Chemical and Petroleum Engineering, Vol. 39, Nos. 3–4, 2003

COMPRESSORS, PUMPS, REFRIGERATION ENGINEERING PARAMETER OPTIMIZATION FOR A CENTRIFUGAL COMPRESSOR WITH REDUCED INPUT PRESSURE*

O. E. Vasin,1 O. A. Varivoda,1 B. S. Revzin,2 and A. V. Tarasov2

When the output decreases in a multipoint gas-transport system, the major compressor stations (CS) work at reduced inlet pressures. Then the centrifugal pumps (CP) designed to raise the pressure by a factor ε = 1.44–1.5 do not provide a gas pressure at the output close to the calculated or cost-minimum value per unit transportation. It is therefore important to determine the parameters of the interchangeable flow sections (IFS) of the CP to provide an increased degree of compression ε = 1.7 (with the IFS located in the standard body), and it is also important to examine the gas-dynamic characteristics of the IFS used to upgrade CP. Also, the compressors need to be matched to the power turbine (PT) as regards speed. The following criteria are proposed for comparing the gas-dynamic characteristics of IFS: • efficiency under the nominal and deviating conditions of CP operation; • the number of working gas pumps (GP) required to provide the necessary gas flow rate; • the width of the CP working zone, which is estimated from the characteristic shape coefficients [1]; and • matching the speed characteristics of the IFS and the PT providing the drive. Particular interest arises from determining the number of working pumps to provide the required commercial gas flow over the pipe and to match the speed characteristics of the turbines. We constructed load characteristics of these pumps (IFS with gas turbine drive), which relate the pressure p1 at the compressor input, the volume throughput Q1 in m3/min and the commercial throughput Qc in million m3/day, together with the pressure increase factor ε. We used the normal formula for the load characteristics: Ni = GHa /ηme, Ha =

(

)

k zRT1 ε ( k −1) / kη − 1 , k −1

where Ni is the internal power required by the centrifugal pump in kW, G the mass flow rate through the pump in kg/sec, Ha the actual working head in the pump, kJ/kg, T1 the gas temperature at the compressor inlet in K, ηmax and η the polytropic efficiency of the pump, and k = 1.312 the adiabatic index for natural gas containing ~98.5% CH4. *

Read at the 7th International Symposium “Compressor and Compressor Equipment Users and Producers” [in Russian], St. Petersburg (2001).

1

Limited Liability Company Tyumentransgaz. Ural State Technical University – Ural Polytechnical Institute.

2

Translated from Khimicheskoe i Neftegazovoe Mashinostroenie, No. 4, pp. 24–25, April, 2003.

0009-2355/03/0304-0225$25.00 ©2003 Plenum Publishing Corporation

225

Fig. 1. Dependence of the commercial throughput of the aggregate Qc on pressure p1 at the inlet for T1 = 269 K, η = 0.8, and zR = 465 J/(kg·K): a) Ni = 17 MW; b) Ni = 16 MW.

In the determination of the actual head, k has little effect on Ha, and it is therefore not necessary to revise that parameter during the calculation. The mass flow rate and the commercial throughput are related by Qc = (3600·24·G)/ρn, where ρn = 0.682 kg/m3 is the density of the gas under normal conditions. We specify the polytropic efficiency as η = 0.8 and (k – 1)/ (kη) = 0.29 to get Qc =

86400 N i ηme . 0.682Ha

Under the nominal working conditions (zR = 465 J/(kg·K), Ni = 17000 kW, ηme = 0.95, and T1 = 269 K), we have Qc = 3.775(ε0.29 – 1) We thus calculated the dependence of the commercial throughput with ηtot = 0.8 and Ni = 15–17 MW on the inlet pressure. Figure 1 shows load characteristics for Ni = 16 and 17 MW. These curves enable one to determine the number of working pumps required to provide the given flow rate in relation to the inlet pressure and the degree of compression (with various output pressures p2). For example, from Fig. 1 with p1 = 4 MPa, and p2 = 6 MPa (ε = 1.5), one pumping unit can supply 30 million m3/day, and to forward 600 million m3/day requires 20 units. With Ni = 17 MW and ε = 1.7 (outlet pressure p2 = 7 MPa), one unit can provide a commercial flow rate of about 22.7 million m3/day at p1 ≥ 4.1 MPa. These relationships can be used to compare the IFS characteristics with the actual working conditions. An important criterion is the match between the speeds of the turbines. As replacing the flow section of a pump by an IFS with a higher degree of compression does not involve changing the drive, we need to estimate the scope for bringing in the speeds of the turbines together. The specific speed coefficient for the PT and CP is Kn =

226

0.5 n V1 n , 30 H 0.75

Fig. 2. Dependence of efficiency in two-stage CP unit working with one-stage PT on the coefficient Kn with flow coefficient ΦCP = 0.05: 1–5) ΦCP respectively 0.05; 0.1; 0.2; 0.3; 0.4; dot-dash and dashed lines calculated speed coefficients correspondingly for the NK-16 PT and a TsN-16-76-1.44 pump (modification 2).

where n is the speed of the CP or PT rotor, V1 the volume flow rate of the working body at the inlet to the turbine in m3/sec, and H the head set up by the CP (heat transfer from the PT) in J/kg. We now consider a NTs-16-76 compressor providing ε = 1.7. With a nominal consumed power of 16 MW and p1 = 3.7 MPa, it provides a throughput of about 400 m3/min. The speed of the IFS rotor is equal to the nominal speed of the PT one. The specific speed coefficient is then 0.173. For the PT, the NK-16 engine has the analogous parameter of 0.36. For comparison, the standard flow section in the N-16-76-1.44 (modification one) is Kn ≈ 0.21. Figure 2 shows efficiency curves for a unit composed of a two-stage CP with bladed cones (NTs-16-76-1.44) with a one-stage PT and NK-16 motor. The specific speed coefficients of the turbines differ considerably, and KnCP/KnPT ≈ 0.6. As the pressure ratio increases and the volume flow rate decreases, that coefficient for the CP will fall, and the match between the machines will be perturbed. To provide the best match between them, we use the maximum efficiency for the power turbine-centrifugal compressor combination (≥0.74, see Fig. 2), where for the CP one needs to assume a rotor speed higher than the standard one. For an IFS with ε = 1.7 and Ni = 16000 kW, Knopt = 0.2–0.25, and with the above constants for the natural gas, the optimal speed range for the IFS rotor is 7000–8000 rpm. If on the other hand one considers replacement of the gas pump, the technical specification must include matching between the pump and the power turbine not only as regards speed but also as regards specific coefficient. For example, it has been shown [2] that a two-stage pumping turbine in such a unit can implement all possible states of operation and forms of matching to the CP.

REFERENCES 1.

2.

P. N. Zaval’nyi, B. S. Revzin, and A. V. Tarasov, “Optimizing the characteristics of natural gas compressors for gas pipeline compressor stations,” in: Proceedings of the 4th International Symposium “Compressor and Compressor Equipment Users and Producers” [in Russian], St. Petersburg (1998). A. V. Tarasov and B. S. Revzin, “Optimizing a free turbine in a gas transport unit drive,” in: Proceedings of the 46th Session on Gas Turbine Problems [in Russian], Samara (1999).

227