Methods and arrangements for controlling a pretreatment process of biomass by colour analysis

Abstract

A method for controlling a pretreatment part in a manufacturing process of chemical bioproducts comprises measuring at least one colour coordinate value of a product that has undergone a hemihydrolysis reaction and steam explosion as parts of said pretreatment part of said manufacturing process. A decision rule is applied to make a control decision at least partly on the basis of the at least one measured colour coordinate value. One or more control actions are implemented in one or more process steps of said pretreatment part in accordance with said control decision.

Description

[0001] METHODS AND ARRANGEMENTS FOR CONTROLLING A PRETREATMENT PROCESS OF BIOMASS BY COLOUR ANALYSIS

[0002] TECHNICAL FIELD

[0003] The disclosure relates in general to controlling an industrial-scale manufacturing process of chemical bioproducts . In particular, the disclosure relates to the application of specific measurement methods at various parts of the process and to the control decisions that can be made on the basis of such measurements .

[0004] BACKGROUND OF THE INVENTION

[0005] The production of biomass-based chemicals may use for example wood particles as the main raw material . In a biomass-to-sugar process the wood particles or other biomass may be subj ected to pretreatment , which may comprise washing and impregnating with water, acid catalyst , and / or other liquids , and subj ecting to elevated temperature and pressure, in order to prepare the material for later steps of the process . The later steps may involve for example enzymatic hydrolysis , from which the sugars ( carbohydrates ) may be fed further to other processes . Such other process may involve the production of for example glycols . The enzymatic hydrolysis step may also produce lignin as one of its outputs .

[0006] In its known form, controlling the pretreatment involves a number of uncertainties . The pretreatment proces s may be des igned and optimi zed for certain nominal characteristics of the raw material , but inherent variations in what actually enters the process cause continuous fluctuation in how successful the pretreatment is in achieving its obj ectives . Effects of such fluctuation are seen for example in the material flow coming out of a hemihydrolysis reactor where the impregnated wood particles are subj ected to elevated temperature and pressure and subsequent steam explosion at the output of the reactor . I f the reaction is too severe , the product coming out of the reactor is too finely grained; in other words , the particle s i ze dis tribution of the product shows too large proportions in the smallest si ze bins , which makes the product more difficult to handle . Also , too severe a reaction may produce excessive amounts of unwanted chemical constituents like acetic acid and / or furfural , which have disadvantageous effects in the later stages of the process , while the yield of desired chemical constituents like C5 carbohydrates becomes low . Towards the other extreme , if the reaction in the hemihydrolysis and steam explosion stage is not severe enough, the product coming out of the reactor is too coarsely grained for effective use in the later stages of the process , and the yield of desired chemical constituents like C5 carbohydrates is again lower than would be possible .

[0007] It would be advantageous to have a possibility to react in real time ( or at least as quickly as possible ) to detected deviations from the expected proceeding of the process . However , it i s diff icult to obtain accurate knowledge of the current status of each step in the process fast enough . Any measurement method that is to be applied must be applicable for prolonged operation in the harsh conditions of an industrial environment , which typically makes it difficult or impossible to utili ze instruments built for use in laboratory conditions .

[0008] SUMMARY

[0009] According to a first aspect there is provided a method for controll ing a pretreatment part in a manufacturing process of chemical bioproducts . The method comprises measuring at least one colour coordinate value of a product that has undergone a hemihydrolysi s reaction and steam explosion as parts of said pretreatment part of said manufacturing process , applying a decision rule to make a control decision at least partly on the basis of the at least one measured colour coordinate value , and implementing one or more control actions in one or more process steps of said pretreatment part in accordance with said control decision .

[0010] According to an embodiment , said measuring of at least one colour coordinate value comprises measuring an a* value , a b* value , or an L* value , where said a* value , b* value , and L* value are coordinate values relative to the green-red opponent colours , the yellowblue opponent colours , and the black-white opponent colours of an CIELAB colour space respectively . This involves at least the advantage that a well-defined colour measurement system can be utili zed, which gives reasonably clear and unambiguous indications of the current severity of the reaction in the hemihydrolysis reactor .

[0011] According to an embodiment , said measuring of at least one colour coordinate value comprises measuring of at least one colour coordinate value comprises measuring an R value , a G value , or a B value , where said R, G, and B values are coordinate values of an RGB colour coordinate system indicative of relative amounts of red, green, and blue respectively . This involves at least the advantage that a well-defined colour measurement system can be utili zed, which gives reasonably clear and unambiguous indications of the current severity of the reaction in the hemihydrolysis reactor .

[0012] According to an embodiment , said measuring of at least one colour coordinate value comprises measuring a C value , an M value , a Y value , or a K value , where said C, M, Y, and K values are coordinate values of a CMYK colour coordinate system indicative of relative amounts of cyan, magenta, yellow, and key colours respectively . This involves at least the advantage that a well-defined colour measurement system can be utili zed, which gives reasonably clear and unambiguous indications of the current severity of the reaction in the hemihydrolysis reactor .

[0013] According to an embodiment , said applying of a decision rule to make said control decision involves making a decision intended to maintain at least one said colour coordinate value in , or to direct said at least one colour coordinate value towards , a predetermined target range in subsequent similar measurements . This involves at least the advantage that consistent control decisions can be made with a reasonably good assumption of the results being as desired .

[0014] According to an embodiment , the method comprises , in the following order : taking a sample of said product that has undergone said hemihydrolysis reaction and steam explosion, processing said sample to standardi ze its measurement characteristics for the measurement , and performing said measuring of said at least one colour coordinate value . This involves at least the advantage that consistency of measurement results is improved .

