Harmonic　state　estimation　is　a　key　part　of　power　system　operation　management,　which　can　be　used　to　monitor　the　harmonic　in　the　power　grid　and　provide　an　important　reference　for　the　stable　operation　of　the　power　grid.　At　present,　the　number　of　phasor　measurement　unit　(PMU)　configurations　are　difficult　to　satisfy　the　observability　requirements　of　state　estimation.　It　is　necessary　to　adapt　power　quality　monitoring　device　(PQMD)　data　to　improve　the　redundancy　of　measurement　and　make　the　harmonic　state　estimation　possible.　However,　the　non-synchronized　monitoring　data　characteristics　of　PQMD　make　the　fused　measurement　data　still　have　deviations,　which　will　lead　to　a　large　error　in　harmonic　estimation　state.　Fusing　PMU　and　PQMD　measurement　data　and　minimizing　the　asynchronous　measurement　bias　of　PQMD　measurement　data,　as　well　as　suppressing　the　influence　of　this　measurement　bias　in　state　estimation,　will　provide　a　more　effective　means　for　grid　harmonic　analysis.　Therefore,　the　paper　proposes　an　interval　robust　harmonic　state　estimation　method　based　on　PMU　and　PQMD　measurement　data　fusion.　Firstly,　the　detection　period　of　PQMD　is　selected　with　the　overlap　index,　and　the　reference　period　is　selected　with　the　maximum　overlap　as　the　target.　The　selected　reference　period　is　used　as　the　measurement　buffer　of　PMU,　and　the　PQMD　measurement　data　in　this　period　is　fused　to　form　an　interval　mixed　measurement　set.　The　harmonic　power　measurements　of　PQMD　are　converted　into　equivalent　harmonic　current　phasor　measurement　by　the　measurement　transformation　strategy,　which　is　updated　with　state　quantity　in　iterative　solution.　Secondly,　the　projection　statistical　method　is　used　to　calculate　the　initial　weight　of　the　measurement,　and　the　overlap　index　is　introduced　into　the　Huber　weight　function　to　adjust　the　measurement　weight.　The　measurement　with　low　overlap　and　large　residual　is　given　a　small　weight　to　suppress　the　influence　of　measurement　deviation,　and　further　improve　the　robustness　of　the　algorithm.　Finally,　the　measurement　points　are　preferring　according　to　the　weights,　and　the　measurement　subset　with　the　least　deviation　of　non-synchronous　measurement　is　obtained.　The　harmonic　state　estimation　model　is　solved　by　iterative　reweighted　least　square　method,　and　the　harmonic　state　range　of　the　whole　network　is　obtained.　The　simulation　results　show　that　when　the　load　fluctuation　is　10%　and　the　average　overlap　degree　is　0.85,　the　estimated　error　of　the　proposed　method　is　1.92%　in　the　upper　bound　and　3.24%　in　the　lower　bound.　The　error　of　phase　angle　upper　bound　estimation　is　2.27%,　and　the　error　of　lower　bound　estimation　is　4.22%.　When　the　overlap　degree　is　reduced　to　0.6,　the　average　error　of　amplitude　and　phase　angle　of　the　proposed　algorithm　is　less　than　6%.　When　the　level　of　load　fluctuation　increases　to　40%,　the　average　estimation　error　of　state　quantity　is　less　than　5%.　The　following　conclusions　can　be　obtained　through　simulation　analysis:　(1)　The　overlap　index　is　used　to　quantify　the　measurement　deviation　of　PQMD　to　improve　the　weight　matrix　in　the　state　estimation,　which　can　effectively　suppress　the　influence　of　the　measurement　deviation　on　state　estimation.　In　addition,　the　interval　mixed　measurement　subset　is　obtained　by　preferring　the　measurement　points　according　to　the　weight　coefficient,　which　can　further　reduce　the　non-synchronous　measurement　deviation　of　the　measurement　set,　improve　the　reliability　of　the　interval　mixed　measurement　set　and　the　accuracy　of　state　estimation.　(2)　Converting　the　interval　weight　matrix　and　Jacobian　matrix　into　a　definite　value　can　further　reduce　the　conservatism　of　the　solution　interval.　(3)　The　proposed　algorithm　can　effectively　reduce　the　estimation　error　under　different　measurement　deviations　and　different　load　fluctuation　sizes　which　has　robustness.　©　2024　China　Machine　Press.　All　rights　reserved.