I will attempt to clarify the potential value of this by using average combat mechanics, and then see where the multiple warding bonds will provide value.
Overall, your party and the enemy will do average damage of:
(Probability of Hit) x (Average Damage) x (number of rounds)
Note that the Probability of Hit is related to (target AC - Attack Bonus), and is 50% when (target AC - Attack Bonus) = 10.5
The winner of an attrition based conflict (see Lanchester Model of Attrition Warfare) will be the side that can inflict average damage until the enemy is dead, and then have some residual strength of their own. In D&D parlance, that means that:
(Allies HP) / (Enemy Rate of Damage) > (Enemies HP) / (Allies Rate of Damage)
From this, your Warding Bond will not affect your Allies Rate of Damage, or your Enemies HP, so you are working entirely on the left hand side of the inequality. In order to be successful, your dwarf must survive enough damage to provide your allies resistance so that you get a two fold bonus to the inequality (since Enemy Rate of Damage is reduced by 1/2).
In addition, the AC bonus that you provide to those affected with the spell gives them an additional 5% reduction in the enemies probability of damaging them per round. This will further serve to reduce the damage taken on the left hand side of the equation.
The answer to your question lies in comparing all of the parameters that I have listed. I will facilitate this by providing four bounding conditions:
The enemy has vastly more HP than your dwarf, relative to the average rate of damage, and they do damage at a higher rate than your allies. This is the suicidal case you reference.
The enemy has far more HP than your dwarf, but they do damage at a much lower rate.
The enemy has far fewer HP than your dwarf, but they do damage at a much higher rate.
The enemy has far fewer HP than your dwarf and they do damage at a much lower rate.
In case 1, it is unlikely (5%) that you will see the benefit of the spell, and your dwarf is very much at risk. In case 4, it is also unlikely that you will see the benefit of the spell, largely because your allied force didn't need it.
In cases 2 and 3, the benefit of the spell can swing the tide. There isn't a clear cut use case, since it will depend on the specific engagement. In these cases, if the enemies average damage is more than double your force's average damage, then the spell is a bad idea (so is combat).
In this case, we now switch back to the first formula.
(number of rounds for attrition) = (Enemy HP) / (Probability of Hit * Average Damage)
using a Ph of 50%, the dwarf will survive (most of the time) when:
(Enemy HP) / (0.5 * Friendly Avg Damage) < (Dwarf HP) / (0.45 * 0.5 * Enemy Avg Damage)
(Note that the 0.45 is for the AC change, and the 0.5 is the resistance).
This tells us that the probability is that the dwarf will survive ~50% of the time when you have ~1/2 as many hit points as the opposing force. If you have more HP than that, you will survive more. If you roll better (or the enemy rolls worse) you will survive more.
From this, you could consider spreading yourself "that" thin. Pick the hardest hitting allies who are the weakest, and start bonding them until you get to a number (understanding that the enemies rate of damage to the dwarf will go up linearly with each ally you bond) that is close to the ratio of your HP to the entire enemy force's HP.
The dwarf is not fighting. If the dwarf is ALSO taking direct damage, it complicates the equations a bit (makes them second order), and I would need a lot more detail to solve them (including the specific parameters of the engagement on both sides).
There are no area based spells being cast against the multiple bonded allies
No consideration was taken for the enhancement in saving throws for bonded allies (requires much more specific model).