[0015] According to an embodiment , said processing of said sample comprises preparing a standardi zed surface , such as a flat surface , of said sample . This involves at least the advantage that consistency of measurement results is improved .

[0016] According to an embodiment , said processing of said sample comprises standardi zing a moisture content of said sample . This involves at least the advantage that consistency of measurement results is improved .

[0017] According to an embodiment , the method comprises , in the following order : placing a predetermined amount of said product in an open vessel to form said sample , flattening an exposed surface of said sample , applying heat to said sample to dry said sample, allowing said sample to cool , and performing said measuring of said at least one colour coordinate value on the flattened exposed surface of said sample . This involves at least the advantage that consistency of measurement results is improved .

[0018] According to an embodiment , said pretreatment part in the manufacturing process of chemical bioproducts comprises impregnating wood particles in an acidic impregnating solution before subj ecting the impregnated wood particles to said hemihydrolysis reaction . Said implementing of one or more control actions may then comprise at least one of : changing an acid concentration of the impregnating solution, changing a residence time of the wood particles in said impregnating solution, changing a storage time of the impregnated wood particles between said impregnating and said subj ecting of the impregnated wood particles to said hemihydrolysis reaction . This involves at least the advantage that control actions can be taken in a logical and verifiable way .

[0019] According to an embodiment said implementing of one or more control actions comprises changing a selected feedstock composition, wherein said feedstock comprises said wood particles before said impregnating . This involves at least the advantage that control actions can be taken in a logical and verifiable way .

[0020] According to an embodiment , said implementing of one or more control actions comprises at least one of : changing a residence time of the process stream in said hemihydrolysis reaction, changing a temperature to which the process stream is subj ected in said hemihydrolysis reaction, changing a pressure to which the process stream is subj ected in said hemihydrolysis reaction . This involves at least the advantage that control actions can be taken in a logical and verifiable way .

[0021] According to an embodiment , said measuring of at least one colour coordinate value comprises measuring two or more colour coordinate values . Said applying of a decision rule may then comprise making said control decision at least partly on the bas is of the measured two or more colour coordinate values . This involves at least the advantage that statistical representativeness of the measurement results is improved .

[0022] According to an embodiment , said applying of a decision rule comprises making said control decision on the basis of a further result calculated from the measured colour coordinate values . This involves at least the advantage that statistical representativeness of the measurement results is improved .

[0023] According to an embodiment , said applying of a decision rule to make said control decision involves making a decision intended to maintain said further result in, or to direct said further result towards , a predetermined target range in subsequent similar measurements . This involves at least the advantage that consistent control decisions can be made with a reasonably good assumption of the results being as desired .

[0024] According to an embodiment , said measuring of at least one colour coordinate value comprises measuring an a* value and a b* value , where said a* value and b* value are coordinate values relative to the green-red opponent colours and the yellow-blue opponent colours of an CIELAB colour space respectively . Said applying of a decision rule may then comprise making said control decision intended to maintain a sum of the measured a* and b* values in, or to direct said sum towards , a predetermined target range in subsequent similar measurements . This involves at least the advantage that a representative indication of properties of the sample can be used as a basis for making control decisions .

[0025] According to an embodiment , said applying of a decision rule comprises using said measured at least one colour coordinate value as an input to an artificial- intelligence-based decision-making system and allowing said artificial-intelligence-based decision-making system to make said control decision . This involves at least the advantage that such complicated systematic trends may be utili zed that would be too difficult to find and utili ze otherwise .

[0026] According to an embodiment , said artificial- intelligence-based decision-making system uses also other inputs than said measured at least one colour coordinate value to produce said control decision, said other inputs being indicative of further observed characteristics of the manufacturing process . This involves at least the advantage that even more complicated decision-making rules can be applied in a consistent and reliable manner .

[0027] According to a second aspect , there is provided an apparatus for controlling a pretreatment part in a manufacturing process of chemical bioproducts . The apparatus comprises measurement means configured to measure at least one colour coordinate value of a product that has under-gone a hemihydrolysis reaction and steam explosion as parts of said pretreatment part of said manufacturing process , decision-making means configured to receive the measured at least one colour coordinate value and to apply a decision rule to make a control decision at least partly on the basis of the at least one measured colour coordinate value , and control means configured to receive , as a result of said control decision, instructions to implement one or more control actions in one or more proces s steps of said pretreatment part in accordance with said control decision, and configured to act accordingly to implement said one or more control actions .

[0028] According to a third aspect , there is provided the use of a method of any kind described above to control a manufacturing process of chemical bioproducts .

[0029] BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The accompanying drawings , which are included to provide a further understanding of the invention and constitute a part of this specification, illustrate embodiments of the invention and together with the description help to explain the principles of the invention . In the drawings :

[0031] Figure 1 is a high-level block diagram of a manufacturing process of chemical bioproducts , figure 2 illustrates the process steps of an exemplary pretreatment process , figure 3 illustrates three steps of a control method, figure 4 illustrates five steps of a method, figure 5 illustrates measurement results from example cases , figure 6 illustrates measurement results from example cases , figure 7 illustrates measurement results from example cases , figure 8 illustrates measurement results from example cases , figure 9 illustrates measurement results from example cases , figure 10 illustrates measurement results from example cases , figure 11 illustrates measurement results from example cases , figure 12 illustrates measurement results from example cases , figure 13 illustrates measurement results from example cases , figure 14 illustrates measurement results from example cases , figure 15 illustrates measurement results from example cases , figure 16 illustrates measurement results from example cases , figure 17 illustrates measurement results from example cases , and figure 18 illustrates measurement results from example cases.

