Modelica MSL中的动态管道模型,有限体积法
我试图使用Modelica对由弹性管道组成的系统进行建模。 现在,我正在尝试使用与Modelica.Fluid库中相同的方法(有限体积,交错)实现我自己的动态管道模型(刚性,但不是弹性),但当然不包括所有选项 这个模型应该更容易理解,因为它是一个平面模型,而不是从其他类扩展而来。这一点很重要,因为即使没有Modelica专有技术,我的同事也可以理解该模型,我可以说服他们Modelica是满足我们目的的适当工具 作为一个测试案例,我使用带有阶跃信号(水锤)的质量流源。 我的模型给出的结果与Modelica.Fluid组件不同。 我真的很感激,如果有人能帮助我,了解发生了什么事 测试系统如下所示:Modelica MSL中的动态管道模型,有限体积法,modelica,dymola,openmodelica,Modelica,Dymola,Openmodelica,我试图使用Modelica对由弹性管道组成的系统进行建模。 现在,我正在尝试使用与Modelica.Fluid库中相同的方法(有限体积,交错)实现我自己的动态管道模型(刚性,但不是弹性),但当然不包括所有选项 这个模型应该更容易理解,因为它是一个平面模型,而不是从其他类扩展而来。这一点很重要,因为即使没有Modelica专有技术,我的同事也可以理解该模型,我可以说服他们Modelica是满足我们目的的适当工具 作为一个测试案例,我使用带有阶跃信号(水锤)的质量流源。 我的模型给出的结果与Mode
model Pipe_FVM_staggered
// Import
import SI = Modelica.SIunits;
import Modelica.Constants.pi;
// Medium
replaceable package Medium = Modelica.Media.Interfaces.PartialMedium "Medium in the component"
annotation (choicesAllMatching = true);
// Interfaces, Ports
Modelica.Fluid.Interfaces.FluidPort_a port_a(redeclare package Medium = Medium) annotation (Placement(transformation(extent={{-110,-10},{-90,10}})));
Modelica.Fluid.Interfaces.FluidPort_b port_b(redeclare package Medium = Medium) annotation (Placement(transformation(extent={{90,-10},{110,10}})));
// Parameters
parameter Integer n(min=2) = 3 "Number of cells"; // No effect yet, only for icon
parameter SI.Length L = 1 "Length";
parameter SI.Diameter D = 0.010 "Diameter";
parameter SI.Height R = 2.5e-5 "Roughness";
parameter Boolean use_fixed_zeta = false "Use fixed zeta value instead of Moody chart";
parameter SI.CoefficientOfFriction zeta = 1;
// Initialization
parameter Medium.Temperature T_start = 293.15 "Start temperature" annotation(Dialog(tab="Initialization"));
parameter Medium.MassFlowRate mflow_start = 1 "Start mass flow rate in design direction" annotation(Dialog(tab="Initialization"));
parameter Medium.AbsolutePressure p_a_start = 2e5 "Start pressure p[1] at design inflow" annotation(Dialog(tab="Initialization"));
parameter Medium.AbsolutePressure p_b_start = 1e5 "Start pressure for p[n+1] at design outflow" annotation(Dialog(tab="Initialization"));
// parameter Medium.AbsolutePressure p_start = (p_a_start + p_b_start)/2 annotation(Dialog(tab="Initialization"));
parameter Medium.AbsolutePressure p_start[:] = linspace(p_a_start, p_b_start, n) annotation(Dialog(tab="Initialization"));
// parameter Medium.SpecificEnthalpy h_start[:] = Medium.specificEnthalpy_pTX(p_start, T_start, Medium.X_default);
parameter Medium.SpecificEnthalpy h_start = Medium.specificEnthalpy_pTX((p_a_start + p_b_start)/2, T_start, Medium.X_default) annotation(Dialog(tab="Initialization"));
parameter SI.AbsolutePressure dp_nominal = 1e5;
parameter SI.MassFlowRate m_flow_nominal = 1;
// Variables general
SI.Length dL = L/n;
SI.Area A(nominal=0.001) = D^2*pi/4;
SI.Volume V = A * dL;
// Variables cell centers: positiv in direction a -> b
Medium.AbsolutePressure p[n](start = p_start, each stateSelect=StateSelect.prefer) annotation(Dialog(tab="Initialization", showStartAttribute=true, enable=false));
Medium.SpecificEnthalpy h[n](each start = h_start, each stateSelect=StateSelect.prefer) annotation(Dialog(tab="Initialization", showStartAttribute=true, enable=false));
Medium.ThermodynamicState state[n] = Medium.setState_phX(p,h);
SI.Mass m[n] = rho .* V;
Medium.Density rho[n] = Medium.density(state);
SI.InternalEnergy U[n] = m .* u;
