Almaarefa University
The authoritative source for research information and outputs is Elsevier Scopus, with researcher profile verification from the Deanship of Scientific Research.
ArticleOpen access2026SDG 70 citation

Global thermodynamic consistency: A linear-algebraic diagnostic framework

Refka Sai ✉،Kamel Smida

Results in Engineering · vol. 32 · 112122-112122 · 2026

Abstract

Thermodynamic cycle models are routinely assembled from heterogeneous sources – steam-table lookups, manufacturer data sheets, handbook correlations, and hand-computed preliminary estimates – and an error introduced in any one source (a misread enthalpy, a sign error in a heat balance, a wrong coefficient-of-performance value) can propagate silently through every downstream calculation. Engineers currently guard against this by checking governing equations pairwise, but pairwise checking has a structural blind spot: a set of equations can pass every pairwise comparison and still be globally infeasible as a system. We develop a linear-algebraic diagnostic framework that closes this gap. The governing equations of a steady-state thermodynamic cycle – energy balances over turbines, compressors, boilers, condensers, and expansion devices, together with isentropic-process relations and the coefficient-of-performance definition – form a linear system A x = b whose state vector x collects enthalpies, heat transfer rates, and work quantities at each process node. We show that global inconsistency of this system is equivalent to infeasibility of A x = b , and that infeasibility is always witnessed by a vector in the left nullspace of A that is not orthogonal to b – an algebraically certifiable condition we call a nullspace inconsistency certificate , computable from a single thin SVD and fundamentally inaccessible to pairwise checking. Building on this certificate, we prove that pairwise consistency checks are provably insufficient to detect global thermodynamic inconsistency; we establish existence and uniqueness of a minimum-norm repair vector that returns the right-hand-side data to the nearest feasible point on the constraint manifold; we derive equation-localization weights that rank which governing equation is the primary source of a fault; and we construct a differentiable repair operator whose fixed points correspond to globally self-consistent cycle models. The complete detection-localization-repair pipeline runs in O ( kn 2 ) time via a thin SVD – negligible overhead for cycle models of practical size – and requires no information beyond the equations and data an engineer already has on hand. We demonstrate the framework on the Rankine steam-power cycle and the vapor-compression refrigeration cycle, where it correctly detects, localizes, and corrects hidden inconsistencies – including a 50 kJ kg − 1 enthalpy error in Rankine cycle steam tables and a simultaneous condenser-balance fault with an erroneous COP value in the refrigeration cycle – that pass all pairwise equation checks yet render the global model infeasible. We further discuss how the certificate behaves under measurement and property-estimation uncertainty, and compare the approach against residual-based checking, data-reconciliation, and optimization-based fault detection. The framework requires only the linear (or linearized) governing equations and standard property data already used in routine cycle calculations, making it directly deployable as an auditing layer in existing engineering workflows.

Details

Journal
Results in Engineering
Publisher
Elsevier BV
Volume
32
Issue
—
Pages
112122-112122
ISSN
2590-1230
Year
2026
Language
en
License
cc-by-nc-nd
Open access
gold
Citations
0
Source
scopus

Topics

Computer scienceWork (physics)Context (archaeology)MathematicsComponent (thermodynamics)Stability (learning theory)

Thermodynamic and Exergetic Analyses of Power and Cooling Systems (Engineering) · Fault Detection and Control Systems (Engineering) · Turbomachinery Performance and Optimization (Engineering)

Cite this

APA
Sai, R., Smida, K. (2026). Global thermodynamic consistency: A linear-algebraic diagnostic framework. Results in Engineering, 32, 112122-112122. https://doi.org/10.1016/j.rineng.2026.112122
Harvard
Sai, R., Smida, K. 2026, 'Global thermodynamic consistency: A linear-algebraic diagnostic framework', Results in Engineering, vol. 32, 112122-112122. Available at: https://doi.org/10.1016/j.rineng.2026.112122
Vancouver
Sai R, Smida K. Global thermodynamic consistency: A linear-algebraic diagnostic framework. Results in Engineering. 2026;32, 112122-112122.
BibTeX
@article{sai2026114,
  title = {Global thermodynamic consistency: A linear-algebraic diagnostic framework},
  author = {Refka Sai and Kamel Smida},
  year = {2026},
  journal = {Results in Engineering},
  volume = {32},
  pages = {112122-112122},
  publisher = {Elsevier BV},
  doi = {10.1016/j.rineng.2026.112122},
  url = {https://doi.org/10.1016/j.rineng.2026.112122}
}
RIS
TY  - JOUR
AU  - Refka Sai
AU  - Kamel Smida
TI  - Global thermodynamic consistency: A linear-algebraic diagnostic framework
PY  - 2026
JO  - Results in Engineering
VL  - 32
SP  - 112122-112122
DO  - 10.1016/j.rineng.2026.112122
UR  - https://doi.org/10.1016/j.rineng.2026.112122
ER  - 

Related outputs