[0032] DETAILED DESCRIPTION

[0033] In the context of this text, cellulose is taken to mean at least one or even all of: fibers, fiber particles, cellulose, glucane, oligomeric glucose. In the context of this text, hemicellulose is taken to mean at least one or even all of: xylan (like glucuronoxylan and arabinoxylan) , xylooligomers, other hemicellulosic oligomeric sugars.

[0034] In the context of this text, wood material may be selected from a group consisting of hardwood, softwood, and their combination. The wood material may e.g. originate from pine, poplar, beech, aspen, spruce, or birch. The wood material may also be any combination or mixture of these. Preferably the wood material is broadleaf wood due to its relatively high inherent sugar content, but the use of other kinds of wood is not excluded. The expressions wood particles, wood chips, and chips are used interchangeably and mean wood material that has been mechanically broken, chipped, and / or crushed to have a distribution of particle sizes mainly ranging from a few millimetres to a few centimetres.

[0035] Fig. 1 illustrates schematically a manufacturing process of chemical bioproducts from wood material. The process can be roughly divided into a wood handling phase 101, a wood-to-sugar phase 102, and a sugar-to- chemical phase 103.

[0036] The wood handling phase 101 comprises mainly mechanical processing such as debarking 111 and chipping 112.

[0037] The wood-to-sugar phase 102, which is also called the wood-to-sugar process, comprises a pre-treat- ment part where the wood chips from the wood handling phase 101 are taken through impregnating 121, hemihydrolysis 122, and steam explosion 123 in order to break down the structure of the wood material and to remove the C5 sugars . Impregnating is typically part of processes that utili ze an acid catalyst , so it may be omitted in processes that rely upon autohydrolysis . The main process stream continues into enzymatic hydrolysis 124 , where the aim is to convert polysaccharides into C6 monomers , essentially converting glucan into glucose . Lignin and other remaining solids are removed after the enzymatic hydrolysis , and the obtained C6 sugars are fed further to a sugar-to-chemical phase 103 . The removed lignin may be utili zed further in other processes .

[0038] The subsequent utili zation of the sugars in the sugar-to-chemical phase 103 may comprise steps such as purif ication 131 of the sugars (both C5 and / or C6 carbohydrates ) and one or several sugar conversion proces ses 132 . The sugar conversion proces ses 132 may include processes such as fermentation to produce alcohols or catalytical hydrotreatment to produce glycols .

[0039] Fig . 2 illustrates an example of a product flow through various stages that all belong to the pretreatment part of fig . 1 . Washing 201 is done with water, removing some mainly inorganic impurities such as sand . Washed wood particles are taken to steam treatment 202 for the purpose of removing air from inside the wood particles and to preheat them to an elevated temperature . Steam-treated wood particles are taken to dilute acid treatment 203 for impregnating them with a dilute acid solution . The aim of the dilute acid treatment 203 is to make the di lute acid solution penetrate into the wood particles as evenly as possible .

[0040] The acid-impregnated wood particles are taken to hemihydrolysis at 204 where they are under elevated pressure and temperature . At the output of the hemihydrolysis 204 the wood particles undergo a steam explosion that breaks their structure . The output stream from the hemihydrolysis and steam explosion 204 goes through steam separation (not separately shown) to mixing 205 where water is added and the resulting mass is homogeni zed mechanically to break up agglomerates . Solids and liquids may then be separated at 206 for feeding into later process stages .

[0041] It has now been found that it is possible to draw relatively reliable conclusions about the severity of the hemihydrolysis reaction and use such conclusions to control and optimi ze the process , by analysing the colour of the product that has undergone the hemihydrolysis reaction and steam explosion as part of the pretreatment . Based on thi s f inding, a method has been developed for controlling a pretreatment part in a manufacturing process of chemical bioproducts . The method is generally illustrated in fig . 3 and comprises

[0042] - measuring at least one colour coordinate value of a product that has undergone a hemihydrolysis reaction and steam explosion as parts of said pretreatment part of said manufacturing process ( step 301 ) ,

[0043] - applying a decision rule to make a control decision at least partly on the basis of the at least one measured colour coordinate value ( step 302 ) , and

[0044] - implementing one or more control actions in one or more process steps of said pretreatment part in accordance with said control decision ( step 303 ) .

[0045] Colour coordinate values represent a systematic way of describing a colour . Several colour coordinate systems have been developed, such as CIELAB, CIELUV, CIEUVW, RGB, CMYK, YIQ, YPbPr, xvYCC, HSV ( or HSB) , HSL, HEX, LCH ( ab) , CIEXYZ , and LRV . Commercial systems for systematically expressing colours are the Munsell colour system, the Pantone Matching System, and the Natural Colour System . A common feature of all such systems is the possibility to express a measured or otherwise observed colour as a relatively limited set of numerical values known as colour coordinate values .

[0046] As an example , the CIELAB colour coordinate system uses three colour coordinates L* , a* , and b* . The CIELAB (or CIE L* a* b*, where CIE means the Commission Internationale de 1 'Eclairage) colour system represents quantitative relationship of colours on three axes: L* value indicates lightness, and a* and b* are chromaticity coordinates. On the colour space diagram, L* is represented on a vertical axis with values from 0 (black) to 100 (white) . The a* value indicates a red- green component of a colour, where positive a* and negative a* indicate red and green values, respectively. The yellow and blue components are represented on the b* axis as positive b* and negative b* values, respectively.