Medium.SpecificInternalEnergy u[n] = Medium.specificInternalEnergy(state);
Medium.Temperature T[n] = Medium.temperature(state);
Medium.DynamicViscosity mu[n] = Medium.dynamicViscosity(state);
SI.Velocity v[n](nominal=0.2) = 0.5 * (mflow[1:n] + mflow[2:n+1]) ./ rho ./ A;
SI.Power Wflow[n];
SI.MomentumFlux Iflow[n] = v .* v .* rho * A;
// Variables faces: positiv in direction a -> b
Medium.MassFlowRate mflow[n+1](each start = mflow_start, each stateSelect=StateSelect.prefer, nominal=0.25) annotation(Dialog(tab="Initialization", showStartAttribute=true, enable=false));
Medium.EnthalpyFlowRate Hflow[n+1];
SI.Momentum I[n-1] = mflow[2:n] * dL;
SI.Force Fp[n-1];
SI.Force Ff[n-1];
SI.PressureDifference dpf[n-1](each start = (p_a_start - p_b_start)/(n-1), nominal=0.01e5) annotation(Dialog(tab="Initialization", showStartAttribute=true, enable=false));
equation
der(m) = mflow[1:n] - mflow[2:n+1]; // Mass balance
der(U) = Hflow[1:n] - Hflow[2:n+1] + Wflow; // Energy balance
der(I) = Iflow[1:n-1] - Iflow[2:n] + Fp - Ff; // Momentum balance, staggered
Hflow[1] = semiLinear(mflow[1], inStream(port_a.h_outflow), h[1]);
Hflow[2:n] = semiLinear(mflow[2:n], h[1:n-1], h[2:n]);
Hflow[n+1] = semiLinear(mflow[n+1], h[n], inStream(port_b.h_outflow));
Wflow[1] = v[1] * A .* ( (p[2] - p[1])/2 + dpf[1]/2);
Wflow[2:n-1] = v[2:n-1] * A .* ( (p[3:n]-p[1:n-2])/2 + (dpf[1:n-2]+dpf[2:n-1])/2);
Wflow[n] = v[n] * A .* ( (p[n] - p[n-1])/2 + dpf[n-1]/2);
Fp = A * (p[1:n-1] - p[2:n]);
Ff = A * dpf; // dpf = Ff ./ A;
if use_fixed_zeta then
dpf = 1/2 * zeta/(n-1) * (mflow[2:n]).^2 ./ ( 0.5*(rho[1:n-1] + rho[2:n]) * A * A);
else
dpf = homotopy(
actual = Modelica.Fluid.Pipes.BaseClasses.WallFriction.Detailed.pressureLoss_m_flow(
m_flow = mflow[2:n],
rho_a = rho[1:n-1],
rho_b = rho[2:n],
mu_a = mu[1:n-1],
mu_b = mu[2:n],
length = dL,
diameter = D,
roughness = R,
m_flow_small = 0.001),
simplified = dp_nominal/(n-1)/m_flow_nominal*mflow[2:n]);
end if;
// Boundary conditions
mflow[1] = port_a.m_flow;
mflow[n] = -port_b.m_flow;
p[1] = port_a.p;
p[n] = port_b.p;
port_a.h_outflow = h[1];
port_b.h_outflow = h[n];
initial equation
der(mflow[2:n]) = zeros(n-1);
der(p) = zeros(n);
der(h) = zeros(n);
annotation (Icon(coordinateSystem(preserveAspectRatio=false), graphics={Rectangle(
extent={{-100,60},{100,-60}},
fillColor={255,255,255},
fillPattern=FillPattern.HorizontalCylinder,
lineColor={0,0,0}),
Line(
points={{-100,60},{-100,-60}},
color={0,0,0},
thickness=0.5),
Line(
points={{-60,60},{-60,-60}},
color={0,0,0},
thickness=0.5),
Line(
points={{-20,60},{-20,-60}},
color={0,0,0},
thickness=0.5),
Line(
points={{20,60},{20,-60}},
color={0,0,0},
thickness=0.5),
Line(
points={{60,60},{60,-60}},
color={0,0,0},
thickness=0.5),
Line(
points={{100,60},{100,-60}},
color={0,0,0},
thickness=0.5),
Line(
points={{60,-80},{-60,-80}},
color={0,128,255},
visible=showDesignFlowDirection),
Polygon(
points={{20,-65},{60,-80},{20,-95},{20,-65}},
lineColor={0,128,255},
fillColor={0,128,255},
fillPattern=FillPattern.Solid,
visible=showDesignFlowDirection),
Text(
extent={{-150,100},{150,60}},
lineColor={0,0,255},
textString="%name"),
Text(
extent={{-40,22},{40,-18}},
lineColor={0,0,0},
textString="n = %n")}), Diagram(
coordinateSystem(preserveAspectRatio=false)));
end Pipe_FVM_staggered;
11个单元格的结果如下:
如您所见,MSL组件的压力峰值较高,频率/周期也不相同。当我选择更多的单元格时,错误就会变小
我很确定我使用的是完全相同的方程。
这可能是数字原因(我尝试使用标称值)的原因吗?
