Why kiln-dried boards crack and warp: a review of drying stress and strain

Introduction

Kiln drying is where much of the value of sawn timber is made or lost. Boards that come out cupped, twisted, checked or honeycombed cost sawmills money in downgrading, rework and customer complaints. Most of these defects have one root cause: stresses and strains that build up inside the wood as water leaves it unevenly.

The review "Drying Stress and Strain of Wood: A Review", by Qin Yin and Hong-Hai Liu (Nanjing Forestry University, with the second author also affiliated with Northeast Forestry University), appeared in Applied Sciences in 2021 (volume 11, issue 11, article 5023, 19 pages). It gathers what is known about how drying stress arises, how total strain breaks down into components, how stress and strain are measured, and how artificial neural networks (ANNs) have been applied to wood drying. This article summarises that review for sawmill and kiln operators.

Research methodology

This is a narrative literature review, not an experimental study. The authors state that few reviews of drying stress have appeared over the previous ten years, and they set out to summarise theory and test methods. The text does not describe how sources were searched for, screened or selected, so the coverage should be read as the authors' own selection rather than a systematic review.

The paper is organised in six parts:

1. Introduction, including the unresolved concept of the fibre saturation point (FSP).
2. Wood drying stress: defects and stress, the generation mechanism, and advances in stress models.
3. The composition of drying strain: free shrinkage, elastic, viscoelastic creep and mechano-sorptive creep strain.
4. Testing methods for stress and strain, both common and modern.
5. Artificial neural networks in wood drying.
6. Conclusions and prospects.

Key findings

How drying stress arises

Drying stress comes from moisture content and temperature gradients combined with the anisotropic shrinkage of wood. The review sorts drying defects into two groups: those caused by anisotropic shrinkage (cupping, bowing, twisting, spring) and those caused by uneven drying (surface, internal and end checks, end splits, honeycombing and case-hardening).

Stress develops in three stages, according to the literature summarised:

• Early on, the surface dries first and is in tension, so surface checks may start.
• In the middle stage the shell has set and case-hardening develops.
• Later, stress reverses, with the core in tension and the shell in compression, so internal checks may occur.

Some examples quoted: in work on Betula platyphylla, when a moisture gradient existed and moisture content was below 26%, actual shrinkage strain decreased while cracking and delamination got worse; in a test on Pinus sylvestris, higher temperatures reached stress equilibrium faster.

The fibre saturation point

The review points out that the FSP is still debated. Stamm placed it at about 30% moisture content; Hoffmeyer and co-workers estimated 38.5 to 42.5% (equilibrium moisture content at 100% relative humidity), and Engelund described a transition zone rather than a single point. This matters because many models assume that shrinkage begins at the FSP.

Strain components

Total strain is described as the sum of thermal strain (considered negligible), free shrinkage, elastic strain, viscoelastic creep and mechano-sorptive creep.

• Elastic strain follows Hooke's law. The modulus of elasticity is only weakly affected by moisture above the FSP, and it falls roughly linearly with rising temperature in the studies cited (for example, tests between minus 26 and 66 degrees C at 4% and 12% moisture content).
• Viscoelastic creep is small and depends on moisture content, temperature and loading time.
• Mechano-sorptive creep occurs when load and moisture change together. In poplar tests at 85, 105, 115 and 125 degrees C with 1.0 m/s air speed, strain levelled off after a maximum, and high temperature could produce permanent set. A comparison by Bengtsson found less bending creep in wood dried at high temperature (maximum 115 degrees C for about 30 h) than at low temperature (70 degrees C for 168 h).

Reducing stress

The review cites pre- and post-treatments that reduce only part of the problem: steam pre-treatment of rubberwood for 10 h; high-temperature, low-humidity pre-treatment of tree discs; and thermal post-treatment at 200 degrees C for 1 h in superheated steam, which was reported to reduce residual stress.

Measuring stress and strain

The authors conclude that accuracy and speed of measurement are the key challenges, and that slicing and prong methods cannot capture how stress changes during drying.

Neural networks

ANNs, especially the back-propagation (BP) multi-layer network, are presented as suitable for the non-linear relations in wood drying. Known weaknesses are fluctuating convergence speed, local minima and the difficulty of choosing hidden layers and nodes. Cited applications include a larch moisture content prediction with 96.86% fitting accuracy, hardness prediction with R squared above 0.99, and a near-infrared-based ANN that predicted surface drying stress better than partial least squares. The review notes that few studies use ANNs to predict drying strain.

Practical applications for sawmills

Points from the paper

• Defects are tied to identifiable stages of the drying process, so schedule design should consider the early tension phase at the surface and the later stress reversal.
• Conventional prong and slicing tests remain the common tools in research and industry, yet they only describe stress at a moment in time, not during the run.
• Newer sensing methods (image correlation, near-infrared spectroscopy) and ANN prediction are described as promising for real-time monitoring and prediction.

Portal interpretation (not claims made by the authors)

• Kiln operators can use the three-stage description as a diagnostic guide: surface checks point to the early stage, case-hardening to the middle, internal checks and honeycombing to later reversal. Routine prong tests on sample boards remain a low-cost check of whether conditioning has relieved case-hardening.
• Because the review is based on mixed species and laboratory conditions, schedule changes should be trialled in the mill's own kiln and species before being adopted. Numbers such as 200 degrees C post-treatment are research settings, not recommended practice.
• Mills thinking about image-based or near-infrared monitoring should treat these as developing technologies. The review points to them as future directions rather than as proven kiln-control tools.

Limitations

The authors name their own limits: models rest on simplifying assumptions and so differ from experiments, Kapp's method is qualitative, prong and slicing tests cannot follow stress during drying, and few studies predict drying strain with ANNs. They state that their next step is to verify an ANN model for drying strain.

From our reading, additional limits apply. The review does not describe how literature was selected, so it may be incomplete or biased toward certain research groups. Many findings come from single studies on particular species (Betula platyphylla, Pinus sylvestris, poplar, rubberwood), and numbers cannot be transferred freely. The review is mostly descriptive and offers no quantitative synthesis or comparison of methods. This summary draws on the open-access full text; the figures and equations of the original are not reproduced here.

Conclusions

The review concludes that current pre-treatments and techniques only partly reduce drying stress defects and distortion, that prong and slicing methods still dominate despite their inability to measure stress during drying, and that digital image and near-infrared methods are gaining ground. The authors call for coupling heat and mass transfer theory with cell-wall mechanics, building real-time image monitoring of strain and cracking, and developing more reasonable prediction models. For sawmills, it is a good conceptual map of why drying defects occur and which measuring tools exist, but it is not a drying-schedule manual.

Original research

• Authors: Qin Yin, Hong-Hai Liu
• Title: Drying Stress and Strain of Wood: A Review
• Journal: Applied Sciences, 2021, volume 11, issue 11, article 5023
• DOI: 10.3390/app11115023
• Link: https://www.mdpi.com/2076-3417/11/11/5023
• Licence: Creative Commons Attribution (CC BY), open access

This article is an independent summary by the portal; the original paper should be consulted for full details.