[0047] The value space of CIELAB colour coordinates may be represented in a Cartesian coordinate system with a sphere of radius 50 and centre point at (0, 0, 50) , so that the Cartesian x, y, and z axes correspond to a*, b*, and L* values, respectively. The x-y plane, or a*- b* plane, may be called the chromaticity plane. At the centre point of the chromaticity plane is neutral or achromatic. The distance from the central axis represents the chroma or saturation of the colour, while the angle on the chromaticity plane represents the hue.

[0048] The sphere-like form of the value space of CIELAB colour coordinates means that there is a maximum value range for the a* and b* values at the middle L* value (L=50) while at the extreme L* values (L=0 and L=100) the value range for the a* and b* values reduces to a point.

[0049] In many cases, the value ranges (particularly the maximum a* and b* value ranges) may be scaled and / or otherwise modified to better match some available cal- culational resolution. For example, the basic [-50, 50] maximum range of the a* and b* values may be replaced with the range [0, 100] , which in said Cartesian coordinate system means transferring the perpendicular L* axis to point (50, 50) on the chromaticity plane. Additionally or alternatively, if a digital processing system has 8 bits allocated for expressing each of the a* and b* values, the binary value range [00000000, 11111111] for each of them may correspond to a decimal value range [-127, 128] or [0, 255] , which can then be mapped to the basic [-50, 50] or [0, 100] range with a trivial mathematical operation. In this text, the ranges of the a* and b* values are considered in the decimal value range [-127, 128] .

[0050] The L*, a*, and b* values can be transcribed to parameters specific to an area of application. For example, in dermatology the L* value correlates with the level of pigmentation of the skin. The a* value correlates with erythema and the b* value correlates with pigmentation and tanning.

[0051] The light source used has significance to measuring colour coordinate values. In the experiments during which the present invention was made, a light source with the standardised D50 output spectrum and 2 degrees output angle was used. It is possible to use a light source of some other kind, like a light source with the standardised D65 output spectrum for example. As the selected light source may affect the measurement results, it is advisable to check, with the aid of paral- lelly made laboratory analyses of the product output from the hemihydrolysis reaction for example, that a selected target value or target range for a measured colour coordinate corresponds properly to the desired characteristics of the product.

[0052] If the CIELAB colour coordinate system is used, step 301 in fig. 3 involves most advantageously measuring an a* value. This is because it has been found that of the L*, a*, and b* values of samples of the kind considered here, the a* value is the most stable against random variation. For example, the L* value depends heavily on the moisture content of the measured sample and consequently changes when the surface of a moist sample is allowed to dry in ambient air. It appears that the L* value is sensitive to the amount of light scattering that occurs at the sample surface , whi le the a* and b* values are les s sensitive . Thi s dependence of the L* value on scattering may be associated with the dependence of moisture content mentioned above , because how moist or dry the sample is has an effect on the microstructure of the sample surface .

[0053] On one hand, there i s a certain optimum level of severity of the hemihydrolysis reaction . On the other hand, a relatively unambiguous dependence has been observed between the severity of the reaction and the measured colour coordinate value of the resulting product . This dependence may be analysed by running the pretreatment process at various levels of reaction severity, obtaining corresponding samples of the pretreated product , performing careful laboratory analyses of the samples to reveal their mechanical and chemical properties ( for which certain optima are known) , and measuring their appropriate colour coordinate values . Thus , it may be possible def ine a predetermined target range in which the colour coordinate value should be as an indication of the hemihydrolysis reaction having optimum severity .

[0054] Laboratory analyses of the product of the hemihydrolysis reaction take more time and are more laborious than a simple measurement of the colour coordinate , but they give more accurate , absolute information about those characteristics that tell , how severe the reaction has been . It may be possible to repeat a round of laboratory analyses every now and then and use the updated knowledge to update the target range in which the colour coordinate value should be as an indication of the hemihydrolysis reaction having optimum severity . Thus , the target range may change over time .

[0055] According to an advantageous embodiment , step 302 in fig . 3 involves making a decision intended to maintain the colour coordinate value ( for example : the a* value) in the predetermined target range in subsequent similar measurements. Alternatively, if the measured colour coordinate value was not currently within the predetermined target range, step 302 may involve making a decision intended to direct the colour coordinate value (for example: the a* value) towards the predetermined target range in subsequent similar measurements .

[0056] The target value or target range, in which the colour coordinate value should preferably be, may change depending on what specific characteristics are expected of the product output from the hemihydrolysis reaction. For example, one may aim at a target amount 90% of monomeric xylose compared to total amount of soluble xylose and xylan. In such a case, an example of a target value for the measured a* value could be 15.0. In the graphical representations of acquired actual measurement results shown later, fig. 5 shows how the measured a* values settle roughly on a falling line, according to which an a* value of about 15 seems to correlate with a measured amount of about 90% of monomeric xylose compared to total amount of soluble xylose and xylan.

[0057] According to another example, one may aim at a target amount 70% of monomeric xylose compared to total amount of soluble xylose and xylan. In such a case, an example of a target value for the measured a* value could be 17.0 (see again fig. 5) .