我还为Modelica.Fluid组件提供了我自己的“固定zeta”流动模型,以便在固定压力损失系数zeta的情况下进行比较
我的管道模型的代码非常短,如果我能让它像这样工作,那就太好了:
model Pipe_FVM_staggered
// Import
import SI = Modelica.SIunits;
import Modelica.Constants.pi;
// Medium
replaceable package Medium = Modelica.Media.Interfaces.PartialMedium "Medium in the component"
annotation (choicesAllMatching = true);
// Interfaces, Ports
Modelica.Fluid.Interfaces.FluidPort_a port_a(redeclare package Medium = Medium) annotation (Placement(transformation(extent={{-110,-10},{-90,10}})));
Modelica.Fluid.Interfaces.FluidPort_b port_b(redeclare package Medium = Medium) annotation (Placement(transformation(extent={{90,-10},{110,10}})));
// Parameters
parameter Integer n(min=2) = 3 "Number of cells"; // No effect yet, only for icon
parameter SI.Length L = 1 "Length";
parameter SI.Diameter D = 0.010 "Diameter";
parameter SI.Height R = 2.5e-5 "Roughness";
parameter Boolean use_fixed_zeta = false "Use fixed zeta value instead of Moody chart";
parameter SI.CoefficientOfFriction zeta = 1;
// Initialization
parameter Medium.Temperature T_start = 293.15 "Start temperature" annotation(Dialog(tab="Initialization"));
parameter Medium.MassFlowRate mflow_start = 1 "Start mass flow rate in design direction" annotation(Dialog(tab="Initialization"));
parameter Medium.AbsolutePressure p_a_start = 2e5 "Start pressure p[1] at design inflow" annotation(Dialog(tab="Initialization"));
parameter Medium.AbsolutePressure p_b_start = 1e5 "Start pressure for p[n+1] at design outflow" annotation(Dialog(tab="Initialization"));
// parameter Medium.AbsolutePressure p_start = (p_a_start + p_b_start)/2 annotation(Dialog(tab="Initialization"));
parameter Medium.AbsolutePressure p_start[:] = linspace(p_a_start, p_b_start, n) annotation(Dialog(tab="Initialization"));
// parameter Medium.SpecificEnthalpy h_start[:] = Medium.specificEnthalpy_pTX(p_start, T_start, Medium.X_default);
parameter Medium.SpecificEnthalpy h_start = Medium.specificEnthalpy_pTX((p_a_start + p_b_start)/2, T_start, Medium.X_default) annotation(Dialog(tab="Initialization"));
parameter SI.AbsolutePressure dp_nominal = 1e5;
parameter SI.MassFlowRate m_flow_nominal = 1;
// Variables general
SI.Length dL = L/n;
SI.Area A(nominal=0.001) = D^2*pi/4;
SI.Volume V = A * dL;
// Variables cell centers: positiv in direction a -> b
Medium.AbsolutePressure p[n](start = p_start, each stateSelect=StateSelect.prefer) annotation(Dialog(tab="Initialization", showStartAttribute=true, enable=false));
Medium.SpecificEnthalpy h[n](each start = h_start, each stateSelect=StateSelect.prefer) annotation(Dialog(tab="Initialization", showStartAttribute=true, enable=false));
Medium.ThermodynamicState state[n] = Medium.setState_phX(p,h);
SI.Mass m[n] = rho .* V;
Medium.Density rho[n] = Medium.density(state);
SI.InternalEnergy U[n] = m .* u;
Medium.SpecificInternalEnergy u[n] = Medium.specificInternalEnergy(state);
Medium.Temperature T[n] = Medium.temperature(state);
Medium.DynamicViscosity mu[n] = Medium.dynamicViscosity(state);
SI.Velocity v[n](nominal=0.2) = 0.5 * (mflow[1:n] + mflow[2:n+1]) ./ rho ./ A;
