XC::Response *XC::DispBeamColumn2d::setResponse(const std::vector<std::string> &argv, Information &eleInfo)
{
const int argc= argv.size();
const size_t numSections= getNumSections();
// global force -
if(argv[0] == "forces" || argv[0] == "force"
|| argv[0] == "globalForce" || argv[0] == "globalForces")
return new ElementResponse(this, 1, P);
// local force -
else if(argv[0] == "localForce" || argv[0] == "localForces")
return new ElementResponse(this, 2, P);
// chord rotation -
else if(argv[0] == "chordRotation" || argv[0] == "chordDeformation"
|| argv[0] == "basicDeformation")
return new ElementResponse(this, 3, Vector(3));
// plastic rotation -
else if(argv[0] == "plasticRotation" || argv[0] == "plasticDeformation")
return new ElementResponse(this, 4, Vector(3));
// section response -
else if(argv[0] == "section" || argv[0] == "-section") {
if(argc <= 2)
return 0;
size_t sectionNum = atoi(argv[1]);
if(sectionNum > 0 && sectionNum <= numSections)
{
std::vector<std::string> argv2(argv);
argv2.erase(argv2.begin(),argv2.begin()+2);
return theSections[sectionNum-1]->setResponse(argv2, eleInfo);
}
else
return 0;
}
else
return 0;
}
int XC::DispBeamColumn2d::update(void)
{
// Update the transformation
theCoordTransf->update();
// Get basic deformations
const Vector &v= theCoordTransf->getBasicTrialDisp();
const double L = theCoordTransf->getInitialLength();
const double oneOverL = 1.0/L;
const size_t numSections= getNumSections();
const Matrix &pts = quadRule.getIntegrPointCoords(numSections);
// Loop over the integration points
for(size_t i = 0; i < numSections; i++) {
int order = theSections[i]->getOrder();
const ID &code = theSections[i]->getType();
Vector e(workArea, order);
double xi6 = 6.0*pts(i,0);
int j;
for(j = 0; j < order; j++) {
switch(code(j)) {
case SECTION_RESPONSE_P:
e(j) = oneOverL*v(0); break;
case SECTION_RESPONSE_MZ:
e(j) = oneOverL*((xi6-4.0)*v(1) + (xi6-2.0)*v(2)); break;
default:
e(j) = 0.0; break;
}
}
// Set the section deformations
theSections[i]->setTrialSectionDeformation(e);
}
return 0;
}
int
XC::DispBeamColumn2d::updateParameter (int parameterID, Information &info)
{
// If the parameterID value is not equal to 1 it belongs
// to section or material further down in the hierarchy.
const size_t numSections= getNumSections();
if(parameterID == 1) {
this->rho = info.theDouble;
return 0;
}
else if(parameterID > 0 ) {
// Extract the section number
int sectionNumber = (int)( floor((double)parameterID) / (100000) );
int ok = -1;
for(size_t i=0; i<numSections; i++) {
if(sectionNumber == theSections[i]->getTag()) {
ok = theSections[i]->updateParameter(parameterID, info);
}
}
if(ok < 0) {
std::cerr << "XC::DispBeamColumn2d::updateParameter() - could not update parameter. " << std::endl;
return ok;
}
else {
return ok;
}
}
else {
std::cerr << "XC::DispBeamColumn2d::updateParameter() - could not update parameter. " << std::endl;
return -1;
}
}
int
XC::DispBeamColumn2d::activateParameter(int passedParameterID)
{
// Note that the parameteID that is stored here at the
// element level contains all information about section
// and material tag number:
parameterID = passedParameterID;
const size_t numSections= getNumSections();
if(passedParameterID == 0 ) {
// "Zero out" all flags downwards through sections/materials
for(size_t i=0; i<numSections; i++) {
