Developing Software to Calculate Wood Kiln Schedules

The wood drying process is complex, and involves many inter-related variables. So creating software to determine drying schedules is a challenge. June 13, 2014

I am a software developer and have been approached by a kiln owner to develop an application for calculating wood kiln schedules for various hardwoods, softwoods and exotic species. After considerable research on the topic my questions are: How come this application does not already exist and is there a market for this application?

Forum Responses
(Commercial Kiln Drying Forum)
From contributor B:
Basic USDA FPL book schedules are available via the FPL website. There have also been several standalone programs over the years which can develop book based schedules. Most experienced operators use these schedules as a base, then modifying such as initial and ending temp

From Contributor W:
I'm the woodworker/kiln operator that has approached the questioner about a software program. Our purpose was to build this software so that each operator would be able to modified the program to their own kiln, start temps, end temps, barometric pressure, specialty dry rates, air velocities, efficiency losses per each kiln based on the R-values and calculated leaks, etc.

For example my DH kiln has a max operating temp of 135F so I have to convert the USDA book schedule to fit my kiln and hope that I didn't make a calculation error. This program would also be able to incorporate a precise adjusted RH to EMC at any temp automatically instead of manually looking it up on a chart. It also would be able to formulate the theoretical kiln schedule by incorporating all the operators’ personal parameters and allow the kiln operator to precisely plot the actual schedule to make sure that the schedule is on track and not drying too fast for a poor outcome or drying too slow for time and money efficiencies.

Another large area for my needs is to track my cost. I have a small kiln and I don't always fill it with every load, and as any operator knows outside air temp and initial lumber temp changes the cost of initial heat up drastically. This program would be able to calculate the cost to dry the particular load based on species, size of charge, outside air temp, starting MC, ending MC, etc. As with any software program once the operator tailors it to their own needs, all the results, graphs, charts, etc. are instantly calculated resulting in a better time efficiency and limits human error. So, we are looking for input on if there is a market out there that is also looking for all this info that I am looking for.

From Gene Wengert, forum technical advisor:

When we wrote the book that Contributor B mentioned I had a program (FORTRAN language) that came up with the schedule, including EMC, RH and temperatures in F and C. The book has hundreds of schedules. For DH you can check the accuracy of your schedule by comparing EMC or RH with the published values. There was also an energy usage prediction program using kiln schedule, outside air temperature, MC, etc.

We do not know how to factor in many of the variables you mention, such as air velocity and barometric pressure. We do not know how to plot the schedule as a function of time (hardwoods use MC based schedules; time depends on thickness of the lumber, green MC, air flow and so on) and compare to actual to see if it is going too fast or slow. Most kilns have plenty of power so R-factor is not an issue in time, but it does affect energy use a few percentage points. Equations do exist for calculating EMC, RH, etc. As energy cost is a large component (50%) in drying costs, it is easy and probably essential to measure energy use to help calculate costs. Much energy is lost through the doors, vents, floor and frame, so Wall or door R-value alone is not enough. Drying cost is based primarily on drying time and time is based on the initial and final MC. However, knowing the schedule is not enough to calculate and predict drying time. With all the unknowns, a prediction would be filled with uncertainty, so the usefulness would be limited to a commercial operation.

So, based on my views as listed above, I do not see a market for such a program. In my experience, I seldom see human error as causing problems in drying. Rather, problems are more often due to lack of management direction or lack of understanding.

From Contributor W:
I would be a fool to ask for your professional opinion on drying lumber and then just dismiss it, so it's not looking like we will continue at building a program unless a bunch of other operators show interest or I can convince you that it can be done! Now I'm a new kiln operator and I'm really trying to learn everything I can about drying lumber so that I can be more efficient and do the best job I can, so I'm asking all of this to better understand this whole process. So far all I’ve received is guidance from a sawmill/kiln owner/operator and from the USDA Dry Kiln Operator's Manual, and from reading your post I guess I've been missing something here.

Between all the charts, graphs, schedules that are published, all of this has come from some type of mathematical equation and/or formula based on either scientific proof or simply from real-life trial and error. So we already have all this info that is scattered all over, where a software program would compile everything including the variables. For example, someone (maybe or most likely even you) at some time already calculated the max rate that water can be drawn out of each type of species of wood at a safe rate as to not collapse the cells, etc. The equations for relating temperature, humidity, and moisture content have already been published in the USDA Dry Kiln Operator's Manual. If there is variable in that formula due to Vapor Pressure due to change in elevation or barometric pressures, and being the relationship between vapor pressure and temperature is non-linear then the software program would also be able to use the Clausius–Clapeyron relation formula to recalculate for these variables (I'm just using this as an example).

