Diffusion lammps (efficient_neuronet) Browse

pair_smd_ulsph.cpp
354e20f43157efficient_neuronet

pair_smd_ulsph.cpp
View Options

			/* ----------------------------------------------------------------------
			*
			* * Smooth Mach Dynamics *
			*
			* This file is part of the USER-SMD package for LAMMPS.
			* Copyright (2014) Georg C. Ganzenmueller, georg.ganzenmueller@emi.fhg.de
			* Fraunhofer Ernst-Mach Institute for High-Speed Dynamics, EMI,
			* Eckerstrasse 4, D-79104 Freiburg i.Br, Germany.
			*
			* ----------------------------------------------------------------------- */

			/* ----------------------------------------------------------------------
			LAMMPS - Large-scale Atomic/Molecular Massively Parallel Simulator
			http://lammps.sandia.gov, Sandia National Laboratories
			Steve Plimpton, sjplimp@sandia.gov

			Copyright (2003) Sandia Corporation. Under the terms of Contract
			DE-AC04-94AL85000 with Sandia Corporation, the U.S. Government retains
			certain rights in this software. This software is distributed under
			the GNU General Public License.

			See the README file in the top-level LAMMPS directory.
			------------------------------------------------------------------------- */

			#include <math.h>
			#include <float.h>
			#include <stdlib.h>
			#include <string.h>
			#include "pair_smd_ulsph.h"
			#include "atom.h"
			#include "domain.h"
			#include "force.h"
			#include "update.h"
			#include "modify.h"
			#include "fix.h"
			#include "comm.h"
			#include "neighbor.h"
			#include "neigh_list.h"
			#include "neigh_request.h"
			#include "memory.h"
			#include "error.h"
			#include <stdio.h>
			#include <iostream>
			#include "smd_material_models.h"
			#include "smd_math.h"
			#include "smd_kernels.h"

			using namespace SMD_Kernels;
			using namespace std;
			using namespace LAMMPS_NS;
			using namespace SMD_Math;

			#include <Eigen/Eigen>
			using namespace Eigen;

			#define ARTIFICIAL_STRESS false
			#define FORMAT1 "%60s : %g\n"
			#define FORMAT2 "\n.............................. %s \n"

			PairULSPH::PairULSPH(LAMMPS *lmp) :
			Pair(lmp) {

			// per-type arrays
			Q1 = NULL;
			eos = viscosity = strength = NULL;
			c0_type = NULL;
			c0 = NULL;
			Lookup = NULL;
			artificial_stress = NULL;
			artificial_pressure = NULL;

			nmax = 0; // make sure no atom on this proc such that initial memory allocation is correct
			stressTensor = L = K = NULL;
			shepardWeight = NULL;
			smoothVel = NULL;
			numNeighs = NULL;
			F = NULL;
			rho = NULL;
			effm = NULL;

			velocity_gradient_required = false; // turn off computation of velocity gradient by default
			density_summation = velocity_gradient = false;

			comm_forward = 18; // this pair style communicates 18 doubles to ghost atoms
			updateFlag = 0;
			}

			/* ---------------------------------------------------------------------- */

			PairULSPH::~PairULSPH() {
			if (allocated) {
			//printf("... deallocating\n");
			memory->destroy(Q1);
			memory->destroy(rho0);
			memory->destroy(eos);
			memory->destroy(viscosity);
			memory->destroy(strength);
			memory->destroy(c0_type);
			memory->destroy(Lookup);
			memory->destroy(artificial_pressure);
			memory->destroy(artificial_stress);

			delete[] onerad_dynamic;
			delete[] onerad_frozen;
			delete[] maxrad_dynamic;
			delete[] maxrad_frozen;

			delete[] K;
			delete[] shepardWeight;
			delete[] c0;
			delete[] smoothVel;
			delete[] stressTensor;
			delete[] L;
			delete[] numNeighs;
			delete[] F;
			delete[] rho;
			delete[] effm;

			}
			}

			/* ----------------------------------------------------------------------
			*
			* Re-compute mass density from scratch.
			* Only used for plain fluid SPH with no physical viscosity models.
			*
			---------------------------------------------------------------------- */

			void PairULSPH::PreCompute_DensitySummation() {
			double *radius = atom->radius;
			double **x = atom->x;
			double *rmass = atom->rmass;
			int *type = atom->type;
			int ilist, jlist, *numneigh;
			int **firstneigh;
			int nlocal = atom->nlocal;
			int inum, jnum, ii, jj, i, itype, jtype, j;
			double h, irad, hsq, rSq, wf;
			Vector3d dx, xi, xj;

			// set up neighbor list variables
			inum = list->inum;
			ilist = list->ilist;
			numneigh = list->numneigh;
			firstneigh = list->firstneigh;

			// zero accumulators
			for (i = 0; i < nlocal; i++) {
			rho[i] = 0.0;
			//shepardWeight[i] = 0.0;
			}

			/*
			* only recompute mass density if density summation is used.
			* otherwise, change in mass density is time-integrated
			*/
			for (i = 0; i < nlocal; i++) {
			itype = type[i];
			if (setflag[itype][itype] == 1) {
			// initialize particle density with self-contribution.
			h = 2.0 * radius[i];
			hsq = h * h;
			Poly6Kernel(hsq, h, 0.0, domain->dimension, wf);
			rho[i] = wf * rmass[i]; // / shepardWeight[i];
			//printf("SIC to rho is %f\n", rho[i]);
			}
			}

			for (ii = 0; ii < inum; ii++) {
			i = ilist[ii];
			itype = type[i];
			jlist = firstneigh[i];
			jnum = numneigh[i];
			irad = radius[i];

			xi << x[i][0], x[i][1], x[i][2];

			for (jj = 0; jj < jnum; jj++) {
			j = jlist[jj];
			j &= NEIGHMASK;

			xj << x[j][0], x[j][1], x[j][2];
			dx = xj - xi;
			rSq = dx.squaredNorm();
			h = irad + radius[j];
			hsq = h * h;
			if (rSq < hsq) {

			jtype = type[j];
			Poly6Kernel(hsq, h, rSq, domain->dimension, wf);

			if (setflag[itype][itype] == 1) {
			rho[i] += wf * rmass[j]; // / shepardWeight[i];
			}

			if (j < nlocal) {
			if (setflag[jtype][jtype] == 1) {
			rho[j] += wf * rmass[i]; // / shepardWeight[j];
			}
			}
			} // end if check distance
			} // end loop over j
			} // end loop over i
			}

			/* ----------------------------------------------------------------------
			*
			* Compute shape matrix for kernel gradient correction and velocity gradient.
			* This is used if material strength or viscosity models are employed.
			*
			---------------------------------------------------------------------- */

			void PairULSPH::PreCompute() {
			double **atom_data9 = atom->smd_data_9;
			double *radius = atom->radius;
			double **x = atom->x;
			double **x0 = atom->x0;
			double **v = atom->vest;
			double *vfrac = atom->vfrac;
			int *type = atom->type;
			int ilist, jlist, *numneigh;
			int **firstneigh;
			int nlocal = atom->nlocal;
			int inum, jnum, ii, jj, i, itype, j, idim;
			double wfd, h, irad, r, rSq, wf, ivol, jvol;
			Vector3d dx, dv, g, du;
			Matrix3d Ktmp, Ltmp, Ftmp, K3di, D;
			Vector3d xi, xj, vi, vj, x0i, x0j, dx0;
			Matrix2d K2di, K2d;

