There are a number of articles on the internet as well as one in AMRM (June 2006) suggesting different methods, and the one that I have used is an amalgamation of several. Even so, a bit of experimentation and decision making was necessary before my methodology was finalised.
I suppose the first decision to make was whether to adopt a prefabricated approach where the fence is constructed 'off-site' and then placed on the layout; or to place the individual posts and then fix the simulated chain wire to these. I chose the prefabricated approach because I thought it would be easier to manage the construction as well as less chance of collateral damage to other scenic elements. I accepted that placing two fence 'modules' side by side creates an unrealistic double post arrangement.
The second decision to make concerned materials. Initially, I tried using thin round styrene but after assembling one section, I found that it lacked sufficient rigidity. I settled on 0.8mm dia brass wire for both the posts and the horizontal top and bottom runners. This wire is a little oversized for the vertical posts and significantly oversized for the horizontal members but when it is all in place on the layout, the scale difference is not really visible. Furthermore, it provides a fairly robust frame for later manipulation and placement.
It was also necessary decide on the material to simulate the chain wire mesh. A number of websites suggest the use of a fabric called tulle. Fortunately my wife had a few offcuts so I tried it out on one section. I thought that it looked more like oversized chicken wire rather than the diamond shape of the chain wire mesh. So I opted to use the other common solution, commercial nylon flywire. Again, this is oversized but it gives the overall appearance of a chain wire fence.
For the longer fencing runs, I created modules each of six panels. Other shorter sections of the fence were made to measure. Each module consisted on single top and bottom runners and the appropriate number of vertical posts. I appreciate that many fences replace the bottom runner with a tensioned wire but this is not universal and the lack of a bottom runner reduces the rigidity of the frame. Besides, it's easy to hide the bottom runner behind additional weeds and foliage.
Each length of wire was longer than required so that it could be cut to size later. I notched both the horizontal and vertical members to allow them to sit flush and then each joint was soldered. The basic frame is shown opposite just prior to gluing the mesh to the frame. I found superglue was best for this activity.
Once the glue had dried, surplus mesh was cut away and the wire trimmed to length. I used elasticised cotton to create the three strands of barb wire that frequently tops this type of fence. Often this part of the fence is bent forward and this would be a simple matter to do once the module was completed. However, I preferred to leave the barb wire section in the vertical plane.
Once completed, each module received a coat of grey primer followed by a wash of silver and reddy brown.
Gates were fashioned in a similar manner but with an added diagonal brace on each panel. For my model, each gate was fitted in the open position.
The final photo shows most of the fence and gates in place. The bottom of each post can be pushed into the ground and then glued in place . Once the fixed panels are in place, the gates are glued into their desired position.
In reality, after spending a fair bit of time over the last few month building fences of various types, they almost seem to disappear into the scene as if they had always been there and there is little to show for this labour. I suppose, that is the consequence of 'modelling the ordinary'!
Additional Information
In response to Col's request the following three photos show more distant views of the chain wire fence. The tulle material has been used on the first nine fence panels counting from the left.