In regards to ice loading on conveyor structures, we take the approach that they are a real load and as such must be accounted for. This means that we consider the ice load when needed on the members that will see the ice load. The ice load is a distributed load to the members that see ice. Just because eventually the load will get to the panel point does not mean that it does not induce bending or catenary tension in the member to get there first. It depends on the structure and location as to the issue of controlling the design or if it will even have any effect on the design. My experience has proven that most of the time (95% or more) it does not change anything. Most typical small box trusses it will not matter. Large gallery trusses may be a different story.
In regards to tension only members on long span box trusses. For a 226' span I am not sure I would be comfortable with tension only plan bracing, at least on the top chords. From a lateral load analysis I see no issue. From a top chord buckling perspective I am not as comfortable with it. It takes a certain stiffness and strength to brace the top chord. Certainly the L3x3x1/4 is probably strong enough in tension but you have to think about the compressive deflection of the top frame. The axial deflection of the top chord will cause the tension angles to go into compression and likely buckle based on your slenderness ratio. To brace the chords the frame would need to drift enough to remove the buckle (compressive stress) from the angle and force a tension load into it. This may be more movement than what you think. On the other hand the bottom chord will also have axial deflection. Remember that the diagonals are attached to the panel points and provide stiffness to the joint. The deflection of the chords due to axial load will induce tension in the diagonals as they try to restrain the joint from moving as well. You may find that the axial load on the diagonals is much higher than what you would anticipate if only accounting for the lateral load. The member may not fail but the connections could. This would all have to be analyzed. I have seen trusses like what you are describing and they seem to work. They may work for a lot of reasons other than what the designer was originally assuming. I do not use tension only bracing in the plan views for these type of structures. I am not personally comfortable with it based on quite of bit of research that I have reviewed. I would guess your chord forces are 400k - 500k based on what little information I have. If you model the truss in a 3D program with the tension only members it will likely give you instabilities on the P-delta analysis. This should get you thinking.
Wind loads are like ice loads and must be accounted for. Wind does not jump around and magically go to the panel point. It blows directly on the exposed members based on the angle of attack, shielding, wind speed, and shape factor. It may be reasonable to assume from an analysis standpoint that the truss is only loaded at the panel point and you may have a safe structure. The reality in the field is that the wind blows on the exposed members. From a practical standpoint we calculate the wind load and distribute it equally to all exposed members on windward and leeward faces of the box truss based on the individual exposed area of each member. It is under loading the windward face and overloading the leeward face. Our justification is that we do not know how much shielding will actually happen and at what level so no matter what we do it's a WAG at best. We just feel like we are in the ballpark on it.
From a code perspective it would not be unreasonable to assume that no code was followed and the structures were hopefully designed by an engineer that made good judgements. If the structures were not stamped by a PE it would not be uncommon that many important features would have been simple overlooked. We see this a lot. It is not uncommon to see companies using mechanical engineers to design these type of conveyors. This is not a statement to start a war about who is smarter in engineering. It is just an observation of mine that I have seen engineers designing structural engineering type structures that are not qualified to do so. It doesn't mean that a mechanical guy can't do it, but they may miss some important design issues from lack of knowledge. Don't ask me to design your huge crusher and I will not ask you to design my long span box truss. Bottom line, code or no code, you need to account for any foreseeable loads that the structure will be exposed to and have enough reserve in the structure to stay in the air and function. This is the requirement of any engineer in any field. It does not matter what you design for or what code is used. If the structure performs over it's life as expected without issues than you have done your job. If it does not then you will be dragged into depositions for three years and deal with whatever comes with a structural failure. IF you follow a code then at least you can argue that you met the standard of minimum care of the time the structure was designed. It may or may not be enough to keep you out of trouble.