			// zero accumulators
			for (i = 0; i < nlocal; i++) {
			itype = type[i];
			if (setflag[itype][itype]) {
			if (gradient_correction_flag) {
			K[i].setZero();
			} else {
			K[i].setIdentity();
			}
			L[i].setZero();
			F[i].setZero();
			}
			}

			// set up neighbor list variables
			inum = list->inum;
			ilist = list->ilist;
			numneigh = list->numneigh;
			firstneigh = list->firstneigh;

			for (ii = 0; ii < inum; ii++) {
			i = ilist[ii];
			itype = type[i];
			jlist = firstneigh[i];
			jnum = numneigh[i];
			irad = radius[i];
			ivol = vfrac[i];

			// initialize Eigen data structures from LAMMPS data structures
			for (idim = 0; idim < 3; idim++) {
			x0i(idim) = x0[i][idim];
			xi(idim) = x[i][idim];
			vi(idim) = v[i][idim];
			}

			for (jj = 0; jj < jnum; jj++) {
			j = jlist[jj];
			j &= NEIGHMASK;

			for (idim = 0; idim < 3; idim++) {
			x0j(idim) = x0[j][idim];
			xj(idim) = x[j][idim];
			vj(idim) = v[j][idim];
			}

			dx = xj - xi;

			rSq = dx.squaredNorm();
			h = irad + radius[j];
			if (rSq < h * h) {

			r = sqrt(rSq);
			jvol = vfrac[j];

			// distance vectors in current and reference configuration, velocity difference
			dv = vj - vi;
			dx0 = x0j - x0i;

			// kernel and derivative
			spiky_kernel_and_derivative(h, r, domain->dimension, wf, wfd);
			//barbara_kernel_and_derivative(h, r, domain->dimension, wf, wfd);

			// uncorrected kernel gradient
			g = (wfd / r) * dx;

			/* build correction matrix for kernel derivatives */
			if (gradient_correction_flag) {
			Ktmp = -g * dx.transpose();
			K[i] += jvol * Ktmp;
			}

			// velocity gradient L
			Ltmp = -dv * g.transpose();
			L[i] += jvol * Ltmp;

			// deformation gradient F in Eulerian frame
			du = dx - dx0;
			Ftmp = dv * g.transpose();
			F[i] += jvol * Ftmp;

			if (j < nlocal) {

			if (gradient_correction_flag) {
			K[j] += ivol * Ktmp;
			}

			L[j] += ivol * Ltmp;
			F[j] += ivol * Ftmp;
			}
			} // end if check distance
			} // end loop over j
			} // end loop over i

			/*
			* invert shape matrix and compute corrected quantities
			*/

			for (i = 0; i < nlocal; i++) {
			itype = type[i];
			if (setflag[itype][itype]) {
			if (gradient_correction_flag) {
			pseudo_inverse_SVD(K[i]);
			K[i] = LimitEigenvalues(K[i], 2.0);
			L[i] *= K[i];
			F[i] *= K[i];
			} // end if (gradient_correction[itype]) {

			/*
			* accumulate strain increments
			* we abuse the atom array "atom_data_9" for this purpose, which was originally designed to hold the deformation gradient.
			*/
			D = update->dt * 0.5 * (L[i] + L[i].transpose());
			atom_data9[i][0] += D(0, 0); // xx
			atom_data9[i][1] += D(1, 1); // yy
			atom_data9[i][2] += D(2, 2); // zz
			atom_data9[i][3] += D(0, 1); // xy
			atom_data9[i][4] += D(0, 2); // xz
			atom_data9[i][5] += D(1, 2); // yz

			} // end if (setflag[itype][itype])
			} // end loop over i = 0 to nlocal

			}

			/* ---------------------------------------------------------------------- */

			void PairULSPH::compute(int eflag, int vflag) {
			double **x = atom->x;
			double **v = atom->vest;
			double **vint = atom->v; // Velocity-Verlet algorithm velocities
			double **f = atom->f;
			double *vfrac = atom->vfrac;
			double *de = atom->de;
			double *rmass = atom->rmass;
			double *radius = atom->radius;
			double *contact_radius = atom->contact_radius;
			double **atom_data9 = atom->smd_data_9;

			int *type = atom->type;
			int nlocal = atom->nlocal;
			int i, j, ii, jj, jnum, itype, jtype, iDim, inum;
			double r, wf, wfd, h, rSq, ivol, jvol;
			double mu_ij, c_ij, rho_ij;
			double delVdotDelR, visc_magnitude, deltaE;
			int ilist, jlist, *numneigh;
			int **firstneigh;
			Vector3d fi, fj, dx, dv, f_stress, g, vinti, vintj, dvint;
			Vector3d xi, xj, vi, vj, f_visc, sumForces, f_stress_new;
			Vector3d gamma, f_hg, dx0, du_est, du;
			double r_ref, weight, p;
			//int periodic = (domain->xperiodic \|\| domain->yperiodic \|\| domain->zperiodic);

			double ini_dist;
			Matrix3d S, D, V, eye;
			eye.setIdentity();
			int k;
			SelfAdjointEigenSolver < Matrix3d > es;

			if (eflag \|\| vflag)
			ev_setup(eflag, vflag);
			else
			evflag = vflag_fdotr = 0;

			if (atom->nmax > nmax) {
			//printf("... allocating in compute with nmax = %d\n", atom->nmax);
			nmax = atom->nmax;
			delete[] K;
			K = new Matrix3d[nmax];
			delete[] shepardWeight;
			shepardWeight = new double[nmax];
			delete[] c0;
			c0 = new double[nmax];
			delete[] smoothVel;
			smoothVel = new Vector3d[nmax];
			delete[] stressTensor;
			stressTensor = new Matrix3d[nmax];
			delete[] L;
			L = new Matrix3d[nmax];
			delete[] numNeighs;
			numNeighs = new int[nmax];
			delete[] F;
			F = new Matrix3d[nmax];
			delete[] rho;
			rho = new double[nmax];
			delete[] effm;
			effm = new double[nmax];
			}

			// zero accumulators
			for (i = 0; i < nlocal; i++) {
			shepardWeight[i] = 0.0;
			smoothVel[i].setZero();
			numNeighs[i] = 0;

			h = 2.0 * radius[i];
			r = 0.0;
			spiky_kernel_and_derivative(h, r, domain->dimension, wf, wfd);
			}

			/*
			* if this is the very first step, zero the array which holds the accumulated strain
			*/
			if (update->ntimestep == 0) {
			for (i = 0; i < nlocal; i++) {
			itype = type[i];
			if (setflag[itype][itype]) {
			for (j = 0; j < 9; j++) {
			atom_data9[i][j] = 0.0;
			}
			}
			}
			}

			if (density_summation) {
			//printf("dens summ\n");
			PreCompute_DensitySummation();

			for (i = 0; i < nlocal; i++) { //compute volumes from rho
			itype = type[i];
			if (setflag[itype][itype]) {
			vfrac[i] = rmass[i] / rho[i];
			}
			}

			}

			if (velocity_gradient) {
			PairULSPH::PreCompute(); // get velocity gradient and kernel gradient correction
			}

			PairULSPH::AssembleStressTensor();

			/*
			* QUANTITIES ABOVE HAVE ONLY BEEN CALCULATED FOR NLOCAL PARTICLES.
			* NEED TO DO A FORWARD COMMUNICATION TO GHOST ATOMS NOW
			*/
			comm->forward_comm_pair(this);

			updateFlag = 0;

			/*
			* iterate over pairs of particles i, j and assign forces using pre-computed pressure
			*/

