THE UNIVERSITY OF MANCHESTSER
SCHOOL OF CHEMICAL ENGINEERING AND ANALYTICAL SCIENCES
CAPD Coursework 2
Zhibek Makhtayeva 9795616
Introduction
In the world with increasing demand of energy, it becomes more important to optimize the existing energy and utility systems. For example, heat exchange network development from the embryonic stage or retrofit of current systems are one of the challenges of chemical engineering nowadays. Heat exchange network design might have various objective functions, such as minimum utility costs, minimum investment costs or minimization of the number of possible matches in the configuration. To address these challenges, optimisation tools and techniques can be applied. The problem needs to be decomposed …show more content…
There are 11 heat exchangers. Several streams are splitted, except hot stream 2 and cooling water. The total area of heat exchangers is 494.366 m2.
Figure 5. Heat exchange network (vertical heat transfer)
Slack variables were calculated by GAMS, the values can be found in Table 3. As it can be observed, values are small, but it is essential to add slack variables in order to optimise the objective function in GAMS.
Table 3. Values of slack variables
〖S1〗_(i,k)
0 kW 〖S2〗_(i,k)
0.004 kW 〖S3〗_(j,k)
0 kW 〖S4〗_(j,k)
0.004 kW
The heat loads of 11 heat exchangers are shown in Table 4. As it was mentioned in Introduction section, Qhot oil= 15 kW, Qcooling water = 534 kW (SPRINT results) for vertical heat exchange should be the same, because utilities and temperature boundaries of the blocks remain constant.
Table 4. Heat loads of generated heat exchange matches Block 1 (kW) Block 2 (kW) Block 3 (kW)
H1. C1 37.244
H1.C2 254.760
H1.C3 271.996 44.000 4.000
H1.QW 15.000
H2.C1 42.756
H2.C2 277.240
H2.C3 40.000
QS.C1 519.996
QS.C3 14.004
QW – cooling water, QS – hot oil
b) Criss-cross heat …show more content…
Criss-cross heat transfer has one more heat exchanger, but the heat load on it is 0.004 kW, whereas the heat loads on both utilities increase by 0.004 kW. The areas do not differ significantly , the difference is 4.104 m2 , which is generated due to the additional heat exchanger. Moreover, for criss-cross heat transfer, heat recovery is 971.996 kW, which is a lower value than for vertical heat transfer, because slack variables were compensated by the use of utilities. Therofore, the total utility cost is approximately the same, whereas the annualised capital cost difference is about 300 $/yr. The heat load compensated by utilities in the basence of slack variables can be considered as a round-off and added to constraint equation of energy balances of hot and cold streams. In that case, both types of heat transfer models will have the same design, which means that the vertical heat transfer model is the optimal solution for the minimisation of total annualised cost. Moreover, the most contribution comes from the minimisation of utilities, taking into consideration that heat transfer was allowed across the
SCHOOL OF CHEMICAL ENGINEERING AND ANALYTICAL SCIENCES
CAPD Coursework 2
Zhibek Makhtayeva 9795616
Introduction
In the world with increasing demand of energy, it becomes more important to optimize the existing energy and utility systems. For example, heat exchange network development from the embryonic stage or retrofit of current systems are one of the challenges of chemical engineering nowadays. Heat exchange network design might have various objective functions, such as minimum utility costs, minimum investment costs or minimization of the number of possible matches in the configuration. To address these challenges, optimisation tools and techniques can be applied. The problem needs to be decomposed …show more content…
There are 11 heat exchangers. Several streams are splitted, except hot stream 2 and cooling water. The total area of heat exchangers is 494.366 m2.
Figure 5. Heat exchange network (vertical heat transfer)
Slack variables were calculated by GAMS, the values can be found in Table 3. As it can be observed, values are small, but it is essential to add slack variables in order to optimise the objective function in GAMS.
Table 3. Values of slack variables
〖S1〗_(i,k)
0 kW 〖S2〗_(i,k)
0.004 kW 〖S3〗_(j,k)
0 kW 〖S4〗_(j,k)
0.004 kW
The heat loads of 11 heat exchangers are shown in Table 4. As it was mentioned in Introduction section, Qhot oil= 15 kW, Qcooling water = 534 kW (SPRINT results) for vertical heat exchange should be the same, because utilities and temperature boundaries of the blocks remain constant.
Table 4. Heat loads of generated heat exchange matches Block 1 (kW) Block 2 (kW) Block 3 (kW)
H1. C1 37.244
H1.C2 254.760
H1.C3 271.996 44.000 4.000
H1.QW 15.000
H2.C1 42.756
H2.C2 277.240
H2.C3 40.000
QS.C1 519.996
QS.C3 14.004
QW – cooling water, QS – hot oil
b) Criss-cross heat …show more content…
Criss-cross heat transfer has one more heat exchanger, but the heat load on it is 0.004 kW, whereas the heat loads on both utilities increase by 0.004 kW. The areas do not differ significantly , the difference is 4.104 m2 , which is generated due to the additional heat exchanger. Moreover, for criss-cross heat transfer, heat recovery is 971.996 kW, which is a lower value than for vertical heat transfer, because slack variables were compensated by the use of utilities. Therofore, the total utility cost is approximately the same, whereas the annualised capital cost difference is about 300 $/yr. The heat load compensated by utilities in the basence of slack variables can be considered as a round-off and added to constraint equation of energy balances of hot and cold streams. In that case, both types of heat transfer models will have the same design, which means that the vertical heat transfer model is the optimal solution for the minimisation of total annualised cost. Moreover, the most contribution comes from the minimisation of utilities, taking into consideration that heat transfer was allowed across the