You're welcome! Glad I could do this for those who will appreciate it. So, Woodgeek... what do you think about the way the curve is turning? This is unexpected to me, perhaps you can shed some insight into this. You seem to have a much better grasp of the physics and math behind diffusion rates. Am I incorrect in seeing that there seems to be an exponential component to the graph, at least for the middle part of the curve? Now it seems to be almost liner, with a loss of nearly 1/2 ounce of water each day. Ain't supposed to happen that way AFAIK, but the scale disagrees. Your thoughts?

http://en.wikipedia.org/wiki/Wood_drying#A_simple_model_for_wood_drying M is the MC at time t, M_e is the equilibrium MC, M_0 is the beginning MC, and tor is a constant in this case, so as you suspected it's an exponential decay down to M_e. Interestingly, they give a figure of about 10x faster drying along the grain than across it.

Theory of course has its limits, but if we started with (1) a 'half-infinite' slab (basically, big and flat), that was (2) uniformly charged with water that (3) (effectively) diffused within the slab, AND (4) any water that reached the surface effectively disappeared, then the result is easy. The cumulative amount of water lost is NOT an exponential function of time, but goes like time^0.5. Interestingly, this formula predicts that the initial drying is not only fast, it is infinitely fast at t=0, and moreover, that the curve will stretch out much longer in time than it would if it were exponential (or you fitted the original data to an exponential). I think assumptions (2) and (3) are pretty aok. (4) requires a little thought--at finite RH there is an equilibrium MC that is reached, perhaps we can adjust our definition of mobile water to be the water that is excess of this equilibrium amount. (1) is ok so long as the thickness of the (growing) dry zone is less than the radius of the split, so it is ok for early times, and then at late times the rate of loss should go faster than the planar model (cuz we run out of water when the dry layer reaches the center). In the past you have suggested that the mobile water diffusivity must depend on MC, like there is cellular water and bound water. I don't think so, as a matter of opinion, I would expect water to move between these two populations rapidly--they would move together with an effective diffusivity. I guess that much of the mass movement is due to the small fraction of the water in vapor form diffusing in the pore spaces, and so should be roughly proportional to the saturation vapor pressure of water. The little published data I have looked at does suggests that the time to air dry dimensional lumber goes like the square of the minimum dimension, a 2" thick plank will take 4x longer to dry than a 1" plank. (for heat diffusion reasons, it might also take four times longer to burn). It also seems to take about 2-3x longer at winter temps (and 0%RH) than at summer temps (at 0%RH), consistent with my idea that is proportional to the vapor pressure of water at those two temps. I was a little busy at work recently, so didn't get to plot up/fit your data. I predict that if you plot the square of the cumulative water loss versus time, it will start out pretty linear, and flatten out at the longest time (when you reach the center of the split). I will not be cranky if the wood drying process differs from my simplistic expectations....

Yes, that's the one I was looking for! Thanks so much. I was trying to apply a standard exponential decay function starting with the original water weight like you would with a pharmaceutical half-life in the bloodstream, but when I put it into my TI-85 and graphed it, the amount remaining at each time interval wasn't coming close to what the results were showing. I'll have to play with your formula some more. I'll also have to bone up on graphing with the TI-85. It's been ten years since I've had to do that for anything. Funny, I remember finding that Wiki entry and posting a link on the forum, but then couldn't find it in my bookmarks later on. Yes, the accepted estimate for water movement along the long grain is 10-15 times as fast as across the thickness. My own experience has shown that wood dries a lot faster when cut shorter. I believe that rate changes after all the free water is removed from the wood, but I'm not sure. I'll have to read that Wiki entry again later on. Does anyone know of a down-loadable graphing calculator that I can use to enter these functions into so that I can try to display the graphs on this thread?

My experience drying wood in my garage is pretty similar. The garage is heated to around 70 by my Tarm boiler, and I run a de-humidifier when the RH goes above 50%. I also aim a big box fan at the wood. If it's below zero I turn on the radiant heat to warm the slab up a little. On Jan 8 I stacked a very green cord of hard maple, cherry, and yellow birch in the garage. I split the wood fine, no bigger than 4-5". I weighed one of the biggest chunks of hard maple at 7 lbs. 14 oz. Over the next 18 days the weight dropped to 5 lbs. 14 oz, so it lost a full two pounds of water weight. I don't have a moisture meter so i don't know the moisture content, but I'm burning it mixed in with some really dry wood. The smaller splits burn fine with no sizzling, the bigger chunks burn with some sizzling. I think the biggest factor in drying wood quickly is splitting it very fine. Then you need heat and air movement from a fan.

It would be cool to see if you can get the exponential thingy to work. I'm skeptical--the major way one gets such forms is if the rate of loss is proportional to the amount remaining, which would be correct if the water remaining were 'well mixed' and its vapor pressure was proportional to the amount left. I would expect neither of these to be true--but I guess I should go read the wiki link. Just wanted to throw down a friendly gauntlet.

I am not very analytical, nor good at graphs etc. I bring in approx. 1/2 cord of wood into the living room(stove room). Now I have two sets of tubs I store the wood in/on. So when one set of bins is empty, I reload and start burning the other set of bins. This always gives the new load plenty of time to dry any moisture that may be in the splits to the point that it all burns very nicely. I get approx. 3 weeks of burning out of a full load in the house. And it also helps humidify the house for a few days. Not very scientific, but works for me.

BK, I'd like to thank you for taking the time to do this experiment and documenting it. I can't say I'm completely surprised at the results, but I too would have guessed it would have taken a little more time than it did to get down to 20% MC, at least based on my own informal drying I do near the stove. Then again, I do tend to burn larger splits and most of it is oak... so I'm sure your decay would have been a lot less with such a subject. Looking forward to your continued results from these weekly weigh-ins. If you happen to do another, I'd love to see what would happen with 10 and 25 lb. splits of red oak! Thanks again for all your efforts!

Week Five Update... Test split is at 7.56 lbs, down 2.99 lbs since the start. Dry-basis MC is 12.8%, wet-basis MC is 11.4%. I tried to mess around with some of the more sophisticated drying models, but there are too many variables and unknowns. When I look at the data, I can see that a consistent daily drying ratio began to emerge after the initial drying period was over (faster initial rate driven in part by the more rapid end-grain drying). This varied from day to day, partly because I am unable to measure the weight of the split to the nearest thousandth, so I had to continually round off. The other reason is probably that the RH in the room varied from day to day, due to changes in outdoor absolute moisture content, indoor air demands and the periodic addition of wet wood to the room. I decided to average this ratio and use the most basic exponential growth function starting on the first day that the ratio appeared to be showing some consistency. This day was on Day 10 of the experiment, when the accumulated weight loss amounted to 1.77 lbs. The column in light green contains all of the daily drying ratios. I averaged these and got a daily weight loss ratio of 0.971439. I entered this into my TI-85 as Y= 1.77(.971439^X), set the range for decimals and traced along the graph to get the value at each integer along the "X" axis (each day). The results are is the pukey lime green columns on the right. This simple exponential decay model proved to be remarkably predictive in this case. The largest deviations were -0.04 lbs, and they all occurred during the time period shortly after bringing in large amounts of green wood. The charts show a similar deviation from a smooth exponential curve during this general time period. I continued tracing along the graph for the next seven days. I used the outputs to make daily drying predictions for the next week. I will continue to record the actual daily weight loss (real slow at this point) to see how close to the predictions the model continues to be. I will report back next week with the results. If all goes according to the predictions, we will have a furniture grade kiln-dried specimen in six weeks total time. Cheers!