			// set up neighbor list variables
			inum = list->inum;
			ilist = list->ilist;
			numneigh = list->numneigh;
			firstneigh = list->firstneigh;

			for (ii = 0; ii < inum; ii++) {
			i = ilist[ii];
			itype = type[i];
			jlist = firstneigh[i];
			jnum = numneigh[i];
			ivol = vfrac[i];

			// initialize Eigen data structures from LAMMPS data structures
			for (iDim = 0; iDim < 3; iDim++) {
			xi(iDim) = x[i][iDim];
			vi(iDim) = v[i][iDim];
			vinti(iDim) = vint[i][iDim];
			}

			for (jj = 0; jj < jnum; jj++) {
			j = jlist[jj];
			j &= NEIGHMASK;

			xj(0) = x[j][0];
			xj(1) = x[j][1];
			xj(2) = x[j][2];

			dx = xj - xi;
			rSq = dx.squaredNorm();
			h = radius[i] + radius[j];
			if (rSq < h * h) {

			// initialize Eigen data structures from LAMMPS data structures
			for (iDim = 0; iDim < 3; iDim++) {
			vj(iDim) = v[j][iDim];
			vintj(iDim) = vint[j][iDim];
			}

			r = sqrt(rSq);
			jtype = type[j];
			jvol = vfrac[j];

			// distance vectors in current and reference configuration, velocity difference
			dv = vj - vi;
			dvint = vintj - vinti;

			// kernel and derivative
			spiky_kernel_and_derivative(h, r, domain->dimension, wf, wfd);
			//barbara_kernel_and_derivative(h, r, domain->dimension, wf, wfd);

			// uncorrected kernel gradient
			g = (wfd / r) * dx;

			delVdotDelR = dx.dot(dv) / (r + 0.1 * h); // project relative velocity onto unit particle distance vector [m/s]

			S = stressTensor[i] + stressTensor[j];

			if (artificial_pressure[itype][jtype] > 0.0) {
			p = S.trace();
			if (p > 0.0) { // we are in tension
			r_ref = contact_radius[i] + contact_radius[j];
			weight = Kernel_Cubic_Spline(r, h) / Kernel_Cubic_Spline(r_ref, h);
			weight = pow(weight, 4.0);
			S -= artificial_pressure[itype][jtype] * weight * p * eye;
			}
			}

			/*
			* artificial stress to control tensile instability
			* Only works if particles are uniformly spaced initially.
			*/
			if (artificial_stress[itype][jtype] > 0.0) {
			ini_dist = contact_radius[i] + contact_radius[j];
			weight = Kernel_Cubic_Spline(r, h) / Kernel_Cubic_Spline(ini_dist, h);
			weight = pow(weight, 4.0);

			es.compute(S);
			D = es.eigenvalues().asDiagonal();
			for (k = 0; k < 3; k++) {
			if (D(k, k) > 0.0) {
			D(k, k) -= weight * artificial_stress[itype][jtype] * D(k, k);
			}
			}
			V = es.eigenvectors();
			S = V * D * V.inverse();
			}

			// compute forces
			f_stress = -ivol * jvol * S * g; // DO NOT TOUCH SIGN

			/*
			* artificial viscosity -- alpha is dimensionless
			* MonaghanBalsara form of the artificial viscosity
			*/

			c_ij = 0.5 * (c0[i] + c0[j]);
			LimitDoubleMagnitude(delVdotDelR, 1.1 * c_ij);

			mu_ij = h * delVdotDelR / (r + 0.1 * h); // units: [m * m/s / m = m/s]
			rho_ij = 0.5 * (rmass[i] / ivol + rmass[j] / jvol);
			visc_magnitude = 0.5 * (Q1[itype] + Q1[jtype]) * c_ij * mu_ij / rho_ij;
			f_visc = -rmass[i] * rmass[j] * visc_magnitude * g;

			if ((Lookup[HOURGLASS_CONTROL_AMPLITUDE][itype] > 0.0) && (Lookup[HOURGLASS_CONTROL_AMPLITUDE][jtype] > 0.0)) {
			f_hg = ComputeHourglassForce(i, itype, j, jtype, dv, dx, g, c_ij, mu_ij, rho_ij);

			} else {
			f_hg.setZero();
			}

			sumForces = f_stress + f_visc + f_hg;

			// energy rate -- project velocity onto force vector
			deltaE = sumForces.dot(dv);

			// apply forces to pair of particles
			f[i][0] += sumForces(0);
			f[i][1] += sumForces(1);
			f[i][2] += sumForces(2);
			de[i] += deltaE;

			// accumulate smooth velocities
			shepardWeight[i] += jvol * wf;
			smoothVel[i] += jvol * wf * dvint;
			numNeighs[i] += 1;

			if (j < nlocal) {
			f[j][0] -= sumForces(0);
			f[j][1] -= sumForces(1);
			f[j][2] -= sumForces(2);
			de[j] += deltaE;

			shepardWeight[j] += ivol * wf;
			smoothVel[j] -= ivol * wf * dvint;
			numNeighs[j] += 1;
			}

			// tally atomistic stress tensor
			if (evflag) {
			ev_tally_xyz(i, j, nlocal, 0, 0.0, 0.0, sumForces(0), sumForces(1), sumForces(2), dx(0), dx(1), dx(2));
			}
			}

			}
			}

			for (i = 0; i < nlocal; i++) {
			itype = type[i];
			if (setflag[itype][itype] == 1) {
			if (shepardWeight[i] != 0.0) {
			smoothVel[i] /= shepardWeight[i];
			} else {
			smoothVel[i].setZero();
			}
			} // end check if particle is SPH-type
			} // end loop over i = 0 to nlocal

			if (vflag_fdotr)
			virial_fdotr_compute();

			}

			/* ----------------------------------------------------------------------
			Assemble total stress tensor with pressure, material sterength, and
			viscosity contributions.
			------------------------------------------------------------------------- */
			void PairULSPH::AssembleStressTensor() {
			double *radius = atom->radius;
			double *vfrac = atom->vfrac;
			double *rmass = atom->rmass;
			double *eff_plastic_strain = atom->eff_plastic_strain;
			double **tlsph_stress = atom->smd_stress;
			double *e = atom->e;
			int *type = atom->type;
			int i, itype;
			int nlocal = atom->nlocal;
			Matrix3d D, Ddev, W, V, sigma_diag;
			Matrix3d eye, stressRate, StressRateDevJaumann;
			Matrix3d sigmaInitial_dev, d_dev, sigmaFinal_dev, stressRateDev, oldStressDeviator, newStressDeviator;
			double plastic_strain_increment, yieldStress;
			double dt = update->dt;
			double vol, newPressure;
			double G_eff = 0.0; // effective shear modulus
			double K_eff; // effective bulk modulus
			double M, p_wave_speed;
			double rho, effectiveViscosity;
			Matrix3d deltaStressDev;

			dtCFL = 1.0e22;
			eye.setIdentity();

			for (i = 0; i < nlocal; i++) {
			itype = type[i];
			if (setflag[itype][itype] == 1) {
			newStressDeviator.setZero();
			newPressure = 0.0;
			stressTensor[i].setZero();
			vol = vfrac[i];
			rho = rmass[i] / vfrac[i];
			effectiveViscosity = 0.0;
			K_eff = 0.0;
			G_eff = 0.0;

