I agree with all of the above comments except for the "common sense" reference by SprinklerDesigner2.
Grid designs are not allowed; looped systems are allowed (with supporting tests or calculations as mentioned above), however, I do not think it is a good idea to use gridded or looped designs for dry systems. In my opinion, NFPA 13 should be altered so that grided and looped systems are not allowed for dry systems. Except for very small systems, the actual trip test times will exceed the maximum trip time required by code. The computer calculations are usually incorrect as Tindey's example suggests. The required trip time (for constant water flow at the most remote point) varies depending on several factors. Refer to Section 7.2.3.6 of NFPA 13 for specific details.
Gridded dry pipe systems are worse than looped dry pipe systems, but both cause problems and lesson the likelihood of a controlled fire.
The problem should be obvious to anyone who has conducted numerous trip tests on these systems and/or or anyone who has a really good understanding of hydraulics. The air pockets constantly shift due to "balancing" of the system pressure in various portions of the grid and/or loop. During trip tests, the water is delayed and then the inspector's test connection has intermittent discharge of water then air then water then air......
This is not a good scenario during a real life fire incident!
But we can save $$$$$ on our material costs and add a little profit for our company..........is it really worth the increased liability and risk? What are the increased costs if a truly qualified inspector conducts a trip test and the system requires 3.5 minutes before CONSTANT water flow reaches the most remote point of the system?? Accelerators and exhausters will not always reduce the trip time enough to mask the real problem (caused by the shifting air pockets). Just use a tree design for dry pipe systems.......I assume most Fire Protection PE's would agree.