error when using SVector and SMatrix

[johhell]

in my example I tried to simulate an induction machines in a 3~ system with position dependent coupling/inductance's.

part of the model for testing

    netz = Spannung | Map(U=:(U1v*sqrt(2.0/3.0))),
    equations = :[
        w = der(phi)
        wt = Omegarated*time
        Lsr  = Interpol1(phi)
        dLsr = Interpol2(phi)
        Lrs  = transpose(Lsr)
        dLrs = transpose(dLsr)
==>     netz.u =r1*i1 + Lss*der(i1)
        u2 = w*dLrs*i1 + Lrs*der(i1)
#         Moment = dot(i1, dLsr*i2)
        Moment = 0.0
        Imech*der(w) = Moment - LAST

        power = dot(netz.u, i1)

code "saveCodeOnFile"

...
        Lsr = Interpol1(phi)
        dLsr = Interpol2(phi)
        Lrs = transpose(Lsr)
        dLrs = transpose(dLsr)
==>     var"der(i1)" = -((var"netz.u" - _p[:r1]::SMatrix{3, 3, Float64, 9} * i1)) / -(_p[:Lss]::SMatrix{3, 3, Float64, 9})
        u2 = (w * dLrs) * i1 + Lrs * var"der(i1)"
        var"der(w)" = -((Float64)(_p[:LAST])::Float64) / (Float64)(_p[:Imech])::Float64
 ...

simplified: var"der(i1)" = SVector / SMatrix the correct function should be: var"der(i1)" = SMatrix^-1 * SVector

errormessage

ERROR: LoadError: DimensionMismatch("Both inputs should have the same number of columns")
Stacktrace:
  [1] /(A::SVector{3, Float64}, B::SMatrix{3, 3, Float64, 9})
    @ LinearAlgebra ~/julia-1.7.1/share/julia/stdlib/v1.7/LinearAlgebra/src/generic.jl:1149
  [2] getDerivatives(_x::Vector{Float64}, _m::SimulationModel{Float64, Float64}, _time::Float64)
    @ Main ~/.julia/packages/Modia/2gNMw/src/A-CodeGeneration.jl:1685
  [3] #invokelatest#2
    @ ./essentials.jl:716 [inlined]

workaround

in my particular case an additional method was helping.

Base.:/(A::SVector, B::SMatrix) =  B^-1 * A

[johhell]

model for testing

testSMatrix.zip