theSections[i]->activateParameter(passedParameterID);
}
}
else if(passedParameterID == 1) {
// Don't treat the 'rho' for now
}
else {
// Extract section and material tags from the passedParameterID
int activeSectionTag = (int)( floor((double)passedParameterID) / (100000) );
// Go down to the sections and set appropriate flags
for(size_t i=0; i<numSections; i++) {
if(activeSectionTag == theSections[i]->getTag()) {
theSections[i]->activateParameter(passedParameterID);
}
}
}
return 0;
}
// NEW METHOD
int XC::DispBeamColumn2d::commitSensitivity(int gradNumber, int numGrads)
{
// Get basic deformation and sensitivities
const XC::Vector &v = theCoordTransf->getBasicTrialDisp();
static XC::Vector vsens(3);
vsens = theCoordTransf->getBasicDisplSensitivity(gradNumber);
double L = theCoordTransf->getInitialLength();
double oneOverL = 1.0/L;
const size_t numSections= getNumSections();
const Matrix &pts = quadRule.getIntegrPointCoords(numSections);
// Some extra declarations
double d1oLdh = 0.0;
// Check if a nodal coordinate is random
bool randomNodeCoordinate = false;
static XC::ID nodeParameterID(2);
nodeParameterID(0) = theNodes[0]->getCrdsSensitivity();
nodeParameterID(1) = theNodes[1]->getCrdsSensitivity();
if(nodeParameterID(0) != 0 || nodeParameterID(1) != 0) {
vsens += theCoordTransf->getBasicTrialDispShapeSensitivity();
randomNodeCoordinate = true;
const XC::Vector &ndICoords = theNodes[0]->getCrds();
const XC::Vector &ndJCoords = theNodes[1]->getCrds();
double dx = ndJCoords(0) - ndICoords(0);
double dy = ndJCoords(1) - ndICoords(1);
if(nodeParameterID(0) == 1) // here x1 is random
d1oLdh = dx/(L*L*L);
if(nodeParameterID(0) == 2) // here y1 is random
d1oLdh = dy/(L*L*L);
if(nodeParameterID(1) == 1) // here x2 is random
d1oLdh = -dx/(L*L*L);
if(nodeParameterID(1) == 2) // here y2 is random
d1oLdh = -dy/(L*L*L);
}
// Loop over the integration points
for(size_t i = 0; i < numSections; i++) {
int order = theSections[i]->getOrder();
const XC::ID &code = theSections[i]->getType();
Vector e(workArea, order);
double xi6 = 6.0*pts(i,0);
for(int j = 0; j < order; j++) {
switch(code(j)) {
case SECTION_RESPONSE_P:
e(j) = oneOverL*vsens(0)
+ d1oLdh*v(0);
break;
case SECTION_RESPONSE_MZ:
e(j) = oneOverL*((xi6-4.0)*vsens(1) + (xi6-2.0)*vsens(2))
+ d1oLdh*((xi6-4.0)*v(1) + (xi6-2.0)*v(2));
break;
default:
e(j) = 0.0;
break;
}
}
// Set the section deformations
theSections[i]->commitSensitivity(e,gradNumber,numGrads);
}
return 0;
}
const XC::Vector &XC::DispBeamColumn2d::getResistingForceSensitivity(int gradNumber)
{
const size_t numSections= getNumSections();
const Matrix &pts = quadRule.getIntegrPointCoords(numSections);
const Vector &wts = quadRule.getIntegrPointWeights(numSections);
double L = theCoordTransf->getInitialLength();
double oneOverL = 1.0/L;
// Zero for integration
q.Zero();
static XC::Vector qsens(3);
qsens.Zero();
// Some extra declarations
static XC::Matrix kbmine(3,3);
kbmine.Zero();
int j, k;
double d1oLdh = 0.0;
// Check if a nodal coordinate is random
bool randomNodeCoordinate = false;
static XC::ID nodeParameterID(2);
nodeParameterID(0) = theNodes[0]->getCrdsSensitivity();
nodeParameterID(1) = theNodes[1]->getCrdsSensitivity();
if(nodeParameterID(0) != 0 || nodeParameterID(1) != 0) {
randomNodeCoordinate = true;
const XC::Vector &ndICoords = theNodes[0]->getCrds();