Even for energy efficiency portion of a software program that I'm interesting in, once I enter in the volumetric air volume, lumber volume along the R-values of the walls ceilings, doors, floor, I know my own R-values, but if I didn't I could look them up, someone somewhere has already calculated all those numbers. Once all values have been inputted the software would do its thing and calculate everything needed instantly according to all values inputted; outside temp, outside RH, specific gravity, initial MC, initial wood temp, volumtric area of wood, thickness of lumber, amount of water, cost of KW/hour, etc. It could even be linked to your own local weather station to get accurate real-time temps, outside RH and atmospheric pressures etc. and current electrical rates, etc. However I understand there would have to be an estimation on efficiency lost through door seals, cracks, vents etc. That I understand, but so far that is the only questionable area that I can't come up with a formula for other than a standard 8, 10, 12, or 15% loss by guessing. In other words everything that I have read so far can be calculated from a formula, including all the variables, and once the operator enters these variables the software will update everything. So I guess my overall question to you is; what variables in your professional experience cannot be explained in a chart/graph/formula?

From Gene Wengert, forum technical advisor:
Certainly one big issue in drying is the natural variability of wood. This cannot be programmed. Also, we do not know the effect of velocity very precisely. I do think that Lignomat computer control systems can be added to most kilns and they are super. Also, when you get a new kiln, the controls can be excellent from most companies and they will have all sorts of information and calculations.

I think one of the big needs in our industry is for a spread sheet to accept moisture section and sample board weights and then automatically calculate MC, speed, suggested kiln settings. Maybe even make it into an App for I-phone, I-pad, etc. I have seen a few of these, but they are not for sale or are not stand alone products but come with a new kiln package.

If you are drying hardwoods, you need Drying Hardwood Lumber instead of the manual you mention. You can find a copy here at WOODWEB. Although we can calculate a lot of different variables and include effects like barometric pressure and heat loss, the results will not change how we run a kiln. We run the kiln based on the MC of the wettest samples, along with knowledge about wood quality and drying defects. These factors are the big factors and they are quite variable.

From Contributor W:
Gene, again thank you very much for your input. I think you have answered a lot of my lingering questions. As for the effect of velocity, we know the safe rate that moisture can be extracted, and there are formulas that can explain the amount of time it takes for moisture to combine with the air. So if we know how far the air has to travel across the load and its velocity I think we can formulate from all of that info a minimum velocity so that the leading edge of the load isn't discharging more than the back due to the air being over saturated for the temperature. Obviously this would be a theoretical value to be used as a baseline for each individual kiln and species. Anyway, I understand a little better now. The guy that helped me design and get my kiln started said that drying was 50% by book and 50% art, and you are proving that to me.

Now as for those spreadsheets you mentioned, I will talk to the software programmer of ZINCASTLE Software Systems offline. I know he can meet your needs on all of those. As to if it would be financially reasonable I'll let him ask you and other operators all those pertinent questions.

From Gene Wengert, forum technical advisor:
The velocity effect on wet wood is fairly predictable in terms of how the drying rate is affected. What we do not know is how the moisture is traveling within the wood. So, as the wood dries, the velocity factor is less important and we are instead waiting for the moisture to move within the wood and not for the air to scrub the moisture off the surface. At this point, accurate predictability becomes difficult. We have tried to use a form of Fick's Law, but the diffusivity coefficient is changing with moisture as well, in unknown manner, so that is tough. Of course, we can get actual drying data and use that. Once we know the wood's behavior, we can easily calculate the change in RH and drying rate through the load. For a fast drying permeable species like southern pine, we can do a better job of prediction.

From Carl Hagstrom, Systems Administrator at WOODWEB
I've included the link for the manual Gene refers to - you can download a copy by following the link below.

Drying Hardwood Lumber

From Contributor W:
Gene, again thank you. I just read a little about Fick's Law, very interesting, I am surprised that Fick's second law won't be useful where the diffusion coefficient is not a constant, as it is in wood from what I'm getting from your responses - I guess Mother Nature holds the right to break all the laws.

From Gene Wengert, forum technical advisor:
Fick's Law does apply to wood, but the coefficient is so variable (species, grain angle, density, MC, temperature, RH, thickness, shrinkage, and maybe more) that we cannot use a precise value. A typical value will get us into the ballpark, but so will an educated guesses. This was the subject of my PhD 35 years ago.