[0058] According to yet another example, one may aim at a target amount 0.11% of HMF (hydroxymethyl furfural) compared to a measured Brix-value in liquid fraction. In such a case, an example of a target value for the measured a* value could be 16.1. In the graphical representations of acquired actual measurement results shown later, fig. 9 shows how the measured a* values settle roughly on a falling line, according to which an a* value of about 16.1 seems to correlate with an amount about 0.11% of HMF (hydroxymethyl furfural) compared to a measured Brix-value in liquid fraction . As an example of a control decision in such a case , if the measured a* value i s 15 . 0 , one pos sible corrective action could be to decrease the residence time in the hemihydrolysis reaction from 8 minutes down to 5 minutes . This would most probably change the value of the measured colour coordinate ( the a* value ) to the target level .

[0059] The principle of using a measured colour coordinate value as a bas is of deciding about control actions can be applied regardless of the details of how the measurement is made . In general , the measurement may be continuous and directed to a product flow ( or some part thereof ) coming out of the hemihydrolysis stage of the process . Alternatively, the measurement may be batch-like , based on taking a sample of the product and measuring the colour coordinate value from said sample in isolation from the continuous product flow .

[0060] In the following, an example is explained in which the latter alternative is applied, with reference to fig . 4 . This example may be characterised as comprising the steps of taking a sample of said product that has undergone said hemihydrolysis reaction and steam explosion, processing said sample to standardi ze its measurement characteristics for the measurement , and performing said measuring of said at least one colour coordinate value .

[0061] Step 401 represents taking the sample in fig . 4 . In order to exclude as many sources of possible random variation as possible , it is advantageous to ensure that the sampling takes place in the same way each time . For example , taking the sample at step 402 may involve al ways using a sample cup of same si ze and material , and filling the sample cup to the same level each time . Certain further steps of the method may benefit from the sample cup being of a material with high thermal conductivity, such as aluminium . For ease of later handling, step 401 may comprise placing a predetermined amount of the product in an open vessel to form said s amp 1 e .

[0062] Steps 402 , 403 , and 404 can all together be designated as processing the sample to standardi ze its measurement characteristics for the measurement . Not all these steps may be needed, and their order may be dif ferent than in fig . 4 . Also other steps than these may be involved in standardising the measurement characteristics of the sample for measurement .

[0063] Step 402 involves preparing a standardi zed surface , such as a flat surface , of the sample . The measurement of a colour coordinate may be affected by how flat or uneven the measured surface is , so one advantageous poss ibil ity is to always flatten an exposed surface of the sample at step 402 . Thi s may be achieved for example by pressing a flat obj ect , such as a glass plate , against the exposed surface of the sample . An alternative is to always util i ze a selected obj ect with some well-known uneven surface to press against the exposed surface of the sample , to ensure a standardised uneven texture of the sample surface .

[0064] Step 403 involves standardi zing a moisture content of the sample . As already indicated above , the measurement of a colour coordinate may be affected by how moist or dry the sample is , so one advantageous possibil ity is to apply a suf ficient amount of heat to the sample to essentially dry it of all moisture that can be vaporised . An example is to place the sample to an oven, like an oven of 105 degrees centigrade for example , for some standard time like 30 minutes . Here it is advantageous to have the sample-holding vessel made of a thermally conductive material , because that facilitates effective conducting of heat into all parts of the sample . Another possibility of standardi zing moisture content at step 403 involves adding a sufficient amount of purified water or other selected liquid to bring the moisture content of the sample to a selected target level , like dry matter content of 30 % in weight . Standardising the moisture content this way, i . e . by adding liquid, may simultaneously involve an easy way of standardising sample temperature , because one may choose to always add liquid of a particular, selected temperature .

[0065] Step 404 involves standardising a temperature of the sample for the colour coordinate measurement . The method of measuring the colour coordinate may be one that requires the measured sample to have a certain temperature : for example , a measurement head that comes against the sample surface may be one that does not tolerate samples hotter than some predetermined safety limit . Additionally or alternatively, the sample may have such a nature that its observed colour depends on temperature , which is then another reason for always measuring the colour coordinate ( s ) at a standardised temperature . I f the temperature-standardising step 404 is involved the method, and if it was preceded by applying heat to the sample for drying, step 404 may involve for example allowing the sample to cool , either in some dedicated cooler or in ambient conditions , until its temperature is below some limit , like 40 degrees centigrade .

[0066] As suming that all steps 401 to 404 above were followed, step 405 involves performing the measuring of the at least one colour coordinate value on the flattened exposed surface of the sample .

[0067] The principle of utilising one or more measured colour coordinate values of the product that has undergone a hemihydrolysis reaction and steam explosion does not as such limit the ways in which the process is controlled as a consequence . However, as the measured colour coordinate value is here used primarily as an indicator of the severity of the hemihydrolysis reaction, the control actions that are implemented on the basis thereof are most appropriately such that have a predictable effect on the severity of the reaction .

[0068] As one example , implementing said one or more control actions may comprise changing an acid concentration of the impregnating solution that is used to impregnate the wood particles before the hemihydrolysis reaction . As the basic rule , increasing the acid concentration increases the severity of the hemihydrolysis reaction while decreasing the acid concentration decreases it .

[0069] As another example , implementing said one or more control actions may comprise changing the residence time ( of the product that is being processed) in the impregnating solution . As a longer residence time in the impregnating solution tends to allow more acid to be absorbed in the wood chips , the length of the residence time in impregnation has an approximately similar direct dependence relation to the severity of the hydrolysis reaction as acid concentration .