const XC::Vector &ndJCoords = theNodes[1]->getCrds();
double dx = ndJCoords(0) - ndICoords(0);
double dy = ndJCoords(1) - ndICoords(1);
if(nodeParameterID(0) == 1) // here x1 is random
d1oLdh = dx/(L*L*L);
if(nodeParameterID(0) == 2) // here y1 is random
d1oLdh = dy/(L*L*L);
if(nodeParameterID(1) == 1) // here x2 is random
d1oLdh = -dx/(L*L*L);
if(nodeParameterID(1) == 2) // here y2 is random
d1oLdh = -dy/(L*L*L);
}
// Loop over the integration points
for(size_t i= 0; i < numSections; i++) {
int order = theSections[i]->getOrder();
const XC::ID &code = theSections[i]->getType();
double xi6 = 6.0*pts(i,0);
double wti = wts(i);
// Get section stress resultant gradient
const XC::Vector &s = theSections[i]->getStressResultant();
const XC::Vector &sens = theSections[i]->getStressResultantSensitivity(gradNumber,true);
// Perform numerical integration on internal force gradient
//q.addMatrixTransposeVector(1.0, *B, s, wts(i));
double si;
double sensi;
for(j = 0; j < order; j++) {
si = s(j)*wti;
sensi = sens(j)*wti;
switch(code(j)) {
case SECTION_RESPONSE_P:
q(0) += si;
qsens(0) += sensi;
break;
case SECTION_RESPONSE_MZ:
q(1) += (xi6-4.0)*si;
q(2) += (xi6-2.0)*si;
qsens(1) += (xi6-4.0)*sensi;
qsens(2) += (xi6-2.0)*sensi;
break;
default:
break;
}
}
if(randomNodeCoordinate) {
// Perform numerical integration to obtain basic stiffness matrix
//kb.addMatrixTripleProduct(1.0, *B, ks, wts(i)/L);
double tmp;
const XC::Matrix &ks = theSections[i]->getSectionTangent();
Matrix ka(workArea, order, 3);
ka.Zero();
for(j = 0; j < order; j++) {
switch(code(j)) {
case SECTION_RESPONSE_P:
for(k = 0; k < order; k++) {
ka(k,0) += ks(k,j)*wti;
}
break;
case SECTION_RESPONSE_MZ:
for(k = 0; k < order; k++) {
tmp = ks(k,j)*wti;
ka(k,1) += (xi6-4.0)*tmp;
ka(k,2) += (xi6-2.0)*tmp;
}
break;
default:
break;
}
}
for(j = 0; j < order; j++) {
switch (code(j)) {
case SECTION_RESPONSE_P:
for(k = 0; k < 3; k++) {
kbmine(0,k) += ka(j,k);
}
break;
case SECTION_RESPONSE_MZ:
for(k = 0; k < 3; k++) {
tmp = ka(j,k);
kbmine(1,k) += (xi6-4.0)*tmp;
kbmine(2,k) += (xi6-2.0)*tmp;
}
break;
default:
break;
}
}
}
}
static XC::Vector dqdh(3);
const XC::Vector &dAdh_u = theCoordTransf->getBasicTrialDispShapeSensitivity();
//dqdh = (1.0/L) * (kbmine * dAdh_u);
dqdh.addMatrixVector(0.0, kbmine, dAdh_u, oneOverL);
static XC::Vector dkbdh_v(3);
const XC::Vector &A_u = theCoordTransf->getBasicTrialDisp();
//dkbdh_v = (d1oLdh) * (kbmine * A_u);
dkbdh_v.addMatrixVector(0.0, kbmine, A_u, d1oLdh);
// Transform forces
static XC::Vector dummy(3); // No distributed loads
// Term 5
P = theCoordTransf->getGlobalResistingForce(qsens,dummy);
if(randomNodeCoordinate) {
// Term 1
P += theCoordTransf->getGlobalResistingForceShapeSensitivity(q,dummy);
// Term 2
P += theCoordTransf->getGlobalResistingForce(dqdh,dummy);
// Term 4
P += theCoordTransf->getGlobalResistingForce(dkbdh_v,dummy);
}
return P;
}
// AddingSensitivity:BEGIN ///////////////////////////////////
int XC::DispBeamColumn2d::setParameter(const std::vector<std::string> &argv, Parameter ¶m)
{
//
// From the parameterID value it should be possible to extract
// information about:
// 1) Which parameter is in question. The parameter could
// be at element, section, or material level.
// 2) Which section and material number (tag) it belongs to.
