I don't think the references are strong enough to justify the rK.

@@ -12 +12 @@
-<ke_inf>&rK;</ke_inf>
+<ke_inf>&sK;</ke_inf>

https://www.kanjipedia.jp/kanji/0007329200
Kanjipedia describes 沪 as 異体字 and 簡易慣用字体

Editorial policy says it's okay to hide these forms.
https://www.edrdg.org/wiki/index.php/Editorial_policy#Search-only_Forms
>> uncommon itaiji and kyūjitai;

I'm not too fussed either way. I do wonder how accurate the n-grams are. There don't seem to be many usages on twitter, so it's probably not inaccurate to tag it as rare.

jawiki: https://ja.wikipedia.org/wiki/ろ過

K/M n-grams:

濾過 17,518 51.5%
ろ過 16,506 48.5%
沪過      0  0.0% (no initial match)

Since these refs have it as a headword and it's more than just an okurigana variant, should make it visible. Jim added the same kanji over at 「ろ紙」.

@@ -12 +12 @@
-<ke_inf>&sK;</ke_inf>
+<ke_inf>&rK;</ke_inf>

Sankoku and smk display both 濾 and 沪 forms in their headword lines. Meikyo also mentions the 沪 form in its entry for this word.

Google N-gram Corpus Counts
╭─ーー─┬─────────┬───────╮
│ ろ過 │ 520,585 │ 44.4% │
│ 濾過 │ 465,783 │ 39.7% │
│ 沪過 │       0 │  0.0% │ - add, sK
│ ろか │ 185,548 │ 15.8% │
╰─ーー─┴─────────┴───────╯

@@ -8,0 +9,4 @@
+</k_ele>
+<k_ele>
+<keb>沪過</keb>
+<ke_inf>&sK;</ke_inf>

Meikyo vt and vi additions to n,vs entries
-*- via bulkupd.py -*-

@@ -15,0 +16 @@
+<pos>&vt;</pos>