and convection into the open air, during the recrystallization process. The value of EGE is particularly dependent on the atmospheric and operational conditions considered, as observed by simulation of the whole process. 5. Conclusions The Process Integration of three Portuguese industrial processes was analysed in this paper using for that purpose a dynamic model and simulation tools. First it was studied the influence of some variables on the cogeneration system response in GateCycle 5.34.0.r. Then, a dynamic model of the whole integrated process was developed and simulated in gPROMS 2.1.1. Different scenarios were explored with the purpose of maximizing the global process efficiencies, by analysing the influence of several atmospheric and operational conditions on the integrated system. The advantages of the existence of a model that can simulate the entire site, is clearly shown by the various advices given to the industrial executives in order to overcome some adverse atmospheric conditions to achieve the highest possible efficiencies. Furthermore, the better understanding of the integrated system using the Process Simulation of possible scenarios, allows the definition of an improved set of operational conditions to obtain a long-term profitable business. Acknowledgements The authors gratefully acknowledge financial support from the Portuguese National Team on Process Integration (GNIP- Grupo Nacional para a Integração de Processos). References Sartori E., 2000, Solar Energy, 68, nº 1, 77-89. Sartori E., 1991, In Proc. ISES Solar World Congress, Denver, USA, 2347-2351. Incropera, F. P., DeWitt, D. P., 2001, Fundamentals of heat and mass transfer, 5Th

edition, John Wiley & Sons, New York.