//
// To accomplish this the parameterID is given the following value:
// parameterID = type + 1000*matrTag + 100000*sectionTag
// ...where 'type' is an integer in the range (1-99) and added 100
// for each level (from material to section to element).
//
// Example:
// If 'E0' (case 2) is random in material #3 of section #5
// the value of the parameterID at this (element) level would be:
// parameterID = 2 + 1000*3 + 100000*5 = 503002
// As seen, all given information can be extracted from this number.
//
// Initial declarations
int ok = -1;
const size_t numSections= getNumSections();
const size_t argc= argv.size();
// If the parameter belongs to the element itself
if(argv[0] == "rho")
return param.addObject(1, this);
// If the parameter is belonging to a section or lower
else if(argv[0] == "section" || argv[0] == "-section")
{
// For now, no parameters of the section itself:
if(argc<5) {
std::cerr << "For now: cannot handle parameters of the section itself." << std::endl;
return -1;
}
// Get section and material tag numbers from user input
int paramSectionTag = atoi(argv[1]);
// Find the right section and call its setParameter method
for(size_t i=0; i<numSections; i++)
{
if(paramSectionTag == theSections[i]->getTag())
{
std::vector<std::string> argv2(argv);
argv2.erase(argv2.begin(),argv2.begin()+2);
ok = theSections[i]->setParameter(argv2, param);
}
}
// Check if the ok is valid
if(ok < 0) {
std::cerr << "XC::DispBeamColumn2d::setParameter() - could not set parameter. " << std::endl;
return -1;
}
else {
// Return the ok value (according to the above comments)
return ok;
}
}
// Otherwise parameter is unknown for this class
else {
return -1;
}
}
const XC::Vector &XC::DispBeamColumn2d::getResistingForce(void) const
{
const size_t numSections= getNumSections();
const Matrix &pts = quadRule.getIntegrPointCoords(numSections);
const Vector &wts = quadRule.getIntegrPointWeights(numSections);
// Zero for integration
q.Zero();
// Loop over the integration points
for(size_t i= 0; i < numSections; i++) {
int order = theSections[i]->getOrder();
const XC::ID &code = theSections[i]->getType();
double xi6 = 6.0*pts(i,0);
// Get section stress resultant
const XC::Vector &s = theSections[i]->getStressResultant();
// Perform numerical integration on internal force
//q.addMatrixTransposeVector(1.0, *B, s, wts(i));
double si;
for(int j = 0; j < order; j++) {
si = s(j)*wts(i);
switch(code(j)) {
case SECTION_RESPONSE_P:
q(0) += si; break;
case SECTION_RESPONSE_MZ:
q(1) += (xi6-4.0)*si; q(2) += (xi6-2.0)*si; break;
default:
break;
}
}
}
// Add effects of element loads, q = q(v) + q0
q(0) += q0[0];
q(1) += q0[1];
q(2) += q0[2];
// Vector for reactions in basic system
Vector p0Vec= p0.getVector();
P = theCoordTransf->getGlobalResistingForce(q, p0Vec);
// Subtract other external nodal loads ... P_res = P_int - P_ext
//P.addVector(1.0, load, -1.0);
P(0) -= load(0);
P(1) -= load(1);
P(2) -= load(2);
P(3) -= load(3);
P(4) -= load(4);
P(5) -= load(5);
if(isDead())
P*=dead_srf;
return P;
}
const XC::Matrix &XC::DispBeamColumn2d::getInitialBasicStiff(void) const
{
static Matrix kb(3,3);
// Zero for integral
kb.Zero();
const size_t numSections= getNumSections();
const Matrix &pts = quadRule.getIntegrPointCoords(numSections);
const Vector &wts = quadRule.getIntegrPointWeights(numSections);