			//printf("rho = %f\n", rho);

			switch (eos[itype]) {
			default:
			error->one(FLERR, "unknown EOS.");
			break;
			case NONE:
			c0[i] = 1.0;
			break;
			case EOS_TAIT:
			TaitEOS_density(Lookup[EOS_TAIT_EXPONENT][itype], Lookup[REFERENCE_SOUNDSPEED][itype],
			Lookup[REFERENCE_DENSITY][itype], rho, newPressure, c0[i]);
			//printf("new pressure =%f\n", newPressure);

			break;
			case EOS_PERFECT_GAS:
			PerfectGasEOS(Lookup[EOS_PERFECT_GAS_GAMMA][itype], vol, rmass[i], e[i], newPressure, c0[i]);
			break;
			case EOS_LINEAR:
			newPressure = Lookup[BULK_MODULUS][itype] * (rho / Lookup[REFERENCE_DENSITY][itype] - 1.0);
			//printf("p=%f, rho0=%f, rho=%f\n", newPressure, Lookup[REFERENCE_DENSITY][itype], rho);
			c0[i] = Lookup[REFERENCE_SOUNDSPEED][itype];
			break;
			}

			K_eff = c0[i] * c0[i] * rho; // effective bulk modulus

			/*
			* ***************************** STRENGTH MODELS **********************************************
			*/

			if (strength[itype] != NONE) {
			/*
			* initial stress state: given by the unrotateted Cauchy stress.
			* Assemble Eigen 3d matrix from stored stress state
			*/
			oldStressDeviator(0, 0) = tlsph_stress[i][0];
			oldStressDeviator(0, 1) = tlsph_stress[i][1];
			oldStressDeviator(0, 2) = tlsph_stress[i][2];
			oldStressDeviator(1, 1) = tlsph_stress[i][3];
			oldStressDeviator(1, 2) = tlsph_stress[i][4];
			oldStressDeviator(2, 2) = tlsph_stress[i][5];
			oldStressDeviator(1, 0) = oldStressDeviator(0, 1);
			oldStressDeviator(2, 0) = oldStressDeviator(0, 2);
			oldStressDeviator(2, 1) = oldStressDeviator(1, 2);

			D = 0.5 * (L[i] + L[i].transpose());
			W = 0.5 * (L[i] - L[i].transpose()); // spin tensor:: need this for Jaumann rate
			d_dev = Deviator(D);

			switch (strength[itype]) {
			default:
			error->one(FLERR, "unknown strength model.");
			break;
			case STRENGTH_LINEAR:

			// here in a version with pressure part
			// stressRateDev = Lookup[BULK_MODULUS][itype] * d_iso * eye + 2.0 * Lookup[SHEAR_MODULUS][itype] * d_dev;
			// c0[i] = Lookup[REFERENCE_SOUNDSPEED][itype];
			// newPressure = 0.0;

			// here only stress deviator
			stressRateDev = 2.0 * Lookup[SHEAR_MODULUS][itype] * d_dev;
			//cout << "stress rate deviator is " << endl << stressRateDev << endl;
			break;

			case STRENGTH_LINEAR_PLASTIC:
			yieldStress = Lookup[YIELD_STRENGTH][itype] + Lookup[HARDENING_PARAMETER][itype] * eff_plastic_strain[i];
			LinearPlasticStrength(Lookup[SHEAR_MODULUS][itype], yieldStress, oldStressDeviator, d_dev, dt,
			newStressDeviator, stressRateDev, plastic_strain_increment);
			eff_plastic_strain[i] += plastic_strain_increment;

			break;
			}

			//double m = effective_longitudinal_modulus(itype, dt, d_iso, p_rate, d_dev, stressRate_dev, damage);

			StressRateDevJaumann = stressRateDev - W * oldStressDeviator + oldStressDeviator * W;
			newStressDeviator = oldStressDeviator + dt * StressRateDevJaumann;

			tlsph_stress[i][0] = newStressDeviator(0, 0);
			tlsph_stress[i][1] = newStressDeviator(0, 1);
			tlsph_stress[i][2] = newStressDeviator(0, 2);
			tlsph_stress[i][3] = newStressDeviator(1, 1);
			tlsph_stress[i][4] = newStressDeviator(1, 2);
			tlsph_stress[i][5] = newStressDeviator(2, 2);

			// estimate effective shear modulus for time step stability
			deltaStressDev = oldStressDeviator - newStressDeviator;
			G_eff = effective_shear_modulus(d_dev, deltaStressDev, dt, itype);

			} // end if (strength[itype] != NONE)

			if (viscosity[itype] != NONE) {
			D = 0.5 * (L[i] + L[i].transpose());
			d_dev = Deviator(D);

			switch (viscosity[itype]) {
			default:
			error->one(FLERR, "unknown viscosity model.");
			break;
			case VISCOSITY_NEWTON:
			effectiveViscosity = Lookup[VISCOSITY_MU][itype];
			// double shear_rate = 2.0
			// * sqrt(d_dev(0, 1) * d_dev(0, 1) + d_dev(0, 2) * d_dev(0, 2) + d_dev(1, 2) * d_dev(1, 2)); // 3d
			//cout << "shear rate: " << shear_rate << endl;
			//effectiveViscosity = PA6_270C(shear_rate);
			//if (effectiveViscosity > 178.062e-6) {
			// printf("effective visc is %f\n", effectiveViscosity);
			//}
			newStressDeviator = 2.0 * effectiveViscosity * d_dev; // newton original
			//cout << "this is Ddev " << endl << d_dev << endl << endl;
			break;
			}
			} // end if (viscosity[itype] != NONE)

			/*
			* assemble stress Tensor from pressure and deviatoric parts
			*/

			stressTensor[i] = -newPressure * eye + newStressDeviator;

			/*
			* stable timestep based on speed-of-sound
			*/

			M = K_eff + 4.0 * G_eff / 3.0;
			p_wave_speed = sqrt(M / rho);
			effm[i] = G_eff;
			dtCFL = MIN(2 * radius[i] / p_wave_speed, dtCFL);

			/*
			* stable timestep based on viscosity
			*/
			if (viscosity[itype] != NONE) {
			dtCFL = MIN(4 * radius[i] * radius[i] * rho / effectiveViscosity, dtCFL);
			}

			/*
			* kernel gradient correction
			*/
			if (gradient_correction_flag) {
			stressTensor[i] *= K[i];
			}
			}
			// end if (setflag[itype][itype] == 1)
			} // end loop over nlocal

			//printf("stable timestep = %g\n", 0.1 * hMin * MaxBulkVelocity);
			}

			/* ----------------------------------------------------------------------
			allocate all arrays
			------------------------------------------------------------------------- */

			void PairULSPH::allocate() {

			allocated = 1;
			int n = atom->ntypes;

			memory->create(setflag, n + 1, n + 1, "pair:setflag");

			memory->create(Q1, n + 1, "pair:Q1");
			memory->create(rho0, n + 1, "pair:Q2");
			memory->create(c0_type, n + 1, "pair:c0_type");
			memory->create(eos, n + 1, "pair:eosmodel");
			memory->create(viscosity, n + 1, "pair:viscositymodel");
			memory->create(strength, n + 1, "pair:strengthmodel");

			memory->create(Lookup, MAX_KEY_VALUE, n + 1, "pair:LookupTable");

			memory->create(artificial_pressure, n + 1, n + 1, "pair:artificial_pressure");
			memory->create(artificial_stress, n + 1, n + 1, "pair:artificial_stress");
			memory->create(cutsq, n + 1, n + 1, "pair:cutsq"); // always needs to be allocated, even with granular neighborlist

			/*
			* initialize arrays to default values
			*/

			for (int i = 1; i <= n; i++) {
			for (int j = i; j <= n; j++) {
			artificial_pressure[i][j] = 0.0;
			artificial_stress[i][j] = 0.0;
			setflag[i][j] = 0;
			}
			}

			onerad_dynamic = new double[n + 1];
			onerad_frozen = new double[n + 1];
			maxrad_dynamic = new double[n + 1];
			maxrad_frozen = new double[n + 1];

