SEUIF97

This is the Rust implementation of the high-speed IAPWS-IF97 package SEUIF97 with C and Python bindings. It is designed for computation-intensive tasks, such as simulating non-stationary processes, on-line process monitoring, and optimization.

Through the high-speed package, IAPWS-IF97 calculations achieve a 5x to 20x speedup compared to direct implementations using the Rust standard library's powi() within loops for the basic equations of Regions 1, 2, and 3.

SEUIF97 also significantly outperforms various approximate equations and algorithms typically used for fast water and steam property calculations.

Key Acceleration Methods

• Loop Tiling Method: Unleashes the full power of compiler optimizations, surpassing the performance of the single loop.

• Recurrence Method for Multi-Polynomial Evaluation: By leveraging the relationship between polynomials and their derivatives, only a single polynomial needs to be computed directly. The remaining values are derived via multiplication or division by the base. This approach eliminates redundant calculations and significantly improves performance.

Please refer to The acceleration methods for more details on the algorithm

Input Pairs and Properties

This package supports 12 distinct input state pairs for calculating 36 thermodynamic, transport, and derived properties (see Properties).

Input Pairs:

(p,t) (p,h) (p,s) (p,v)
(t,h) (t,s) (t,v)
(p,x) (t,x) (h,x) (s,x)
(h,s)

Usage

Install the crate

cargo add seuif97

The type of functions are provided in the package:

struct o_id_region_args {
   o_id: i32,
   region: i32,
}

fn<R>(f64,f64,R) -> f64
where
    R: Into<o_id_region_args>,

• the first,second input parameters(f64) : the input property pairs
• the third and fourth input parametes:
  • the third : the property ID of the calculated property - o_id
  • the fourth option parameter: the region of IAPWS-IF97
• the return(f64): the calculated property value of o_id

pt<R>(p:f64,t:f64,o_id_region:R)->f64
ph<R>(p:f64,h:f64,o_id_region:R)->f64
ps<R>(p:f64,s:f64,o_id_region:R)->f64
pv<R>(p:f64,v:f64,o_id_region:R)->f64

th<R>(t:f64,h:f64,o_id_region:R)->f64
ts<R>(t:f64,s:f64,o_id_region:R)->f64
tv<R>(t:f64,v:f64,o_id_region:R)->f64

hs<R>(h:f64,s:f64,o_id_region:R)->f64

px(p:f64,x:f64,o_id:i32)->f64
tx(p:f64,x:f64,o_id:i32)->f64
hx(h:f64,x:f64,o_id:i32)->f64
sx(s:f64,x:f64,o_id:i32)->f64

Example

use seuif97::*;
fn main() {
    
    let p:f64 = 3.0;
    let t:f64= 300.0-273.15;
   
    let h=pt(p,t,OH);
    let s=pt(p,t,OS);
    // set the region
    let v=pt(p,t,(OV,1));
    println!("p={p:.6} t={t:.6} h={t:.6} s={s:.6} v={v:.6}");   
}

The C binding

Building the dynamic link library

• cdecl

cargo build -r --features cdecl

• stdcall: Windows API functions(64bit)

cargo build -r --features stdcall

• stdcall: Windows API functions(32bit)

cargo build -r  --target=i686-pc-windows-msvc --features stdcall

The convenient compiled dynamic link libraries are provided in the ./dynamic_lib/

• seuif97.dll: Windows64 and Windows32

• libseuif97.so: Linux64

The functions in C

double pt(double p,double t,short o_id);
double ph(double p,double h,short o_id);
double ps(double p,double s,short o_id);
double pv(double p,double v,short o_id);

double tv(double t,double v,short o_id);
double th(double t,double h,short o_id);
double ts(double t,double s,short o_id);

double hs(double h,double s,short o_id);

double px(double p,double x,short o_id);
double tx(double t,double x,short o_id);
double hx(double h,double x,short o_id);
double sx(double s,double x,short o_id);

Example

• ./demo_using_lib/: C, Python, C#, Excel VBA, Java, Fortran, Golang

#include <stdlib.h>
#include <stdio.h>
#include <string.h>

#define OH 4
#define OS 5

extern double pt(double p,double t,short o_id);

int main(void)
{
    double p = 16.0;
    double t = 530.0;
    double h = pt(p, t, OH);
    double s = pt(p, t, OS);
    printf("p,t %f,%f h= %f s= %f\n", p, t, h, s);
    return EXIT_SUCCESS;
}

The Example of Rankine Cycle Analysis

• The Rankine Cycle Steady-state Simulator in Python，C++ and Modelica

The Python binding

Install from pypi

pip install seuif97

Build the release version locally and install it to the local Python environment

python ./setup.py install

Example

from seuif97 import *

OH=4

p=16.0
t=535.1
# ??(in1,in2,o_id)
h=pt(p,t,OH)
# ??2?(in1,in2)
s=pt2s(p,t)
print(f"p={p}, t={t} h={h:.3f} s={s:.3f}")

Examples

• T-S Diagram

• H-S Diagram

• H-S Diagram of Steam Turbine Expansion

• The Hybrid Steady-state Simulator of Rankine Cycle in Python

Properties

Poperty	Unit	Symbol	o_id	o_id(i32)
Pressure	MPa	p	OP	0
Temperature	°C	t	OT	1
Density	kg/m³	ρ	OD	2
Specific Volume	m³/kg	v	OV	3
Specific enthalpy	kJ/kg	h	OH	4
Specific entropy	kJ/(kg·K)	s	OS	5
Specific exergy	kJ/kg	e	OE	6
Specific internal energy	kJ/kg	u	OU	7
Specific isobaric heat capacity	kJ/(kg·K)	cp	OCP	8
Specific isochoric heat capacity	kJ/(kg·K)	cv	OCV	9
Speed of sound	m/s	w	OW	10
Isentropic exponent		k	OKS	11
Specific Helmholtz free energy	kJ/kg	f	OF	12
Specific Gibbs free energy	kJ/kg	g	OG	13
Compressibility factor		z	OZ	14
Steam quality		x	OX	15
Region		r	OR	16
Isobari cubic expansion coefficient	1/K	ɑv	OEC	17
Isothermal compressibility	1/MPa	kT	OKT	18
Partial derivative (∂V/∂T)p	m³/(kg·K)	(∂V/∂T)p	ODVDT	19
Partial derivative (∂V/∂p)T	m³/(kg·MPa)	(∂v/∂p)t	ODVDP	20
Partial derivative (∂P/∂T)v	MPa/K	(∂p/∂t)v	ODPDT	21
Isothermal throttling coefficient	kJ/(kg·MPa)	δt	OIJTC	22
Joule-Thomson coefficient	K/MPa	μ	OJTC	23
Dynamic viscosity	Pa·s	η	ODV	24
Kinematic viscosity	m²/s	ν	OKV	25
Thermal conductivity	W/(m.K)	λ	OTC	26
Thermal diffusivity	m²/s	a	OTD	27
Prandtl number		Pr	OPR	28
Surface tension	N/m	σ	OST	29
Static Dielectric Constant		ε	OSDC	30
Isochoric pressure coefficient	1/K	β	OPC	31
Isothermal stress coefficient	kg/m³	βp	OBETAP	32
Fugacity coefficient		fi	OFI	33
Fugacity	MPa	f*	OFU	34
Relative pressure coefficient	1/K	αp	OAFLAP	35

Cite as

• Cheng Maohua. (2023). The Rust implementation of the high-speed IAPWS-IF97 package: SEUIF97 (1.0.9). Zenodo. https://doi.org/10.5281/zenodo.8246380