SI.Power Wflow[n];
SI.MomentumFlux Iflow[n] = v .* v .* rho * A;
// Variables faces: positiv in direction a -> b
Medium.MassFlowRate mflow[n+1](each start = mflow_start, each stateSelect=StateSelect.prefer, nominal=0.25) annotation(Dialog(tab="Initialization", showStartAttribute=true, enable=false));
Medium.EnthalpyFlowRate Hflow[n+1];
SI.Momentum I[n-1] = mflow[2:n] * dL;
SI.Force Fp[n-1];
SI.Force Ff[n-1];
SI.PressureDifference dpf[n-1](each start = (p_a_start - p_b_start)/(n-1), nominal=0.01e5) annotation(Dialog(tab="Initialization", showStartAttribute=true, enable=false));
equation
der(m) = mflow[1:n] - mflow[2:n+1]; // Mass balance
der(U) = Hflow[1:n] - Hflow[2:n+1] + Wflow; // Energy balance
der(I) = Iflow[1:n-1] - Iflow[2:n] + Fp - Ff; // Momentum balance, staggered
Hflow[1] = semiLinear(mflow[1], inStream(port_a.h_outflow), h[1]);
Hflow[2:n] = semiLinear(mflow[2:n], h[1:n-1], h[2:n]);
Hflow[n+1] = semiLinear(mflow[n+1], h[n], inStream(port_b.h_outflow));
Wflow[1] = v[1] * A .* ( (p[2] - p[1])/2 + dpf[1]/2);
Wflow[2:n-1] = v[2:n-1] * A .* ( (p[3:n]-p[1:n-2])/2 + (dpf[1:n-2]+dpf[2:n-1])/2);
Wflow[n] = v[n] * A .* ( (p[n] - p[n-1])/2 + dpf[n-1]/2);
Fp = A * (p[1:n-1] - p[2:n]);
Ff = A * dpf; // dpf = Ff ./ A;
if use_fixed_zeta then
dpf = 1/2 * zeta/(n-1) * (mflow[2:n]).^2 ./ ( 0.5*(rho[1:n-1] + rho[2:n]) * A * A);
else
dpf = homotopy(
actual = Modelica.Fluid.Pipes.BaseClasses.WallFriction.Detailed.pressureLoss_m_flow(
m_flow = mflow[2:n],
rho_a = rho[1:n-1],
rho_b = rho[2:n],
mu_a = mu[1:n-1],
mu_b = mu[2:n],
length = dL,
diameter = D,
roughness = R,
m_flow_small = 0.001),
simplified = dp_nominal/(n-1)/m_flow_nominal*mflow[2:n]);
end if;
// Boundary conditions
mflow[1] = port_a.m_flow;
mflow[n] = -port_b.m_flow;
p[1] = port_a.p;
p[n] = port_b.p;
port_a.h_outflow = h[1];
port_b.h_outflow = h[n];
initial equation
der(mflow[2:n]) = zeros(n-1);
der(p) = zeros(n);
der(h) = zeros(n);
annotation (Icon(coordinateSystem(preserveAspectRatio=false), graphics={Rectangle(
extent={{-100,60},{100,-60}},
fillColor={255,255,255},
fillPattern=FillPattern.HorizontalCylinder,
lineColor={0,0,0}),
Line(
points={{-100,60},{-100,-60}},
color={0,0,0},
thickness=0.5),
Line(
points={{-60,60},{-60,-60}},
color={0,0,0},
thickness=0.5),
Line(
points={{-20,60},{-20,-60}},
color={0,0,0},
thickness=0.5),
Line(
points={{20,60},{20,-60}},
color={0,0,0},
thickness=0.5),
Line(
points={{60,60},{60,-60}},
color={0,0,0},
thickness=0.5),
Line(
points={{100,60},{100,-60}},
color={0,0,0},
thickness=0.5),
Line(
points={{60,-80},{-60,-80}},
color={0,128,255},
visible=showDesignFlowDirection),
Polygon(
points={{20,-65},{60,-80},{20,-95},{20,-65}},
lineColor={0,128,255},
fillColor={0,128,255},
fillPattern=FillPattern.Solid,
visible=showDesignFlowDirection),
Text(
extent={{-150,100},{150,60}},
lineColor={0,0,255},
textString="%name"),
Text(
extent={{-40,22},{40,-18}},
lineColor={0,0,0},
textString="n = %n")}), Diagram(
coordinateSystem(preserveAspectRatio=false)));
end Pipe_FVM_staggered;
很长一段时间以来,我一直在努力解决这个问题,因此非常感谢您的任何评论或提示!!