			}

			/* ----------------------------------------------------------------------
			global settings
			------------------------------------------------------------------------- */

			void PairULSPH::settings(int narg, char **arg) {
			if (narg != 3) {
			printf("narg = %d\n", narg);
			error->all(FLERR, "Illegal number of arguments for pair_style ulsph");
			}

			if (comm->me == 0) {
			printf("\n>>========>>========>>========>>========>>========>>========>>========>>========\n");
			printf("... SMD / ULSPH PROPERTIES\n\n");
			}

			if (strcmp(arg[0], "*DENSITY_SUMMATION") == 0) {
			density_summation = true;
			density_continuity = false;
			if (comm->me == 0)
			printf("... density summation active\n");
			} else if (strcmp(arg[0], "*DENSITY_CONTINUITY") == 0) {
			density_continuity = true;
			density_summation = false;
			if (comm->me == 0)
			printf("... density continuity active\n");
			} else {
			error->all(FLERR,
			"Illegal settings keyword for first keyword of pair style ulsph. Must be either DENSITY_SUMMATION or DENSITY_CONTINUITY");
			}

			if (strcmp(arg[1], "*VELOCITY_GRADIENT") == 0) {
			velocity_gradient = true;
			if (comm->me == 0)
			printf("... computation of velocity gradients active\n");
			} else if (strcmp(arg[1], "*NO_VELOCITY_GRADIENT") == 0) {
			velocity_gradient = false;
			if (comm->me == 0)
			printf("... computation of velocity gradients NOT active\n");
			} else {
			error->all(FLERR,
			"Illegal settings keyword for first keyword of pair style ulsph. Must be either VELOCITY_GRADIENT or NO_VELOCITY_GRADIENT");
			}

			if (strcmp(arg[2], "*GRADIENT_CORRECTION") == 0) {
			gradient_correction_flag = true;
			if (comm->me == 0)
			printf("... first order correction of kernel gradients is active\n");
			} else if (strcmp(arg[2], "*NO_GRADIENT_CORRECTION") == 0) {
			gradient_correction_flag = false;
			if (comm->me == 0)
			printf("... first order correction of kernel gradients is NOT active\n");
			} else {
			error->all(FLERR, "Illegal settings keyword for pair style ulsph");
			}

			// error check
			//if ((gradient_correction_flag == true) && (density_summation)) {
			// error->all(FLERR, "Cannot use DENSITY_SUMMATION in combination with YES_GRADIENT_CORRECTION");
			//}

			if (comm->me == 0)
			printf(">>========>>========>>========>>========>>========>>========>>========>>========\n");

			}

			/* ----------------------------------------------------------------------
			set coeffs for one or more type pairs
			------------------------------------------------------------------------- */

			void PairULSPH::coeff(int narg, char **arg) {
			int ioffset, iarg, iNextKwd, itype, jtype;
			char str[128];
			std::string s, t;

			if (narg < 3) {
			sprintf(str, "number of arguments for pair ulsph is too small!");
			error->all(FLERR, str);
			}
			if (!allocated)
			allocate();

			/*
			* if parameters are give in i,i form, i.e., no a cross interaction, set material parameters
			*/

			if (force->inumeric(FLERR, arg[0]) == force->inumeric(FLERR, arg[1])) {

			itype = force->inumeric(FLERR, arg[0]);
			eos[itype] = viscosity[itype] = strength[itype] = NONE;

			if (comm->me == 0) {
			printf("\n>>========>>========>>========>>========>>========>>========>>========>>========\n");
			printf("...SMD / ULSPH PROPERTIES OF PARTICLE TYPE %d\n\n", itype);
			}

			/*
			* read parameters which are common -- regardless of material / eos model
			*/

			ioffset = 2;
			if (strcmp(arg[ioffset], "*COMMON") != 0) {
			sprintf(str, "common keyword missing!");
			error->all(FLERR, str);
			} else {
			}

			t = string("*");
			iNextKwd = -1;
			for (iarg = ioffset + 1; iarg < narg; iarg++) {
			s = string(arg[iarg]);
			if (s.compare(0, t.length(), t) == 0) {
			iNextKwd = iarg;
			break;
			}
			}

			//printf("keyword following *COMMON is %s\n", arg[iNextKwd]);

			if (iNextKwd < 0) {
			sprintf(str, "no KEYWORD terminates COMMON");
			error->all(FLERR, str);
			}

			if (iNextKwd - ioffset != 5 + 1) {
			sprintf(str, "expected 5 arguments following *COMMON but got %d\n", iNextKwd - ioffset - 1);
			error->all(FLERR, str);
			}

			Lookup[REFERENCE_DENSITY][itype] = force->numeric(FLERR, arg[ioffset + 1]);
			Lookup[REFERENCE_SOUNDSPEED][itype] = force->numeric(FLERR, arg[ioffset + 2]);
			Q1[itype] = force->numeric(FLERR, arg[ioffset + 3]);
			Lookup[HEAT_CAPACITY][itype] = force->numeric(FLERR, arg[ioffset + 4]);
			Lookup[HOURGLASS_CONTROL_AMPLITUDE][itype] = force->numeric(FLERR, arg[ioffset + 5]);

			Lookup[BULK_MODULUS][itype] = Lookup[REFERENCE_SOUNDSPEED][itype] * Lookup[REFERENCE_SOUNDSPEED][itype]
			* Lookup[REFERENCE_DENSITY][itype];

			if (comm->me == 0) {
			printf("material unspecific properties for SMD/ULSPH definition of particle type %d:\n", itype);
			printf(FORMAT1, "reference density", Lookup[REFERENCE_DENSITY][itype]);
			printf(FORMAT1, "reference speed of sound", Lookup[REFERENCE_SOUNDSPEED][itype]);
			printf(FORMAT1, "linear viscosity coefficient", Q1[itype]);
			printf(FORMAT1, "heat capacity [energy / (mass * temperature)]", Lookup[HEAT_CAPACITY][itype]);
			printf(FORMAT1, "bulk modulus", Lookup[BULK_MODULUS][itype]);
			printf(FORMAT1, "hourglass control amplitude", Lookup[HOURGLASS_CONTROL_AMPLITUDE][itype]);
			}

			/*
			* read following material cards
			*/

			// if (comm->me == 0) {
			// printf("next kwd is %s\n", arg[iNextKwd]);
			// }
			while (true) {
			if (strcmp(arg[iNextKwd], "*END") == 0) {
			// if (comm->me == 0) {
			// sprintf(str, "found *END");
			// error->message(FLERR, str);
			// }
			break;
			}

			ioffset = iNextKwd;
			if (strcmp(arg[ioffset], "*EOS_TAIT") == 0) {

			/*
			* Tait EOS
			*/

			eos[itype] = EOS_TAIT;
			//printf("reading *EOS_TAIT\n");

			t = string("*");
			iNextKwd = -1;
			for (iarg = ioffset + 1; iarg < narg; iarg++) {
			s = string(arg[iarg]);
			if (s.compare(0, t.length(), t) == 0) {
			iNextKwd = iarg;
			break;
			}
			}

			if (iNextKwd < 0) {
			sprintf(str, "no KEYWORD terminates EOS_TAIT");
			error->all(FLERR, str);
			}

			if (iNextKwd - ioffset != 1 + 1) {
			sprintf(str, "expected 1 arguments following *EOS_TAIT but got %d\n", iNextKwd - ioffset - 1);
			error->all(FLERR, str);
			}

