Data modeling

Hello,

I reproduced your use case 

1) I confirm, you need to clean the table and have only 3 columns. I added the M code to achieve this here below. 

2) It it not necessary to have 2 dimension tables.

let
    Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("i45WCgjydwl1DlFwVNJRcg0M9QwwBDL8HH1dQTQCxeoglDrBhcEajKAajLAodUYxBKrBGKrBGEWpC/EOcCXggFgA", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [Item = _t, Equipment_Type1 = _t, Equipment_Name1 = _t, Equipment_Type2 = _t, Equipment_Name2 = _t, Equipment_Type3 = _t, Equipment_Name3 = _t]),
    #"Unpivoted Other Columns" = Table.UnpivotOtherColumns(Source, {"Item"}, "Attribute", "Value"),
    #"Filtered Rows" = Table.SelectRows(#"Unpivoted Other Columns", each ([Value] <> "")),
    #"Split Column by Character Transition" = Table.SplitColumn(#"Filtered Rows", "Attribute", Splitter.SplitTextByCharacterTransition((c) => not List.Contains({"0".."9"}, c), {"0".."9"}), {"Type", "Index"}),

    #"Removed Numeric" = Table.TransformColumns(#"Split Column by Character Transition", {"Type", each Text.Select(_, {"A".."Z","a".."z"}), type text}),
    #"Pivoted Column" = Table.Pivot(#"Removed Numeric", List.Distinct(#"Removed Numeric"[Type]), "Type", "Value"),

    #"Renamed Columns" = Table.RenameColumns(#"Pivoted Column", {{"EquipmentType", "Equipment Type"}, {"EquipmentName", "Equipment Name"}}),
    #"Removed Columns" = Table.RemoveColumns(#"Renamed Columns",{"Index"})
in
    #"Removed Columns"

If it is answering your query, please mark my reply as the solution. Thanks!