double L = theCoordTransf->getInitialLength();
double oneOverL = 1.0/L;
// Loop over the integration points
for(size_t i= 0;i<numSections;i++) {
int order = theSections[i]->getOrder();
const XC::ID &code = theSections[i]->getType();
Matrix ka(workArea, order, 3);
ka.Zero();
double xi6 = 6.0*pts(i,0);
// Get the section tangent stiffness and stress resultant
const XC::Matrix &ks = theSections[i]->getInitialTangent();
// Perform numerical integration
//kb.addMatrixTripleProduct(1.0, *B, ks, wts(i)/L);
double wti = wts(i)*oneOverL;
double tmp;
int j, k;
for(j = 0; j < order; j++) {
switch(code(j)) {
case SECTION_RESPONSE_P:
for(k = 0; k < order; k++)
ka(k,0) += ks(k,j)*wti;
break;
case SECTION_RESPONSE_MZ:
for(k = 0; k < order; k++) {
tmp = ks(k,j)*wti;
ka(k,1) += (xi6-4.0)*tmp;
ka(k,2) += (xi6-2.0)*tmp;
}
break;
default:
break;
}
}
for(j = 0; j < order; j++) {
switch (code(j)) {
case SECTION_RESPONSE_P:
for(k = 0; k < 3; k++)
kb(0,k) += ka(j,k);
break;
case SECTION_RESPONSE_MZ:
for(k = 0; k < 3; k++) {
tmp = ka(j,k);
kb(1,k) += (xi6-4.0)*tmp;
kb(2,k) += (xi6-2.0)*tmp;
}
break;
default:
break;
}
}
}
return kb;
}
const XC::Matrix &XC::DispBeamColumn2d::getTangentStiff(void) const
{
static Matrix kb(3,3);
this->getBasicStiff(kb);
// Zero for integral
q.Zero();
const size_t numSections= getNumSections();
const Matrix &pts = quadRule.getIntegrPointCoords(numSections);
const Vector &wts = quadRule.getIntegrPointWeights(numSections);
const double L = theCoordTransf->getInitialLength();
const double oneOverL = 1.0/L;
// Loop over the integration points
for(size_t i = 0; i < numSections; i++)
{
int order = theSections[i]->getOrder();
const XC::ID &code = theSections[i]->getType();
Matrix ka(workArea, order, 3);
ka.Zero();
double xi6 = 6.0*pts(i,0);
// Get the section tangent stiffness and stress resultant
const XC::Matrix &ks = theSections[i]->getSectionTangent();
const XC::Vector &s = theSections[i]->getStressResultant();
// Perform numerical integration
//kb.addMatrixTripleProduct(1.0, *B, ks, wts(i)/L);
double wti = wts(i)*oneOverL;
double tmp;
int j, k;
for(j = 0; j < order; j++) {
switch(code(j)) {
case SECTION_RESPONSE_P:
for(k = 0; k < order; k++)
ka(k,0) += ks(k,j)*wti;
break;
case SECTION_RESPONSE_MZ:
for(k = 0; k < order; k++) {
tmp = ks(k,j)*wti;
ka(k,1) += (xi6-4.0)*tmp;
ka(k,2) += (xi6-2.0)*tmp;
}
break;
default:
break;
}
}
for(j = 0; j < order; j++) {
switch (code(j)) {
case SECTION_RESPONSE_P:
for(k = 0; k < 3; k++)
kb(0,k) += ka(j,k);
break;
case SECTION_RESPONSE_MZ:
for(k = 0; k < 3; k++) {
tmp = ka(j,k);
kb(1,k) += (xi6-4.0)*tmp;
kb(2,k) += (xi6-2.0)*tmp;
}
break;
default:
break;
}
}
//q.addMatrixTransposeVector(1.0, *B, s, wts(i));
double si;
for(j = 0; j < order; j++)
{
si = s(j)*wts(i);
switch(code(j)) {
case SECTION_RESPONSE_P:
q(0) += si; break;
case SECTION_RESPONSE_MZ:
q(1) += (xi6-4.0)*si; q(2) += (xi6-2.0)*si; break;
default:
break;
}
}
}
// Add effects of element loads, q = q(v) + q0
q(0) += q0[0];
q(1) += q0[1];
q(2) += q0[2];
// Transform to global stiffness
K = theCoordTransf->getGlobalStiffMatrix(kb, q);
if(isDead())
K*=dead_srf;
return K;
}
const BinarySection *BinaryImage::getSectionByIndex(int idx) const
{
return Util::inRange(idx, 0, getNumSections()) ? m_sections[idx] : nullptr;
}
