QUENCH SIMULATION FOR 16T DIPOLE BUILT AT TEXAS A&M UNIVERSITY Damir Latypov, Peter McIntyre and Weijun Shen Department of Physics, Texas A&M University, College Station, TX 1. QUENCH CODE A 16 Tesla Nb 3 Sn block-coil dual dipole and its prototype are being developed at Texas A&M University [1]. Quench protection system for the magnet is, of course, one of our concerns. Although there is a number of quench codes reported in literature, most of them are written to model a specific magnet. To our knowledge, the only commercially available quench codes are OPUS [2] and QUABER [3]. In OPUS the magnetic field calculations and the quench simulator are incorporated in one finite element package. The local magnetic fields and inductive couplings of the coils are accurately recalculated during the quench. However, being otherwise very general and robust OPUS deals only with the solenoidal magnets. QUABER has only been used at CERN and we have not yet had an opportunity to evaluate it. Therefore to facilitate the design of quench protection we have developed a new code for a magnet of general geometry with an arbitrary number of coils and with the provision for heaters and switches. The algorithm is based on adiabatic assumption which allows to integrate the equation for the temperature distribution in a coil and express it as an implicit function of miits. This function is tabulated for each coil and is used to calculate the peak temperature in the coils. Quench propagation is described in terms of time- dependent longitudinal quench velocities and transversal turn-to-turn quench jump steps. The circuit equations are solved by inverting the circuit inductance matrix and integrating the resulting set of equations by the Runge- Kutta method. The average magnetic fields seen by the coils are calculated using the transfer functions. This algorithm is rather standard and closely resembles the one implemented in QUCERN [4]. However, in implementing it the care has been taken to parametrize the data so that the program can be used for different magnets with little further programming. To check our code we have simulated the coil described in [2]. As with any quench code the results of simulations depend critically on the input assumptions. To make a meaningful test of our program we therefore used the same input data as had been used in [2] for OPUS simulation. We found our results to be in a good agreement with both, OPUS and the experiment. 2. QUENCH RESULTS The prototype 16T dipole is 1 m long, has two bores with three coils and stored energy of 1.03 MJ per bore. A 2D view of the magnet with the field lines is shown in Fig.1. Fig.1. 2D view of the magnet with the field lines. Main coil parameters are summarized in Table 1. The data on cable which will be used for the winding of coils are given in the Table 2. Table 1. Coil Parameters. Coil Outer Middle Inner Op. Current (A) 8650 8650 8650 Peak field (T) 9.45 13.29 16.18 J in Nb 3 Sn (A/mm 2 ) 2300 1462 592 Turns per bore 102 96 50 Inductance per bore (mH) 11.496 6.948 1.078 Table 2. Cable Parameters. Cable Outer Middle Inner Strand diameter (mm) 0.6081 0.6950 1.1583 Cu/Sc ratio 2.095 1.244 0.515 Filament size ( m) 6 6 6 Number of strands 40 35 21 Cable thickness (mm) 1.2571 1.4178 2.2749 Cable width (mm) 12.537 12.537 12.537 3446 0-7803-4376-X/98/$10.00  1998 IEEE