There are two options.
First Option
You can connect your "console apps" to the signalR backplane, and these apps will be able to call client's methods with something like
var hub = GlobalHost.ConnectionManager.GetHubContext<MyHub>();
hub.Clients.All.doSomething("blah-blah");
You can use any port for the signalR config but same backplane configuration and same Hubs. You can use Owin for self hosting. Nobody will connect to these console apps.
Disadvantages:
1) Your console applications needs to reference SignalR.SelfHosting and your hubs classes. This is not good from architecture point of view.
2) Your console apps will listen some ports (since they're signalR servers). Theoretically somebody could connect to this hubs and do something. This is not good neither from architecture point of view no from security.
Second Option
You can implement Hub method like "PushSomeDataToClients" and call it from your console apps using SignalR client as mentioned in @bartbje comment.
Pros: no disadvantages from the first option.
Cons: You need to implement some security stuff to prevent anybody outside the system to call this method. SignalR has a lot of stuff to do this, so just google. For example you can create separate Hub for inter-system communications.
Third Option
Interact with web-server apps with another way then SignalR. Probably you're already using something like rabbitMq or any type of service bus. Also you can implement it using separate ApiController at your web-server app. But it seems to be closer to the second option.
Due to me, I would probably choose the third option since it's clean from architecture point of view.