[0070] As another example , the controlling of the value of at least one process parameter of the pretreatment part may comprise changing a storage time of the impregnated wood particles between the impregnating in the impregnation vessel and the input of said hemihydrolysis reactor . After removing from the impregnation vessel , the impregnated wood particles may not go directly to the hemihydrolysis reactor but to an intermediate silo or other kind of temporary storage . At least to a certain extent , the longer the impregnated wood particles are kept in such a temporary storage , the better the remaining impregnating solution penetrates into their inner parts , which has es sentially the same effect as a longer residence time in the impregnating solution .

[0071] The results obtained from the colour coordinate measurement may also be interpreted as being indicative of suboptimal quality and / or composition of the feedstock, meaning the raw material used in the process like wood particles before impregnating . A manufacturing plant of chemical bioproducts may obtain its feedstock from its own chipping or crushing facility, and / or it may acquire the wood particles or other biomass ready chipped, crushed, or otherwise prepared from one or more subcontractors . The plant may have stockpiles of different kinds of feedstock, and these may differ from each other in e . g . coarseness , plant species , storage age , moisture content , rot content , or the like . Some kind of a feeding mechanism i s used to feed a selected feedstock composition to the pretreatment part of the process . The controlling of the value of at least one process parameter of said pretreatment part may comprise changing the selected feedstock composition . For example , aged wood is known to behave differently in the hemihydrolysis reaction than fresh wood . I f the results obtained from the colour coordinate measurement can be interpreted as being indicative of the hemihydrolysis reaction exhibiting too dominant features typical to aged wood, one may change the selected feedstock composition to include a larger proportion of fresh wood .

[0072] As another example , implementing said one or more control actions may comprise changing the residence time ( of the product that is being processed) in the hemihydrolysis reaction . Again, an increase in the residence time in the hemihydrolysis reaction increases severity and vice versa .

[0073] As another example , implementing said one or more control actions may comprise changing the temperature and / or pressure to which the processed product is subj ected in the hemihydrolysis reaction . Temperature and pressure in the hemihydrolysis reaction are mutually related quantities , so that an increase in temperature typically means also an increase in pressure and vice versa . Again, an increase in either temperature or pressure or both increases severity and vice versa . Controlled quantities , such as the residence time in the hemihydrolysis reaction, typically have certain windows or ranges in which the value of such a quantity must remain under all circumstances . The dynamic state of the process may thus affect the causal relations between measured colour coordinate value ( s ) and the resulting control decision . For example , one may have selected the residence time in the hemihydrolysis reaction as a preferable altered process parameter so that the preferred action to decrease reaction severity in a detected need would be decreasing said residence time . However, if then a measured colour coordinate value would indicate too severe a reaction but the residence time in the hemihydrolysis reaction was already at the lower limit of its allowable range , the resulting control decision may be something else like decreasing acid concentration in impregnation, decreasing residence time in impregnation, or decreasing temperature and / or pressure in the hemihydrolysis reaction .

[0074] Above , for the sake of simplicity, the description has focused upon embodiments in which one measures the value of one colour coordinate ( such as the a* coordinate of the CIELAB colour coordinate system) and uses the measured value as a basis of making control decisions . In another embodiment , the measuring of at least one colour coordinate value may comprise measuring two or more colour coordinate values . Correspondingly, the applying of a decis ion rule may then comprise making said control decision at least partly on the basis of the measured two or more colour coordinate values .

[0075] I f two or more colour coordinate values are used as the basis of decisions , it is possible to use laboratory analyses to define ( and possibly update every now and then) target ranges for both or all of them . The making of control decisions may then involve making decisions intended to maintain each measured colour coordinate value in, or to direct each measured color coordinate value towards , its corresponding predetermined target range in subsequent similar measurements .

[0076] Additionally or alternatively, said applying of a decision rule may comprise making said control decisions on the basis of a further result calculated from the measured colour coordinate values . Calculations may be based on values of one kind ( like only a* values ) or values of several kinds ( such as a* , b* values ) . Such a calculated result may be for example some kind of a (weighted) mean value . Another example of a calculated result is a vector value where the two or more colour coordinate values are the components of the vector in a corresponding vector space . Similar to individual values , there may be target ranges for such calculated results so that the control decisions would then aim to maintain each calculated result in, or to direct each calculated result towards , its corresponding predetermined target range in subsequent similar measurements .

[0077] Especially if two or more measured colour coordinate values are used, and / or if one or more measured colour coordinate values do not constitute the sole basis for making control decisions , the dependencies between measured values and optimal further decisions may become very complicated . In such a case, the previously described applying of a decision rule may comprise using the measured at least one colour coordinate value as an input to an artificial-intelligence-based decision-making system and allowing said artificial-intelligence- based decision-making system to make said control decision . As time goes by and experience accumulates about how each decision affected the (multitude of ) measured value ( s ) , the artificial-intelligence-based decisionmaking system may apply the machine learning principle to optimise the ways in which it makes the decisions . As already hinted above , an artificial-intelligence- based decision-making system may use also other inputs than the measured at least one colour coordinate value to produce its control decisions . Such inputs may be indicative of further observed characteristics of the manufacturing process than j ust colour .

[0078] An apparatus aspect of the invention involves an apparatus for controlling a pretreatment part in a manufacturing process of chemical bioproducts . The apparatus comprises measurement means configured to measure at least one colour coordinate value of a product that has undergone a hemihydrolysis reaction and steam explosion as parts of said pretreatment part of said manufacturing process . The apparatus comprises also decision-making means configured to receive the measured at least one colour coordinate value and to apply a decision rule to make a control decis ion at least partly on the basi s of the at least one measured colour coordinate value . Further, the apparatus comprises control means configured to receive , as a result of said control decision, instructions to implement one or more control actions in one or more proces s steps of said pretreatment part in accordance with said control decision, and configured to act accordingly to implement said one or more control actions .