如果您需要更多信息或测试结果,请告诉我
这是测试示例的代码:
model Test_Waterhammer
extends Modelica.Icons.Example;
import SI = Modelica.SIunits;
import g = Modelica.Constants.g_n;
replaceable package Medium = Modelica.Media.Water.StandardWater;
Modelica.Fluid.Sources.Boundary_pT outlet(
redeclare package Medium = Medium,
nPorts=1,
p=2000000,
T=293.15)
annotation (Placement(transformation(extent={{90,-10},{70,10}})));
inner Modelica.Fluid.System system(
allowFlowReversal=true,
energyDynamics=Modelica.Fluid.Types.Dynamics.SteadyStateInitial,
massDynamics=Modelica.Fluid.Types.Dynamics.SteadyStateInitial,
momentumDynamics=Modelica.Fluid.Types.Dynamics.SteadyStateInitial,
m_flow_start=0.1,
m_flow_small=0.0001)
annotation (Placement(transformation(extent={{60,60},{80,80}})));
Modelica.Fluid.Sources.MassFlowSource_T inlet(
redeclare package Medium = Medium,
nPorts=1,
m_flow=0.1,
use_m_flow_in=true,
T=293.15)
annotation (Placement(transformation(extent={{-50,-10},{-30,10}})));
Modelica.Blocks.Sources.TimeTable timeTable(table=[0,0.1; 1,0.1; 1,0.25;
40,0.25; 40,0.35; 60,0.35])
annotation (Placement(transformation(extent={{-90,10},{-70,30}})));
Pipe_FVM_staggered pipe(
redeclare package Medium = Medium,
R=0.035*0.005,
mflow_start=0.1,
L=1000,
m_flow_nominal=0.1,
D=0.035,
zeta=2000,
n=11,
use_fixed_zeta=false,
T_start=293.15,
p_a_start=2010000,
p_b_start=2000000,
dp_nominal=10000)
annotation (Placement(transformation(extent={{10,-10},{30,10}})));
Modelica.Fluid.Pipes.DynamicPipe pipeMSL(
redeclare package Medium = Medium,
allowFlowReversal=true,
length=1000,
roughness=0.035*0.005,
m_flow_start=0.1,
energyDynamics=Modelica.Fluid.Types.Dynamics.SteadyStateInitial,
massDynamics=Modelica.Fluid.Types.Dynamics.SteadyStateInitial,
momentumDynamics=Modelica.Fluid.Types.Dynamics.SteadyStateInitial,
diameter=0.035,
modelStructure=Modelica.Fluid.Types.ModelStructure.av_vb,
redeclare model FlowModel =
Modelica.Fluid.Pipes.BaseClasses.FlowModels.DetailedPipeFlow (
useUpstreamScheme=false, use_Ib_flows=true),
p_a_start=2010000,
p_b_start=2000000,
T_start=293.15,
nNodes=11)
annotation (Placement(transformation(extent={{10,-50},{30,-30}})));
Modelica.Fluid.Sources.MassFlowSource_T inlet1(
redeclare package Medium = Medium,
nPorts=1,
m_flow=0.1,
use_m_flow_in=true,
T=293.15)
annotation (Placement(transformation(extent={{-48,-50},{-28,-30}})));
Modelica.Fluid.Sources.Boundary_pT outlet1(
redeclare package Medium = Medium,
nPorts=1,
p=2000000,
T=293.15)
annotation (Placement(transformation(extent={{90,-50},{70,-30}})));
equation
connect(inlet.ports[1], pipe.port_a)
annotation (Line(points={{-30,0},{-10,0},{10,0}}, color={0,127,255}));
connect(pipe.port_b, outlet.ports[1])
annotation (Line(points={{30,0},{50,0},{70,0}}, color={0,127,255}));
connect(inlet1.ports[1], pipeMSL.port_a)
annotation (Line(points={{-28,-40},{-10,-40},{10,-40}}, color={0,127,255}));
connect(pipeMSL.port_b, outlet1.ports[1])
annotation (Line(points={{30,-40},{50,-40},{70,-40}}, color={0,127,255}));
connect(timeTable.y, inlet.m_flow_in)
annotation (Line(points={{-69,20},{-60,20},{-60,8},{-50,8}}, color={0,0,127}));
