The American Society of Mechanical Engineers

s spring In some nuclear reactor containment isolation applica­ Greek tions, butterfly valves with pneumatic actuators are used to close pipelines in the event of containment p density pressure transients resulting from postulated acci­ e angular position dents. A model is developed which predicts the closure time history of such valve/actuator systems INTRODUCTION under prescribed transients. Examples are presented which show the importance of the valve and actuator During certain postulated accidents in a nuclear power characteristics. It was found that actuation delays generating station, the pressure inside containment of sufficient magnitude could allow the dynamic torque rises quickly. Emergency core cooling systems (ECCS) to exceed the available actuator torque thus preven­ and containment spray systems (CSS) are activated ting valve closure. within approximately one minute and fifteen minutes, respectively, to remove heat and control the pressure NOMENCLATURE transient. Containment pressure response curves, such as Figure I, are developed thermodynamically with due A piston area consideration of the specific plant configuration. Ap projected disc area Necessarily, various pipelines connect through con­ Cv exhaust or vent valve discharge coefficient tainment and are fitted with containment isolation CT torque coefficient valves which must close to minimize releases to the d shaft diameter environment in the event of an accident. Such valves, D valve diameter their actuators, and control systems are designed for F force fail-safe operation. ;f friction factor equivalent frictional force 50 FV equivalent dynamic force FO spring preload .;;; Jl. spring constant a. 40 '" w· m mass flowrate a:: M equivalent mass of moving parts :> CIl w CONTAINMENT n polytropic coefficient CIl 30 a:: p Q. pressure f~. w 20 ., J AP pressure drop across valve Z t time :::;: TF frictional torque Z ~ TV dynamic torque fz lO t actuation delay 0 d V volume u x linear position 0 10-1 10' Subscripts TIME, SECONDS p pressure side r vent side FIGURE 1 TYPICAL CONTAINMENT PRESSURE RESPONSE Purge valves are a class of containment isolation valves characterized by the fact that they must isolate a compressible gas flow at essentially the pressure difference between containment and the environment. Radiological considerations are used to determine the tolerable maximum closure time of such valves through consideration of total release to the environment. Typically, the valves must be fully closed within 5 to 10 seconds of the initiation of the pressure transient with the actual time being plant specific. In some applications, high-performance butterfly valves (6 to 36 inches) are used in con­ junction with a pneumatic actuator and mechanism which converts rotary shaft motion to linear actuator motion. Both scotch yokes and slider-crank mechanisms are used. The actuator consists of a pneumatic cylinder with piston and piston rod. A coil spring acts on the piston rod, tending to close the attached valve. Pressurization of the pneumatic cylinder compresses the spring and opens the valve. An exhaust valve is located on the pressure side of the pneumatic cylinder. It is held closed by pilot valve pressure which is provided in response to an electrical signal. Removal of the electrical signal at the initiation of the containment pressure transient causes relief of the pilot pressure, allowing the exhaust valve to open. As air escapes through the exhaust valve, the spring tends to close the butterfly valve. Both dynamic torque due to the imposed containment pressure transient causing flow through the butterfly valve, and torque due to seal and bearing friction must, in net, be small enough, if opposing motion, that the spring can close the butterfly valve within the prescribed time. In addition, the dynamic torque may, at times, be in such direction and have such magnitude to assist closure. At other times, however, the dynamic torque may oppose closure. In this case, it becomes very important that sufficient valve closure is achieved before the containment pressure rises to a value which produces enough flow and hence dynamic torque to stall the actuator. In addition, there may be delays in the containment pressure sensing system and pneumatic control systems which, effectively, shift the valve operation schedule further along the containment pressure curve, thus, reducing the stall margin. Typically, analyses of the closure process begin by assuming a valve position vs time relationship. For example, an assumed delay and then linear closure rate over the prescribed window is often used. A table or graph of the valve torque (dynamic and friction) as a function of valve position is then prepared. The torque which can be produced by the actuator at each position is also listed or plotted. The two curves are then compared to determine that the available actuator torque is always greater than the valve torque required. The actuator/valve combination is usually operated without any through flow to verify that the actual delay is less than that assumed and that the stroke time is less than that assumed. It is then argued that such an analysis is conservative and that the valve will close within the required time window. While this may be so in cases where the actuator torque is always considerably greater than the valve torque required, it is not clear that it is adequate when the available torque is only slightly greater than required torque. Furthermore, such analyses give no information regarding the safety margin available in terms of actual valve closure time. In general terms, the actuator available torque known as a function of position whil torque is known as a function of ti assumed time-position relationship allo n. In actuality, however, the actuator dynamically coupled and the escape of actuator governed by thermodynamics. study the closure process in detail, n J equations were developed for the system components anI incorporated into a model which predicted the syste" closure dynamically.