			Lookup[EOS_TAIT_EXPONENT][itype] = force->numeric(FLERR, arg[ioffset + 1]);

			if (comm->me == 0) {
			printf(FORMAT2, "Tait EOS");
			printf(FORMAT1, "Exponent", Lookup[EOS_TAIT_EXPONENT][itype]);
			}
			} // end Tait EOS

			else if (strcmp(arg[ioffset], "*EOS_PERFECT_GAS") == 0) {

			/*
			* Perfect Gas EOS
			*/

			eos[itype] = EOS_PERFECT_GAS;
			//printf("reading *EOS_PERFECT_GAS\n");

			t = string("*");
			iNextKwd = -1;
			for (iarg = ioffset + 1; iarg < narg; iarg++) {
			s = string(arg[iarg]);
			if (s.compare(0, t.length(), t) == 0) {
			iNextKwd = iarg;
			break;
			}
			}

			if (iNextKwd < 0) {
			sprintf(str, "no KEYWORD terminates EOS_PERFECT_GAS");
			error->all(FLERR, str);
			}

			if (iNextKwd - ioffset != 1 + 1) {
			sprintf(str, "expected 1 arguments following *EOS_PERFECT_GAS but got %d\n", iNextKwd - ioffset - 1);
			error->all(FLERR, str);
			}

			Lookup[EOS_PERFECT_GAS_GAMMA][itype] = force->numeric(FLERR, arg[ioffset + 1]);

			if (comm->me == 0) {
			printf(FORMAT2, "Perfect Gas EOS");
			printf(FORMAT1, "Heat Capacity Ratio Gamma", Lookup[EOS_PERFECT_GAS_GAMMA][itype]);
			}
			} // end Perfect Gas EOS
			else if (strcmp(arg[ioffset], "*EOS_LINEAR") == 0) {

			/*
			* Linear EOS
			*/

			eos[itype] = EOS_LINEAR;

			t = string("*");
			iNextKwd = -1;
			for (iarg = ioffset + 1; iarg < narg; iarg++) {
			s = string(arg[iarg]);
			if (s.compare(0, t.length(), t) == 0) {
			iNextKwd = iarg;
			break;
			}
			}

			if (iNextKwd < 0) {
			sprintf(str, "no KEYWORD terminates EOS_LINEAR");
			error->all(FLERR, str);
			}

			if (iNextKwd - ioffset != 0 + 1) {
			sprintf(str, "expected 0 arguments following *EOS_LINEAR but got %d\n", iNextKwd - ioffset - 1);
			error->all(FLERR, str);
			}

			if (comm->me == 0) {
			printf(FORMAT2, "Linear EOS");
			printf(FORMAT1, "Bulk modulus", Lookup[BULK_MODULUS][itype]);
			}
			} // end Linear EOS
			else if (strcmp(arg[ioffset], "*STRENGTH_LINEAR_PLASTIC") == 0) {

			if (velocity_gradient != true) {
			error->all(FLERR, "A strength model was requested but *VELOCITY_GRADIENT is not set");
			}

			/*
			* linear elastic / ideal plastic material model with strength
			*/

			strength[itype] = STRENGTH_LINEAR_PLASTIC;
			velocity_gradient_required = true;
			//printf("reading *LINEAR_PLASTIC\n");

			t = string("*");
			iNextKwd = -1;
			for (iarg = ioffset + 1; iarg < narg; iarg++) {
			s = string(arg[iarg]);
			if (s.compare(0, t.length(), t) == 0) {
			iNextKwd = iarg;
			break;
			}
			}

			if (iNextKwd < 0) {
			sprintf(str, "no KEYWORD terminates STRENGTH_LINEAR_PLASTIC");
			error->all(FLERR, str);
			}

			if (iNextKwd - ioffset != 3 + 1) {
			sprintf(str, "expected 3 arguments following *STRENGTH_LINEAR_PLASTIC but got %d\n", iNextKwd - ioffset - 1);
			error->all(FLERR, str);
			}

			Lookup[SHEAR_MODULUS][itype] = force->numeric(FLERR, arg[ioffset + 1]);
			Lookup[YIELD_STRENGTH][itype] = force->numeric(FLERR, arg[ioffset + 2]);
			Lookup[HARDENING_PARAMETER][itype] = force->numeric(FLERR, arg[ioffset + 3]);

			if (comm->me == 0) {
			printf(FORMAT2, "linear elastic / ideal plastic material mode");
			printf(FORMAT1, "yield_strength", Lookup[YIELD_STRENGTH][itype]);
			printf(FORMAT1, "constant hardening parameter", Lookup[HARDENING_PARAMETER][itype]);
			printf(FORMAT1, "shear modulus", Lookup[SHEAR_MODULUS][itype]);
			}
			} // end *STRENGTH_LINEAR_PLASTIC
			else if (strcmp(arg[ioffset], "*STRENGTH_LINEAR") == 0) {

			if (velocity_gradient != true) {
			error->all(FLERR, "A strength model was requested but *VELOCITY_GRADIENT is not set");
			}

			/*
			* linear elastic / ideal plastic material model with strength
			*/

			strength[itype] = STRENGTH_LINEAR;
			t = string("*");
			iNextKwd = -1;
			for (iarg = ioffset + 1; iarg < narg; iarg++) {
			s = string(arg[iarg]);
			if (s.compare(0, t.length(), t) == 0) {
			iNextKwd = iarg;
			break;
			}
			}

			if (iNextKwd < 0) {
			sprintf(str, "no KEYWORD terminates STRENGTH_LINEAR");
			error->all(FLERR, str);
			}

			if (iNextKwd - ioffset != 1 + 1) {
			sprintf(str, "expected 1 arguments following *STRENGTH_LINEAR but got %d\n", iNextKwd - ioffset - 1);
			error->all(FLERR, str);
			}

			Lookup[SHEAR_MODULUS][itype] = force->numeric(FLERR, arg[ioffset + 1]);

			if (comm->me == 0) {
			printf(FORMAT2, "linear elastic strength model");
			printf(FORMAT1, "shear modulus", Lookup[SHEAR_MODULUS][itype]);
			}
			} // end *STRENGTH_LINEAR
			else if (strcmp(arg[ioffset], "*VISCOSITY_NEWTON") == 0) {

			if (velocity_gradient != true) {
			error->all(FLERR, "A viscosity model was requested but *VELOCITY_GRADIENT is not set");
			}

			/*
			* linear elastic / ideal plastic material model with strength
			*/

			viscosity[itype] = VISCOSITY_NEWTON;
			t = string("*");
			iNextKwd = -1;
			for (iarg = ioffset + 1; iarg < narg; iarg++) {
			s = string(arg[iarg]);
			if (s.compare(0, t.length(), t) == 0) {
			iNextKwd = iarg;
			break;
			}
			}

			if (iNextKwd < 0) {
			sprintf(str, "no KEYWORD terminates VISCOSITY_NEWTON");
			error->all(FLERR, str);
			}

			if (iNextKwd - ioffset != 1 + 1) {
			sprintf(str, "expected 1 arguments following *VISCOSITY_NEWTON but got %d\n", iNextKwd - ioffset - 1);
			error->all(FLERR, str);
			}