connect(inlet1.m_flow_in, inlet.m_flow_in)
annotation (Line(points={{-48,-32},{-60,-32},{-60,8},{-50,8}}, color={0,0,127}));
annotation (Icon(coordinateSystem(preserveAspectRatio=false)), Diagram(
coordinateSystem(preserveAspectRatio=false)),
experiment(
StopTime=15,
__Dymola_NumberOfIntervals=6000,
Tolerance=1e-005,
__Dymola_Algorithm="Dassl"));
end Test_Waterhammer;
我已经用301个单元格运行了测试:
缩放峰值1和2:
解决方案:根据scottG的建议进行修改
model FVM_staggered_Ncells
// Import
import SI = Modelica.SIunits;
import Modelica.Constants.pi;
// Medium
replaceable package Medium = Modelica.Media.Interfaces.PartialMedium "Medium in the component"
annotation (choicesAllMatching = true);
// Interfaces, Ports
Modelica.Fluid.Interfaces.FluidPort_a port_a(redeclare package Medium = Medium) annotation (Placement(transformation(extent={{-110,-10},{-90,10}})));
Modelica.Fluid.Interfaces.FluidPort_b port_b(redeclare package Medium = Medium) annotation (Placement(transformation(extent={{90,-10},{110,10}})));
// Parameters
parameter Integer n(min=2) = 3 "Number of cells"; // No effect yet, only for icon
parameter SI.Length L = 1 "Length";
parameter SI.Diameter D = 0.010 "Diameter";
parameter SI.Height R = 2.5e-5 "Roughness";
parameter Boolean use_fixed_zeta = false "Use fixed zeta value instead of Moody chart";
parameter SI.CoefficientOfFriction zeta = 1;
// Initialization
parameter Medium.Temperature T_start = 293.15 "Start temperature" annotation(Dialog(tab="Initialization"));
parameter Medium.MassFlowRate mflow_start = 1 "Start mass flow rate in design direction" annotation(Dialog(tab="Initialization"));
parameter Medium.AbsolutePressure p_a_start = 2e5 "Start pressure p[1] at design inflow" annotation(Dialog(tab="Initialization"));
parameter Medium.AbsolutePressure p_b_start = 1e5 "Start pressure for p[n+1] at design outflow" annotation(Dialog(tab="Initialization"));
parameter Medium.AbsolutePressure p_start[:] = linspace(p_a_start, p_b_start, n) annotation(Dialog(tab="Initialization"));
// parameter Medium.SpecificEnthalpy h_start[:] = Medium.specificEnthalpy_pTX(p_start, T_start, Medium.X_default);
parameter Medium.SpecificEnthalpy h_start = Medium.specificEnthalpy_pTX((p_a_start + p_b_start)/2, T_start, Medium.X_default) annotation(Dialog(tab="Initialization"));
parameter SI.AbsolutePressure dp_nominal = 1e5;
parameter SI.MassFlowRate m_flow_nominal = 1;
// Variables general
SI.Length dL = L/n;
SI.Length dLs[n-1] = cat(1,{1.5*dL}, fill(dL,n-3), {1.5*dL});
SI.Area A = D^2*pi/4;
SI.Volume V = A * dL;
// Variables cell centers: positiv in direction a -> b
Medium.AbsolutePressure p[n](start = p_start, each stateSelect=StateSelect.prefer) annotation(Dialog(tab="Initialization", showStartAttribute=true, enable=false));
Medium.SpecificEnthalpy h[n](each start = h_start, each stateSelect=StateSelect.prefer) annotation(Dialog(tab="Initialization", showStartAttribute=true, enable=false));
Medium.ThermodynamicState state[n] = Medium.setState_phX(p,h);
SI.Mass m[n] = rho .* V;
Medium.Density rho[n] = Medium.density(state);
