Unless you establish a clear understanding of the load profile and the actual significance of reliability during each segment of the load profile, you don't have the slightest chance of arriving at a coherent and useful conclusion. The previous comments indicating the significance of bottoming cycles or cogeneration options should be considered thoroughly as these can dramatically influence the wisdom of some options that you may wish to consider.
Beware of any form of gas turbines if the load profile dictates that it (or they) must operate for significant periods at reduced loading. Gas turbines, particularly combined cycle units, are capable of very high efficiencies, but those high efficiencies are only realized at full load. Their efficiency drops rapidly as the loading is reduced. It is not unusual for a gas turbine operating at synchronous idle to consume fuel at rates in the range of 80% to 90% of their full load fuel consumption rate.
From your posting, it appears that your system could be operating as a modest size "power island." If this is true and you have some proportionately very large loads that must be started, you would be wise to consider including diesel(s) to provide the required nearly instantaneous power increment while using other prime movers for the bulk of the generation capability. Dual-fuel engines (engine running a very lean mix of gas as the main fuel with about 1% diesel fuel serving as the ignition fuel) can serve this same function if the diesel injectors can be used to provide the large instant fuel requirement until the system can absorb and adjust to the large load increment(s). Unless their dynamic effects are properly considered, starting or dropping relatively large loads can cause very serious stability problems for a relatively small system. Unless there are some very unusual aspects to your system, it is most likely that a mix of several different prime movers will need to be included to optimize cost, load vsriation, and reliability needs.
Valuable advice from a professor many years ago: First, design for graceful failure. Everything we build will eventually fail, so we must strive to avoid injuries or secondary damage when that failure occurs. Only then can practicality and economics be properly considered.
