Re: Boiler plans - Reliable Steam Engine Co
Posted: Fri Aug 07, 2015 4:35 pm
"artificial intelligence", I've been employing it for years, it's nothing to fear...
Note that the concept of artifical intellegence is far different than automatic controls, be they electronic or not. Many of the newer "smart" automatic controls can actually "learn" to change various parameters in a control system based on the recorded history within a system. For example, if the burner controls on a fired boiler respond too quickly, and steam pressure overshoots the desired control pressure, the "smart" controller will observe this, and adjust the rate of firing increase to make system response better behaved. The controller constantly observes the response, and adjusts its functions as needed, however this is not artifical intelligence.
Artifical intelligence refers to the ability to actually think things out, and come up with solutions to problems that had not previously been developed, that were not "programmed into" the computer. The issue here is when a machine starts thinking independently, and makes decisions by itself, and figures out how to implement whatever the machine decides is the proper action, we may be in trouble. Programmed interlocks to prevent this are always stated to make things safe and foolproof, however if you look at all the safety programming that has failed miserably, such as Three Mile Island, Chernobyl, and many other machines, one can imagine a fast intellegent computer figuring out how to circumvent the rules as mere "child's play". The movie "2001" is a good example, where "HAL" (IBM minus one letter) decided the mission was more important than the humans.
Note that the concept of artifical intellegence is far different than automatic controls, be they electronic or not. Many of the newer "smart" automatic controls can actually "learn" to change various parameters in a control system based on the recorded history within a system. For example, if the burner controls on a fired boiler respond too quickly, and steam pressure overshoots the desired control pressure, the "smart" controller will observe this, and adjust the rate of firing increase to make system response better behaved. The controller constantly observes the response, and adjusts its functions as needed, however this is not artifical intelligence.
Artifical intelligence refers to the ability to actually think things out, and come up with solutions to problems that had not previously been developed, that were not "programmed into" the computer. The issue here is when a machine starts thinking independently, and makes decisions by itself, and figures out how to implement whatever the machine decides is the proper action, we may be in trouble. Programmed interlocks to prevent this are always stated to make things safe and foolproof, however if you look at all the safety programming that has failed miserably, such as Three Mile Island, Chernobyl, and many other machines, one can imagine a fast intellegent computer figuring out how to circumvent the rules as mere "child's play". The movie "2001" is a good example, where "HAL" (IBM minus one letter) decided the mission was more important than the humans.