			Lookup[VISCOSITY_MU][itype] = force->numeric(FLERR, arg[ioffset + 1]);

			if (comm->me == 0) {
			printf(FORMAT2, "Newton viscosity model");
			printf(FORMAT1, "viscosity mu", Lookup[VISCOSITY_MU][itype]);
			}
			} // end *STRENGTH_VISCOSITY_NEWTON

			else if (strcmp(arg[ioffset], "*ARTIFICIAL_PRESSURE") == 0) {

			/*
			* use Monaghan's artificial pressure to prevent particle clumping
			*/

			t = string("*");
			iNextKwd = -1;
			for (iarg = ioffset + 1; iarg < narg; iarg++) {
			s = string(arg[iarg]);
			if (s.compare(0, t.length(), t) == 0) {
			iNextKwd = iarg;
			break;
			}
			}

			if (iNextKwd < 0) {
			sprintf(str, "no KEYWORD terminates ARTIFICIAL_PRESSURE");
			error->all(FLERR, str);
			}

			if (iNextKwd - ioffset != 1 + 1) {
			sprintf(str, "expected 1 arguments following *ARTIFICIAL_PRESSURE but got %d\n", iNextKwd - ioffset - 1);
			error->all(FLERR, str);
			}

			artificial_pressure[itype][itype] = force->numeric(FLERR, arg[ioffset + 1]);

			if (comm->me == 0) {
			printf(FORMAT2, "Artificial Pressure is enabled.");
			printf(FORMAT1, "Artificial Pressure amplitude", artificial_pressure[itype][itype]);
			}
			} // end *ARTIFICIAL_PRESSURE

			else if (strcmp(arg[ioffset], "*ARTIFICIAL_STRESS") == 0) {

			/*
			* use Monaghan's artificial stress to prevent particle clumping
			*/

			t = string("*");
			iNextKwd = -1;
			for (iarg = ioffset + 1; iarg < narg; iarg++) {
			s = string(arg[iarg]);
			if (s.compare(0, t.length(), t) == 0) {
			iNextKwd = iarg;
			break;
			}
			}

			if (iNextKwd < 0) {
			sprintf(str, "no KEYWORD terminates ARTIFICIAL_STRESS");
			error->all(FLERR, str);
			}

			if (iNextKwd - ioffset != 1 + 1) {
			sprintf(str, "expected 1 arguments following *ARTIFICIAL_STRESS but got %d\n", iNextKwd - ioffset - 1);
			error->all(FLERR, str);
			}

			artificial_stress[itype][itype] = force->numeric(FLERR, arg[ioffset + 1]);

			if (comm->me == 0) {
			printf(FORMAT2, "Artificial Stress is enabled.");
			printf(FORMAT1, "Artificial Stress amplitude", artificial_stress[itype][itype]);
			}
			} // end *ARTIFICIAL_STRESS

			else {
			sprintf(str, "unknown *KEYWORD: %s", arg[ioffset]);
			error->all(FLERR, str);
			}

			}

			/*
			* copy data which is looked up in inner pairwise loops from slow maps to fast arrays
			*/

			rho0[itype] = Lookup[REFERENCE_DENSITY][itype];
			c0_type[itype] = Lookup[REFERENCE_SOUNDSPEED][itype];
			setflag[itype][itype] = 1;

			/*
			* error checks
			*/

			if ((viscosity[itype] != NONE) && (strength[itype] != NONE)) {
			sprintf(str, "cannot have both a strength and viscosity model for particle type %d", itype);
			error->all(FLERR, str);
			}

			if (eos[itype] == NONE) {
			sprintf(str, "must specify an EOS for particle type %d", itype);
			error->all(FLERR, str);
			}

			} else {
			/*
			* we are reading a cross-interaction line for particle types i, j
			*/

			itype = force->inumeric(FLERR, arg[0]);
			jtype = force->inumeric(FLERR, arg[1]);

			if (strcmp(arg[2], "*CROSS") != 0) {
			sprintf(str, "ulsph cross interaction between particle type %d and %d requested, however, *CROSS keyword is missing",
			itype, jtype);
			error->all(FLERR, str);
			}

			if (setflag[itype][itype] != 1) {
			sprintf(str,
			"ulsph cross interaction between particle type %d and %d requested, however, properties of type %d have not yet been specified",
			itype, jtype, itype);
			error->all(FLERR, str);
			}

			if (setflag[jtype][jtype] != 1) {
			sprintf(str,
			"ulsph cross interaction between particle type %d and %d requested, however, properties of type %d have not yet been specified",
			itype, jtype, jtype);
			error->all(FLERR, str);
			}

			setflag[itype][jtype] = 1;
			setflag[jtype][itype] = 1;

			if ((artificial_pressure[itype][itype] > 0.0) && (artificial_pressure[jtype][jtype] > 0.0)) {
			artificial_pressure[itype][jtype] = 0.5 * (artificial_pressure[itype][itype] + artificial_pressure[jtype][jtype]);
			artificial_pressure[jtype][itype] = artificial_pressure[itype][jtype];
			} else {
			artificial_pressure[itype][jtype] = artificial_pressure[jtype][itype] = 0.0;
			}

			if ((artificial_stress[itype][itype] > 0.0) && (artificial_stress[jtype][jtype] > 0.0)) {
			artificial_stress[itype][jtype] = 0.5 * (artificial_stress[itype][itype] + artificial_stress[jtype][jtype]);
			artificial_stress[jtype][itype] = artificial_stress[itype][jtype];
			} else {
			artificial_stress[itype][jtype] = artificial_stress[jtype][itype] = 0.0;
			}

			if (comm->me == 0) {
			printf(">>========>>========>>========>>========>>========>>========>>========>>========\n");
			}

			}
			}

			/* ----------------------------------------------------------------------
			init for one type pair i,j and corresponding j,i
			------------------------------------------------------------------------- */

			double PairULSPH::init_one(int i, int j) {

			if (!allocated)
			allocate();

			if (setflag[i][j] == 0)
			error->all(FLERR, "All pair coeffs are not set");

			// cutoff = sum of max I,J radii for
			// dynamic/dynamic & dynamic/frozen interactions, but not frozen/frozen

			double cutoff = maxrad_dynamic[i] + maxrad_dynamic[j];
			cutoff = MAX(cutoff, maxrad_frozen[i] + maxrad_dynamic[j]);
			cutoff = MAX(cutoff, maxrad_dynamic[i] + maxrad_frozen[j]);
			//printf("cutoff for pair sph/fluid = %f\n", cutoff);
			return cutoff;
			}

			/* ----------------------------------------------------------------------
			init specific to this pair style
			------------------------------------------------------------------------- */

			void PairULSPH::init_style() {
			int i;

			//printf(" in init style\n");
			// request a granular neighbor list
			int irequest = neighbor->request(this);
			neighbor->requests[irequest]->half = 0;
			neighbor->requests[irequest]->gran = 1;

			// set maxrad_dynamic and maxrad_frozen for each type
			// include future Fix pour particles as dynamic

			for (i = 1; i <= atom->ntypes; i++)
			onerad_dynamic[i] = onerad_frozen[i] = 0.0;

			double *radius = atom->radius;
			int *type = atom->type;
			int nlocal = atom->nlocal;

			for (i = 0; i < nlocal; i++)
			onerad_dynamic[type[i]] = MAX(onerad_dynamic[type[i]], radius[i]);

			MPI_Allreduce(&onerad_dynamic[1], &maxrad_dynamic[1], atom->ntypes, MPI_DOUBLE, MPI_MAX, world);
			MPI_Allreduce(&onerad_frozen[1], &maxrad_frozen[1], atom->ntypes, MPI_DOUBLE, MPI_MAX, world);

			}

			/* ----------------------------------------------------------------------
			neighbor callback to inform pair style of neighbor list to use
			optional granular history list
			------------------------------------------------------------------------- */

			void PairULSPH::init_list(int id, NeighList *ptr) {
			if (id == 0)
			list = ptr;
			}