From Contributor H:
I have been involved with kiln automation and schedule development for the last eight years. I have played around a lot with different approaches to drying, to attempt to take the guess work out of the operator's hands. This proves extremely difficult due to lots and lots of variables. I am currently testing an approach where we run at a single DB setpoint, and observing the rate of change of the WB, and then ramping the WB setpoint up or down (by 0.1 Deg C at a time), according to this rate of change. My assumption is that if the WB does not increase due to moisture coming out of the timber, the drying gradient aka WB depression is not big enough. If it changes to fast on the other hand, the drying rate is too harsh. This approach more or less reduces the amount of variables considerably. Essentially, I think this is what the textbook drying schedules are trying to achieve through ramping the WB down. Has anyone else played around with this?

From Gene Wengert, forum technical advisor:
An approach that considers the drying rate of wood is to look at the change in the DB temperature, which directly indicates the energy loss, as the air travels through the load and of course, energy loss is almost 100% going to evaporation of water. So, the DB drop across the load concept, developed for pine lumber in 1978 in Virginia with Union Camp, is an excellent way to judge the rate of drying of a load. It works very well for softwoods, where the numbers are large enough and the noise is only a small component. However, with hardwoods, there are too many variables and control issues in a large kiln to provide reliable readings - the noise is too large. Plus in hardwoods, we look at individual pieces of lumber (the wettest and driest typically) and make decisions on these extreme pieces rather than the average MC. When we look at a typical kiln, even with proportional controls, we still see variations in conditions due to control issues.

I am somewhat confused by your control based on the rate of change of the WB. So, do you set a given WB and DB and then run the kiln, controlling only the DB at a constant value and let the WB change over time? Then you look at this rate of change? It would seem that if the rate of change is something like the WB is increasing at 0.1 degree per 15 minutes (meaning the RH is increasing), then you may have the correct WB setting (and a reasonable EMC so that you do not over stress the lumber when above 50% MC) but no venting. With venting that will bring the rate of change down to zero. That is how a typical kiln operates and dries lumber efficiently and with good quality - most of the time the rate of change of the WB is zero; it is only for a short period (30 minutes) for five times in the schedule that the WB is changed (and also sloppiness of the control system will cause small changes plus and minus, but average zero). However, even at zero rate of change, we could be drying the lumber too fast or too slowly or just right. Can you explain more completely what you are doing? Are you doing this with softwoods or hardwoods, green or air-dried?

From Contributor H:
First, let me state that I am not a scientist, nor an academic. I am however passionate about timber drying. My observations and assumptions are not based on scientific fact or lab proven methods, rather on observing actions and reactions inside the control process of drying and trying to find common sense reasons why they occur and how to use it to our advantage in timber drying.

In setting up custom drying schedules below 100 Deg C for freshly cut pine species found in South Africa (Pinus patula, P.teada, P.elliotii, and P.radiata) I spend a lot of hours observing the effect of active venting on the drying speed and quality. Starting with prescribed drying schedules as a guide, I specifically looked at the effect different settings had on venting. Even when using in kiln moisture meters, in conjunction with prescribed drying schedules, there were times when the vents would stay closed for hours, or open for extensive periods. This did not give the required results.

I came to the conclusion that when running at a single DB and WB setting, eg 80/63 the initial venting times would be long. i.o.w. while there was still a lot of moisture in the timber, vents would be open for long periods, and closed for short periods. As the timber dries out, this reverses to a state where the vents stay closed for long periods and only open for very short periods. The rate of change I referred to is the speed by which the WB actual temperature increases when vents are fully closed, or how long the vents remain closed before enough moisture has evaporated to reach vent setpoint. When the WB temp does not increase anymore, it would indicate a state of equilibrium, or evaporation equal to leaks in the kiln structure, or at worst case hardening which restricts the flow of moisture to the surface. In all three cases vents would stay closed.

Using these assumptions I programmed the PLC with only 3 basic phases:

1. Heat up – DB equals WB until wood core temp reaches my first drying phases WB/vent Setpoint.

2. Phase 1: Run a set DB & WB - 80/63 until MC content gets to ±27% (in kiln moisture meter).

3. Phase 2: Go into an auto vent phase that would decrease the vent temp if the vent stays closed for too long (settable variable) or increase if it stays open for too long (settable variable). End MC 12%.

The end result was flat timber with a very even moisture distribution and very little kiln defect (I did not find a necessary to run a conditioning phase after drying). I think that the time spend running at a steady RH, helped in smoothing out average MC in the load.

Has anybody tried something similar with schedules? Obviously this cannot be applied to all timbers and all kilns. This was done on a compartment/batch type kiln that dries ± 40m3 of 38mm thickness boards at a time. We have gone up to 95 Deg C, but kept the phase2 RH around 44%. Drying quality decreased slightly, but was still acceptable.