[0079] A use aspect of the invention involves the use of any of the method embodiments described above to control a manufacturing process of chemical bioproducts .

[0080] It is obvious to a person skil led in the art that with the advancement of technology, the basic idea of the invention may be implemented in various ways . The invention and its embodiments are thus not limited to the examples described above , instead they may vary within the scope of the claims .

[0081] EXAMPLE 1

[0082] To examine the feasibility of using a colour coordinate value as an indicator of reaction severity - and consequently as a valid basis for control decisions three series of measurements were made . In each series , the residence time in the hemihydrolysis reactor was consecutively changed to five different values . After each change , the pretreatment process was allowed to run under otherwise unchanged conditions for a period long enough to ensure that the ef fect of the newly selected residence time on reaction severity should be observable at the output of the hemihydrolysis reactor . A sample was taken of the material flow at the output of the hemihydrolysis reactor and prepared for a colour coordinate measurement as described above . Simultaneously, a parallel sample was taken to accurate laboratory measurements , where pulp was extracted and the liquid fraction analysed to measure chemical quantities like the amount of monomeric xylose, monomeric glucose , and total xylose ; Brix value ; amount of HMF (hydroxymethyl furfural ) ; and the like . These may be used to calculate quantities like the monomeric xylose ratio ( relative amount of monomeric xylose over total xylose after acid hydrolysis ) , HMF / xylose ratio ( relative amount of hydroxymethyl furfural over / xylose measured directly from filtrate ) , G2 factor ( relative amount of glucose over xylose measured with HPLC from acid hydrolysed liquid) , and G1 factor ( relative amount of glucose over xylose measured with HPLC directly from filtrate ) .

[0083] The a* colour coordinate value was measured from all 15 samples and plotted against the monomeric xylose ratio , HMF / xylose ratio , G2 factor, and G1 factor . The results are seen in figs . 5 , 6 , 7 , and 8 respectively . In all graphs , the five circles represent the first measurement series , the five triangles represent the second measurement series , and the five crosses represent the third measurement series . A polynomial fit was calculated for each measurement series respectively in figs . 6 , 7 , and 8 , with the dot line representing the first measurement series ( the circles ) , the dash line representing the second measurement series ( the triangles ) and the dash-dot line representing the third measurement series ( the crosses ) . Figs . 5 to 8 all consistently show that the a* colour coordinate value is a relatively unambiguous indicator of reaction severity, because an increase in the a* colour coordinate value unambiguously and monotonously represents a decrease in monomeric xylose ratio ( fig . 5 ) , HMF / xylose ratio ( fig . 6 ) , G2 factor ( fig . 7 ) , and G1 factor ( fig . 8 ) .

[0084] EXAMPLE 2

[0085] To examine the feasibility of using a colour coordinate value as an indicator of reaction severity - and consequently as a valid basis for control decisions - three series of measurements were made with respective three different combinations of temperature in the hemihydrolysis reactor and acid concentration of the impregnating liquid . The reactor was different than in example 1 above , having e . g . a different dry matter content of biomass in reactor . Residence time in reactor was changed to three dif ferent values at each combination of acid concentration and temperature level . After each change of acid concentration and temperature level , the pretreatment process was al lowed to run under otherwise unchanged conditions for a period long enough to ensure that the effect of the newly selected acid concentration and temperature level on reaction severity should be observable at the output of the hemihydrolysis reactor . A sample was taken of the material flow at the output of the hemihydrolysis reactor and prepared for a colour coordinate measurement as described above . Simultaneously, a parallel sample was taken to accurate laboratory extraction and measurements to measure its chemical properties from which were calculated the HMF / xylose ratio ( relative amount of hydroxymethyl furfural over xylose measured directly form filtrate ) , G2 factor ( relative amount of glucose over xylose measured with HPLC from acid hydrolysed liquid) , and G1 factor (relative amount of glucose over xylose measured with HPLC directly from filtrate) .

[0086] Figure 9 shows that there is clear correlation between the measured a* value and relative amount of HMF in pulp. The points plotted with round black dots are from a measurement series in which the acid dosing in impregnation liquid was 1.25 % w / w, temperature in the hemihydrolysis reactor was 185°C, and the residence time in the hemihydrolysis reactor was 5.0 min, 6.5 min, and 8.0 min respectively. The points plotted with X-marks are from a measurement series in which the three different residence times were the same as above but the acid dosing in the impregnation liquid was 1.05 % w / w and the temperature in the hemihydrolysis reactor was 191°C. The points plotted with black triangles are from a measurement series in which the three residence times were again the same but the acid dosing in the impregnation liquid was 0.83 % w / w and the temperature in the hemihydrolysis reactor was 197°C. In each case, an increase in residence time leads consistently to a lower a* value and a larger HMF / BRIX ratio.

[0087] Figure 10 shows same kind of correlation with b* -value. The measurement series are marked with the same symbols as in fig. 9. Similar to fig. 9, also in fig. 10 an increase in residence time leads to a consistent trend of a decreasing b* value and a larger HMF / BRIX ratio at each combination of acid concentration and temperature level.