			/* ----------------------------------------------------------------------
			memory usage of local atom-based arrays
			------------------------------------------------------------------------- */

			double PairULSPH::memory_usage() {

			//printf("in memory usage\n");

			return 11 * nmax * sizeof(double);

			}

			/* ---------------------------------------------------------------------- */

			int PairULSPH::pack_forward_comm(int n, int list, double buf, int pbc_flag, int *pbc) {
			double *vfrac = atom->vfrac;
			double *eff_plastic_strain = atom->eff_plastic_strain;
			int i, j, m;

			//printf("packing comm\n");
			m = 0;
			for (i = 0; i < n; i++) {
			j = list[i];
			buf[m++] = vfrac[j];
			buf[m++] = c0[j]; //2

			buf[m++] = stressTensor[j](0, 0); // pack symmetric stress tensor
			buf[m++] = stressTensor[j](1, 1);
			buf[m++] = stressTensor[j](2, 2);
			buf[m++] = stressTensor[j](0, 1);
			buf[m++] = stressTensor[j](0, 2);
			buf[m++] = stressTensor[j](1, 2); // 2 + 6 = 8

			buf[m++] = F[j](0, 0); // F is not symmetric
			buf[m++] = F[j](0, 1);
			buf[m++] = F[j](0, 2);
			buf[m++] = F[j](1, 0);
			buf[m++] = F[j](1, 1);
			buf[m++] = F[j](1, 2);
			buf[m++] = F[j](2, 0);
			buf[m++] = F[j](2, 1);
			buf[m++] = F[j](2, 2); // 8 + 9 = 17

			buf[m++] = eff_plastic_strain[j]; // 18
			}
			return m;
			}

			/* ---------------------------------------------------------------------- */

			void PairULSPH::unpack_forward_comm(int n, int first, double *buf) {
			double *vfrac = atom->vfrac;
			double *eff_plastic_strain = atom->eff_plastic_strain;
			int i, m, last;

			m = 0;
			last = first + n;
			for (i = first; i < last; i++) {
			vfrac[i] = buf[m++];
			c0[i] = buf[m++]; // 2

			stressTensor[i](0, 0) = buf[m++];
			stressTensor[i](1, 1) = buf[m++];
			stressTensor[i](2, 2) = buf[m++];
			stressTensor[i](0, 1) = buf[m++];
			stressTensor[i](0, 2) = buf[m++];
			stressTensor[i](1, 2) = buf[m++]; // 2 + 6 = 8
			stressTensor[i](1, 0) = stressTensor[i](0, 1);
			stressTensor[i](2, 0) = stressTensor[i](0, 2);
			stressTensor[i](2, 1) = stressTensor[i](1, 2);

			F[i](0, 0) = buf[m++];
			F[i](0, 1) = buf[m++];
			F[i](0, 2) = buf[m++];
			F[i](1, 0) = buf[m++];
			F[i](1, 1) = buf[m++];
			F[i](1, 2) = buf[m++];
			F[i](2, 0) = buf[m++];
			F[i](2, 1) = buf[m++];
			F[i](2, 2) = buf[m++]; // 8 + 9 = 17

			eff_plastic_strain[i] = buf[m++]; // 18
			}
			}

			/*
			* EXTRACT
			*/

			void PairULSPH::extract(const char str, int &i) {
			//printf("in extract\n");
			if (strcmp(str, "smd/ulsph/smoothVel_ptr") == 0) {
			return (void *) smoothVel;
			} else if (strcmp(str, "smd/ulsph/stressTensor_ptr") == 0) {
			return (void *) stressTensor;
			} else if (strcmp(str, "smd/ulsph/velocityGradient_ptr") == 0) {
			return (void *) L;
			} else if (strcmp(str, "smd/ulsph/numNeighs_ptr") == 0) {
			return (void *) numNeighs;
			} else if (strcmp(str, "smd/ulsph/dtCFL_ptr") == 0) {
			//printf("dtcfl = %f\n", dtCFL);
			return (void *) &dtCFL;
			} else if (strcmp(str, "smd/ulsph/updateFlag_ptr") == 0) {
			return (void *) &updateFlag;
			} else if (strcmp(str, "smd/ulsph/effective_modulus_ptr") == 0) {
			return (void *) effm;
			} else if (strcmp(str, "smd/ulsph/shape_matrix_ptr") == 0) {
			return (void *) K;
			}

			return NULL;
			}

			/* ----------------------------------------------------------------------
			compute effective shear modulus by dividing rate of deviatoric stress with rate of shear deformation
			------------------------------------------------------------------------- */

			double PairULSPH::effective_shear_modulus(const Matrix3d d_dev, const Matrix3d deltaStressDev, const double dt, const int itype) {
			double G_eff; // effective shear modulus, see Pronto 2d eq. 3.4.7
			double deltaStressDevSum, shearRateSq, strain_increment;

			if (domain->dimension == 3) {
			deltaStressDevSum = deltaStressDev(0, 1) * deltaStressDev(0, 1) + deltaStressDev(0, 2) * deltaStressDev(0, 2)
			+ deltaStressDev(1, 2) * deltaStressDev(1, 2);
			shearRateSq = d_dev(0, 1) * d_dev(0, 1) + d_dev(0, 2) * d_dev(0, 2) + d_dev(1, 2) * d_dev(1, 2);
			} else {
			deltaStressDevSum = deltaStressDev(0, 1) * deltaStressDev(0, 1);
			shearRateSq = d_dev(0, 1) * d_dev(0, 1);
			}

			strain_increment = dt * dt * shearRateSq;

			if (strain_increment > 1.0e-12) {
			G_eff = 0.5 * sqrt(deltaStressDevSum / strain_increment);
			} else {
			if (strength[itype] != NONE) {
			G_eff = Lookup[SHEAR_MODULUS][itype];
			} else {
			G_eff = 0.0;
			}
			}

			return G_eff;

			}

			/* ----------------------------------------------------------------------
			hourglass force for updated Lagrangian SPH
			input: particles indices i, j, particle types ityep, jtype
			------------------------------------------------------------------------- */

			Vector3d PairULSPH::ComputeHourglassForce(const int i, const int itype, const int j, const int jtype, const Vector3d dv,
			const Vector3d xij, const Vector3d g, const double c_ij, const double mu_ij, const double rho_ij) {

			double *rmass = atom->rmass;
			Vector3d dv_est, f_hg;
			double visc_magnitude;

			dv_est = -0.5 * (F[i] + F[j]) * xij;
			double hurz = dv_est.dot(dv) / (dv_est.norm() * dv.norm() + 1.0e-16);
			if (hurz < 0.0) {

			visc_magnitude = 0.5 * (Q1[itype] + Q1[jtype]) * c_ij * mu_ij / rho_ij;
			f_hg = -rmass[i] * rmass[j] * visc_magnitude * g;
			// printf(" f_hg = %f %f %f\n", f_hg(0), f_hg(1), f_hg(2));
			// printf("\nnegative\n");
			// printf(" dv_est = %f %f %f\n", dv_est(0), dv_est(1), dv_est(2));
			// printf(" dv = %f %f %f\n", dv(0), dv(1), dv(2));
			} else {
			f_hg.setZero();
			}

			return f_hg;

			}

lammps/src/USER-SMD/pair_smd_ulsph.cpp354e20f43157efficient_neuronet

pair_smd_ulsph.cppView Options

lammps/src/USER-SMD/pair_smd_ulsph.cpp
354e20f43157efficient_neuronet

pair_smd_ulsph.cpp
View Options