SI.InternalEnergy U[n] = m .* u;
Medium.SpecificInternalEnergy u[n] = Medium.specificInternalEnergy(state);
Medium.Temperature T[n] = Medium.temperature(state);
Medium.DynamicViscosity mu[n] = Medium.dynamicViscosity(state);
SI.Velocity v[n] = 0.5 * (mflow[1:n] + mflow[2:n+1]) ./ rho ./ A;
SI.Power Wflow[n];
SI.MomentumFlux Iflow[n] = v .* v .* rho * A;
// Variables faces: positiv in direction a -> b
Medium.MassFlowRate mflow[n+1](each start = mflow_start, each stateSelect=StateSelect.prefer) annotation(Dialog(tab="Initialization", showStartAttribute=true, enable=false));
Medium.EnthalpyFlowRate Hflow[n+1];
SI.Momentum I[n-1] = mflow[2:n] .* dLs;
SI.Force Fp[n-1];
SI.Force Ff[n-1];
SI.PressureDifference dpf[n-1](each start = (p_a_start - p_b_start)/(n-1)) annotation(Dialog(tab="Initialization", showStartAttribute=true, enable=false));
equation
der(m) = mflow[1:n] - mflow[2:n+1]; // Mass balance
der(U) = Hflow[1:n] - Hflow[2:n+1] + Wflow; // Energy balance
der(I) = Iflow[1:n-1] - Iflow[2:n] + Fp - Ff; // Momentum balance, staggered
Hflow[1] = semiLinear(mflow[1], inStream(port_a.h_outflow), h[1]);
Hflow[2:n] = semiLinear(mflow[2:n], h[1:n-1], h[2:n]);
Hflow[n+1] = semiLinear(mflow[n+1], h[n], inStream(port_b.h_outflow));
Wflow[1] = v[1] * A .* ( (p[2] - p[1])/2 + dpf[1]/2);
Wflow[2:n-1] = v[2:n-1] * A .* ( (p[3:n]-p[1:n-2])/2 + (dpf[1:n-2]+dpf[2:n-1])/2);
Wflow[n] = v[n] * A .* ( (p[n] - p[n-1])/2 + dpf[n-1]/2);
Fp = A * (p[1:n-1] - p[2:n]);
Ff = A * dpf;
if use_fixed_zeta then
dpf = 0.5 * zeta/(n-1) * abs(mflow[2:n]) .* mflow[2:n] ./ ( 0.5*(rho[1:n-1] + rho[2:n]) * A * A);
else
dpf = homotopy(
actual = Modelica.Fluid.Pipes.BaseClasses.WallFriction.Detailed.pressureLoss_m_flow(
m_flow = mflow[2:n],
rho_a = 0.5*(rho[1:n-1] + rho[2:n]),
rho_b = 0.5*(rho[1:n-1] + rho[2:n]),
mu_a = 0.5*(mu[1:n-1] + mu[2:n]),
mu_b = 0.5*(mu[1:n-1] + mu[2:n]),
length = dLs,
diameter = D,
roughness = R,
m_flow_small = 0.001),
simplified = dp_nominal/(n-1)/m_flow_nominal*mflow[2:n]);
end if;
// Boundary conditions
mflow[1] = port_a.m_flow;
mflow[n+1] = -port_b.m_flow;
p[1] = port_a.p;
p[n] = port_b.p;
port_a.h_outflow = h[1];
port_b.h_outflow = h[n];
initial equation
der(mflow[2:n]) = zeros(n-1);
der(p) = zeros(n);
der(h) = zeros(n);
annotation (Icon(coordinateSystem(preserveAspectRatio=false), graphics={Rectangle(
extent={{-100,60},{100,-60}},
fillColor={255,255,255},
fillPattern=FillPattern.HorizontalCylinder,
lineColor={0,0,0}),
Line(
points={{-100,60},{-100,-60}},
color={0,0,0},
thickness=0.5),
Line(
points={{-60,60},{-60,-60}},
color={0,0,0},
thickness=0.5),
Line(
points={{-20,60},{-20,-60}},
color={0,0,0},
thickness=0.5),
Line(
points={{20,60},{20,-60}},
color={0,0,0},
thickness=0.5),
Line(
points={{60,60},{60,-60}},
color={0,0,0},
thickness=0.5),
Line(
points={{100,60},{100,-60}},
color={0,0,0},
thickness=0.5),
Line(
points={{60,-80},{-60,-80}},
color={0,128,255},
visible=showDesignFlowDirection),
Polygon(
points={{20,-65},{60,-80},{20,-95},{20,-65}},
lineColor={0,128,255},
fillColor={0,128,255},
fillPattern=FillPattern.Solid,
