Economists typically abhor tariffs since they reduce total surplus, harm people/companies in both importing and exporting nations, and generally reduce efficiency. However, despite this rhetoric, it is, in fact, possible for a tariff to be a net-positive policy within a country.
The most common argument in favor of tariffs is to protect strategic industries. Henry Kissinger once said "America has no permanent friends or enemies, only interests". More generally reliance on other countries for food, military production, and other essentials reduces the ability of countries to achieve their foreign policy objectives and be free of foreign influence themselves.
If you think your country's influence is net-positive while other countries' is net-negative (a fairly common belief), you can justify tariffs to keep "essential" industries domestic.
However, I have another argument I haven't heard anyone else give:
Suppose the US imports Rembrandt paintings from the Netherlands. The supply of Rembrandt paintings is perfectly inelastic, which means that a tariff has no deadweight loss. Moreover, the supply being perfectly inelastic implies that 100% of the burden of the tax falls on the suppliers - that is, the Dutch will pay the entire tax to the US. This is unambiguously, a net-positive for the US.
Moreover, on the margin, dead-weight loss always grows quadratically while revenue-from-foreigners grows linearly. This implies (with some calculus) that small tariffs are nearly equivalent to transfers from the exporting countries to importing country and that, from a self-interested perspective, the optimal tariff on imports (and exports) is always non-zero unless the demand (supply) is perfectly inelastic.
And what's more, this line of thinking suggests there are times when even perfect altruists should prefer tariffs: namely when poor nations impose them on richer ones. The justification goes like this: (1) small tariffs are effectively transfers (2) transfers from rich countries to poor countries are good (3) therefore small tariffs imposed by poor countries on rich countries are good.
Now, I should mention that in practice this line of argument probably doesn't work very well. In the real world, when one country levies tariffs, this typically causes other countries to levy their own tariffs in retaliation. Moreover, the fact that small tariffs can be optimal in theory doesn't imply they are in reality since tariffs in the real world impose other costs like bookkeeping costs, enforcement costs, and foreign relations costs.
Still, I'm surprised I haven't heard even one person mention this line of thought. Among economists, by far the most common reason to reject tariffs is that consumers would lose more than the producers gain - I've never heard anyone say that tariffs can often be net-positive for the government implementing them and I, instead, frequently hear that the costs to domestic consumers outweigh the benefits to domestic producers (true, only if one completely ignores the benefits of tax revenue).