The　traditional　harmonic　source　identification　method　cannot　solve　the　problem　of　non-intrusive　identification　of　specific　types　of　harmonic　sources　for　industrial　users.　To　solve　the　above　problem,　this　paper　proposes　a　method　for　identifying　industrial　harmonic　source　load　subtypes　based　on　blind　source　separation.　This　method　facilitates　non-intrusive　identification　of　the　specific　type　of　harmonic　source　load,　solely　relying　on　the　voltage　and　current　data　collected　at　the　industrial　customer＇s　incoming　line.　First,　starting　from　the　load　equivalent　impedance　model,　a　parallel　circuit　model　with　multiple　load　equivalent　impedances　for　industrial　users　is　established.　Next,　ensemble　empirical　mode　decomposition　and　singular　value　decomposition　are　combined　to　determine　the　number　of　source　impedance　signals　that　constitute　the　integrated　equivalent　impedance　signal　of　the　monitoring　points　at　the　user＇s　incoming　line.　Then,　the　fast　independent　component　analysis　is　used　to　separate　the　harmonic　source　load　equivalent　impedance　signal　from　the　integrated　equivalent　impedance　signal.　Finally,　the　frequency　characteristics　of　the　separated　harmonic　source　load　equivalent　impedance　signal　are　matched　with　the　typical　harmonic　source　load,　and　then　the　subtype　identification　is　realized.　Both　simulation　and　actual　cases　show　that　the　proposed　method　can　accurately　identify　the　specific　types　of　multiple　harmonic　source　loads　contained　in　industrial　users,　which　obviously　demonstrates　its　feasibility　and　practical　applicability.　©2024　Chin.Soc.for　Elec.Eng.