visible=showDesignFlowDirection),
Text(
extent={{-150,100},{150,60}},
lineColor={0,0,255},
textString="%name"),
Text(
extent={{-40,22},{40,-18}},
lineColor={0,0,0},
textString="n = %n")}),
Diagram(coordinateSystem(preserveAspectRatio=false)));
end FVM_staggered_Ncells;
正确结果:
好的。。经过一番挖掘,我找到了答案。下面我显示了“收到时”代码,然后在下面进行编辑。希望这能解决所有问题 背景,正如您所知,有一个非常重要的模型结构。您建模的是
av_vb
1。更正流模型的长度
对于av_vb
模型结构的第一卷和最后一卷,变量dL(流段长度)不同。此更正对于运行的案例最为重要
添加以下修改:
// Define the variable
SI.Length dLs[n-1];
SI.Momentum I[n-1] = mflow[2:n] .* dLs; // Changed from *dL to .*dLs
// Add to equation section
dLs[1] = dL + 0.5*dL;
dLs[2:n-2] = fill(dL,n-3);
dLs[n-1] = dL + 0.5*dL;
2。从dpf更改为mflow计算
我运行了一个带有恒定流量计算的简单案例,并检查了结果,发现即使进行了第一次修正,结果也不同。在指定设置下,“一对一”比较将使用mflow=f(dpf)时,似乎使用了dpf=f(mflow)计算。这是因为您选择了momentumDynamics=SteadyStateInitial
,这使得在PartialGenericPipeFlow
中的来自
。如果更改它,则恒定流量示例的结果将相同(两者之间的差异将更容易显示,因为它们不会被随时间变化的流量动力学所掩盖)
而且,我认为使用的平均密度与MSL管道不同。这并没有影响这个例子的计算,所以请仔细检查我的结论
if use_fixed_zeta then
dpf = 1/2*zeta/(n - 1)*(mflow[2:n]) .^ 2 ./ (0.5*(rho[1:n - 1] + rho[2:n])*
A*A);
else
// This was the original
// dpf = homotopy(
// actual = Modelica.Fluid.Pipes.BaseClasses.WallFriction.Detailed.pressureLoss_m_flow(
// m_flow = mflow[2:n],
// rho_a = rho[1:n-1],
// rho_b = rho[2:n],
// mu_a = mu[1:n-1],
// mu_b = mu[2:n],
// length = dLs, //Notice changed dL to dLs
// diameter = D,
// roughness = R,
// m_flow_small = 0.001),
// simplified = dp_nominal/(n-1)/m_flow_nominal*mflow[2:n]);
// This is the correct model for "one-to-one" comparison for the chosen conditions. Averaged rho and mu was used since useUpstreamScheme = false.
mflow[2:n] = homotopy(actual=
Modelica.Fluid.Pipes.BaseClasses.WallFriction.Detailed.massFlowRate_dp(
dpf,
0.5*(rho[1:n - 1] + rho[2:n]),
0.5*(rho[1:n - 1] + rho[2:n]),
0.5*(mu[1:n - 1] + mu[2:n]),
0.5*(mu[1:n - 1] + mu[2:n]),
dLs,
D,
A,
R,
1e-5,
4000), simplified=m_flow_nominal/dp_nominal .* dpf);
end if;
3。正确的端口b.m\U流量参考
这是另一次编辑,不会影响此计算结果,但可能会影响其他计算结果
// Original
mflow[n] = -port_b.m_flow;
// Fixed to reference proper flow variable
mflow[n+1] = -port_b.m_flow;
下面是您生成的相同绘图。情节重叠
您是否愿意为您的示例发布代码?您的测试/比较模型是一个很好的方法!也许您想添加更多的管道模型,例如库中的管道,或或中的管道。MSL中的管道模型有很多选项,我想您已经使用了!?特别是高级->模型结构中的设置,可能还有假设->动力学。使用的图元越多,MSL管道模型应该越精确。因此,如果您的模型给出与MSL模型相同的结果,例如300个元素,那么您的模型似乎是正确的。@当然是!如上所述,我更新了帖子。@matth我将测试您建议的其他库,然后将结果添加到此处。是的,我测试了所有设置并试图了解整个模型,即使这是一个很大的工作:/I我添加了上面301个单元格的结果。你怎么认为?我真的很想理解为什么我得到了与MSL模型相比的这种偏差,因为在我看来,这个问题是数值性质的。这似乎不是一个大错误,但为了积累我的知识,理解它会很好…第1点解决了问题!第2点我没有改变dp=f(m_流量),因为我按照第1点得到了正确的结果。在PartialGenericPipeFlow中,您可以从以下位置找到参数布尔值:dp=momentumDynamics>=Types.Dynamics.SteadyStateInitial“=true,使用m\u flow=f(dp),否则dp=f(m\u flow)”当我使用steadystate初始化时,应该使用dp=f(m\u flow)。您是否使用了SteadyStateInitial?你说得对。平均密度:函数pressureLoss_m_flow内部上游离散。做中心讨论。需要您的修改。第三点:thx@塞吉:很高兴我能帮忙。关于来自\u dp的。通过选择momentumDynamics=Dynamics.SteadyStateInitial
(您这样做了),然后从\u dp=true
中选择。因此,您应该使用m_flow=f(dp),但您使用的是dp=f(m_flow)。如果您选择DynamicFreeInitial或FixedInitial,则从\u dp=false中选择,然后执行dp=f(m\u flow)。这是因为枚举类型是“编号”的。DynamicFreeInitial=1,FixedInitial=2,SteadyStateInitial=3,SteadyState=4。因此from_dp=SteadyStateInitial(3)>