From Gene Wengert, forum technical advisor:
I think that you are using a few words that have a different meaning here in the U.S. So, to help other readers of your posting casehardening is a stress condition and is not related to anything being harder, and also develops very early in drying, has not been shown to restrict moisture movement to the surface or restrict evaporation. Perhaps overly dry wood might seem to do that. In veneer drying and log drying, I have heard the word casehardening used when the surface gets real dry and seems very hard when you try to nail it, but this is not accurate use of the word for lumber drying.

Also, to achieve uniform MC, we equalize (not condition) the load. Conditioning is the aggressive steaming of lumber to remove drying stress. It is done at 4% EMC above the target MC. So, in your case with 12% MC target, we would condition (if possible) at 16% EMC. Many U.S. kilns are not well enough insulated to develop such high humidity’s.

A load of lumber will release three-four U.S. gallons of water per 1% MC loss per MBF. So, in a large kiln, we might be looking at ten gallons of water loss per hour. If the lumber is wet and the kiln is hot and the humidity is well below 100% RH and the fans are working, how can the lumber not be drying? So, if the vents do not open, that means that either drying has stopped or that there is some other way that the evaporated moisture is leaving the kiln, perhaps through leaks or through condensation.

I have seen kilns that run on the basis of constant venting. The idea is that because the drying rate is constant (almost), then a control system that uses constant venting will be a good way to control the drying rate. The control system opens the vents just a little bit so that the vents are open all the time and change the amount they are open very little, hour after hour. Early attempts to do this used a control system that was not adequate, so overall the idea did not work out well in practice. A similar approach was used by Dallas Dedrick in his CRT (Constantly Rising Temperature) patented idea used for softwoods on the West Coast of the USA. The idea was also tried for southern pine, but the reports I received were not as good as hoped for.

From contributor M:
I don't have a lot of experience with drying, but I've been reading up on it and I'm an engineer. What Gene is saying sounds right to me. The third key parameter is air speed. If it is too low, then the water removal rate will be low, even if the DB and WB are on schedule, especially at green MC. There are a lot of ways to determine the rate of drying. You can measure the MC of a sample, or weigh a sample manually or automatically. Or you can measure the volume or weight of the water coming out of a DH. As someone mentioned above, the water evaporation reduces the DB across the stack. You could also theoretically calculate the water removal rate, if you calculated the air volume leaving the vent as the average air speed of the exhaust times the cross-sectional area of the vent. Instead of an open vent, you could also use an exhaust fan with a known flow rate, perhaps with a flap that closed with gravity when the fan was off. The volume exhausted is just this flow rate times the length of time the exhaust fan is on. Technically you'd also need to know the humidity of the intake air, if that's a factor, since it brings some (presumably small) amount of water into the kiln.

I've worked through some of the calculations on how to convert relative humidity and temperature to absolute humidity. That's in units of mass, which makes it easier to account for where the water is going. On the issue of a software app, I'd be interested to be a user and developer if it was open-source. I think the first requirement is just to have a giant look-up-table for existing drying-schedules. After that, system modeling might be useful, but it might also be extraneous. I mean if the schedules are optimized, then there's little point in intentionally deviating from them, right? Maybe where models could be useful is where the drying-schedule is accidentally deviated from, through power failure or equipment failure, etc, and the operator wants to know what action to take to minimize degrade and eventually get back on schedule. It could be thought of as gracefully shutting down the kiln and putting it in standby or air-dry mode. This might also have use for solar kilns that make this transition every day, or for hybrid solar kilns that may have varying schedules that are influenced by resource management.

Another use for a model is in kiln-design and optimization. But that seems more like R&D experimentation activity rather than a production tool, because there are too many variables to accurately model, right? I suppose that role of theory here is to guide the kiln-design and operation in the right direction, but then testing and experimental refinement is required for fine-tuning. I've programmed some basic models from scratch, and that is how I see them being used. I wouldn't dare suggest that someone use a theoretical model in place of an extensively developed and proven schedule, unless the model was so accurate and extensive that it could actually predict the established schedules across all sorts of variables like species, etc.

From Contributor H:
Here are some more thoughts on the subject and to add a few more variables. The kilns ability, heating capacity, air flow and venting capacity can throw the whole program out. There’s no sense in trying to get to a specific set point, but the kiln just can't reach. I guess that would be identifying the kiln's constraints, and then modeling your drying schedule around the specie, and the outcome or end use of the product or balancing the kilns constraints with the outcome - for that specific kiln. You could have two identical kilns, structurally, with the same heating and venting capacity but with different air flow capacity as pointed out. Trying to apply the same drying schedule will just not work. The RH/EMC on the lower air speed kiln will probably always be high, resulting in vents staying open for very long periods or always 100% open equals 0% control. Change just one of the constants like the fillets/sticker thickness and the whole drying curve changes. The variable hysteresis caused by the hundreds of variables will have scientist and software developers scratching their heads for a long time to come.