In　order　to　adapt　the　network　structure　of　a　distribution　network　to　the　access　of　new　energy　technologies　such　as　distributed　generation　(DG),　energy　storage　and　DC　loads,　a　new　network　planning　method　for　AC/DC　hybrid　distribution　networks　is　proposed.　The　method　considered　all　possibilities　of　the　AC/DC　configuration　to　improve　the　economic　benefits　of　the　network　construction,　and　constructed　a　linearized　single-layer　network　planning　model　to　improve　the　global　optimization　capability　in　view　of　the　difficulty　of　obtaining　global　optimal　solutions　for　the　network　planning　model.　Firstly,　considering　that　AC/DC　types　of　buses　and　the　connection　states　between　buses　determine　the　investment　cost　of　converters　and　lines,　which　determines　the　economic　benefits　of　distribution　network　planning,　a　binary　bus　type　matrix　is　used　to　represent　AC/DC　types　of　buses,　a　binary　network　association　matrix　is　developed　to　represent　the　connection　states　between　buses,　and　binary　line　type　variables　are　adopoted　to　represent　line　type　selection　to　construct　network　structure　decision　variables,　achieving　a　description　of　the　overall　network　structure.　Secondly,　with　an　objective　of　converter　installation　cost,　line　construction　cost　and　system　operation　and　maintenance　cost,　a　single-layer　network　planning　model　is　established　considering　elaborated　and　practical　constraints,　such　as　the　number　of　node-connected　lines,　line　selection　constraint,　power　flow,　voltage　source　converter　(VSC)　reactive　power　compensation　capacity,　system　safety　operation,　energy　storage　operation,　network　connectivity　and　so　on.　In　this　proposed　model,　the　AC　and　DC　current　models　is　unified　by　introducing　branch　type　variables,　thus　linking　the　planning　layer　and　operation　layer,　so　that　network　planning　variables　and　operation　optimization　variables　consisting　of　power　purchased　from　the　distribution　network　to　the　grid,　DG　output,　energy　storage　charging　and　discharging　power　and　VSC　power　were　optimized　simultaneously　to　enhance　the　possibility　of　obtaining　a　global　optimal　solution.　Finally,　to　achieve　better　sulution,　power　flow　equations　and　line　transmission　capacity　constraints　are　linearized　by　a　series　of　linearization　methods,　such　that　the　mixed　integer　nonlinear　programming　(MINLP)　problem　was　transformed　into　a　mixed　integer　linear　programming　(MILP)　problem,　and　the　GUROBI　solver　is　invoked　in　Matlab　to　solve　it.　The　validity　of　the　planning　method　was　verified　in　a　13-node　distribution　system.　The　simulation　results　show　that:　(1)　The　AC/DC　hybrid　distribution　network　planning　scheme　can　save　5.05%　of　the　total　investment　and　operation　cost　compared　with　the　pure　AC　distribution　network　planning,　and　the　network　structure　of　the　planned　AC/DC　hybrid　distribution　network　also　has　greater　network　supply　capacity,　which　illustrates　that　the　AC/DC　hybrid　distribution　network　is　more　advantageous　in　the　case　of　mixed　AC/DC　sources　and　loads.　(2)　The　proposed　planning　method　saves　13.54%　of　the　total　investment　and　operation　cost　and　91.32%　of　the　computation　time　compared　with　the　traditional　bi-layer　planning　model　solved　by　genetic　algorithm,　which　indicates　that　the　single-layer　network　planning　method　has　prominent　global　optimization　performance　and　faster　computational　efficiency.　(3)　As　the　method　does　not　require　relaxation　of　constraints,　compared　to　second-order　cone　planning　techniques,　there　are　no　limitations　in　terms　of　grid　topology,　etc,　and　it　is　suitable　for　the　planning　of　both　radial　and　meshed　distribution　networks.　©　2024　China　Machine　Press.　All　rights　reserved.