[0088] In the three measurement series mentioned above, it was found that the tried combinations of temperature and acid concentration in impregnation did not give a completely clear correlation in all process combinations between the measured colour coordinate values of the pulp and glucose as such. However, in the temperature range from 185°C to 191°C the measured colour coordinate values correlated reasonably clearly with the glucose / Brix ratio (or glucose / xylose -ratio) , as is seen in f igs . 11 ( the a* value ) and 12 ( the b* value ) . At the highest temperature ( 197 ° C) the results deviated slightly from those of the lower temperatures , but still also at the highest temperature a reasonably good correlation was found between the measured colour coordinate values and the GB1 , i . e . glucose / Brix ratio ( or glucose / xylose -ratio ) . The use of the dot , X-mark, and triangle symbols is the same as in figs . 9 and 10 above .

[0089] EXAMPLE 3

[0090] An experiment was performed by operating a pretreatment part of a manufacturing process of chemical bioproducts in which wood chips were taken through impregnating, hemihydrolysis reaction, and steam explosion . Various control actions were implemented to intentionally vary the severity of the hemihydrolysis reaction . Five samples of the output product were taken at different times and a chemical analysis was performed on each of them to determine the combined severity of the hemihydrolysis reaction that each sample had undergone . The measured combined severity was in each case represented with a numerical value , which in this set of five samples ranged from 0 . 93 to 1 . 14 . In general , there are industry standards for methods used to describe the measured severity of the hemihydrolysis reaction with a numerical value .

[0091] A part of each sample was processed to standardi ze its measurement characteristics for a measurement of colour coordinates . The processing comprised preparing a flat surface and standardising a moisture content of the respective part of the sample . This was done by placing a predetermined amount of the sampled output produce in an open vessel , flattening an exposed surface thereof , applying heat to dry the sample , and al lowing the sample to cool .

[0092] A number of colour coordinate values were subsequently measured on the flattened exposed surface of each sample . More exactly, the measured colour coordinate values were values of the CIELAB L* , a* , and b* coordinates ; values of the R, G, and B coordinates of an RGB colour coordinate system; and values of the C, M, Y, and K coordinates of a CMYK colour coordinate system . Additionally, some further results were calculated from the measured coordinate values , like the sums a*+b* and L*+a*+b* .

[0093] The purpose of the measurements was to inves- tigate whether any correlation could be found between the measured colour coordinate values and the respective measured combined severity . Additionally, it was investigated whether any mutual correlation could be found between different measured colour coordinate values . Correlation between the calculated sums of colour coordinate values and the measured combined severity was also investigated . The measurements are shown in the following table . The rightmost column in the table shows a value indicative of the correlation between ( i ) the five colour coordinate values or coordinate value sums on that row in the table and ( ii ) the corresponding measured severity values . A negative value in the correlation column indicates inverse proportionality ( increase in colour coordinate ( sum) value corresponds to decrease in measured severity) while a positive value in the correlation column indicates direct proportionality ( increase in colour coordinate ( sum) value corresponds to increase in measured severity) . A correlation value with absolute value equal to 1 would mean perfect correlation whi le value 0 would mean no observable correlation .

[0094] Figs . 13 , 14 , and 15 show graphically the observed dependency of the CIELAB L* value as well as that of the G and B values of the RGB colour coordinate system on the measured severity . A can also be seen from the table , all these three colour coordinate values seem to correlate well with the measured severity . This , in turn, means that at least any of these three colour coordinate values could be used as a basis for making control decisions with reasonable chances of success .

[0095] Fig . 16 shows the mutual correlation between the measured CIELAB L* coordinate value and the measured K coordinate value of the CMYK coordinate system for each of said five samples . Fig . 17 shows the mutual correlation between the measured CIELAB a* coordinate value and the measured C coordinate value of the CMYK coordinate system for each of said five samples . The good mutual correlation within each set of two colour coordinate values suggests that in case it would only be possible to measure one such colour coordinate value , for example due to limitations in the available measurement hardware , the results could still be used as indications of also the other with reasonable certainty . Consequently, if there are decision rules based on e . g . measured C coordinate values of the CMYK coordinate system, one could apply suitably modified versions of such decision rules when using measured CIELAB a* values as inputs .

[0096] EXAMPLE 4

[0097] An experiment was performed by operating a pretreatment part of a manufacturing process of chemical bioproducts in which wood chips were taken through impregnating, hemihydrolysis reaction, and steam explosion. During the experiment, samples of the output product were taken at times 0:00, 2:00, 4:00, 10:00, 11:00, 16:00, 17:00, 18:00, 19:00, and 20:00. The values of process parameters were to be kept essentially constant, but there were unpredictable changes in the quality of the wood chips that were used as feedstock as well as uncertainties concerning the accuracy at which the various other process parameters could be controlled.

[0098] A part of each sample was processed to standardize its measurement characteristics for a measurement of colour coordinates. The processing comprised preparing a flat surface and standardising a moisture content of the respective part of the sample. This was done by placing a predetermined amount of the sampled output produce in an open vessel, flattening an exposed surface thereof, applying heat to dry the sample, and allowing the sample to cool.

[0099] A number of colour coordinate values were subsequently measured at five different locations on the flattened exposed surface of each sample. More exactly, the measured colour coordinate values were values of the CIELAB L*, a*, and b* coordinates. Additionally, the sums a*+b* were calculated. Of the five measurements, arithmetic mean values of L*, a*, b*, and a*+b* were calculated to obtain a corresponding representative value for each sampling moment. Fig. 18 shows the calculated mean a*+b* value as a black dot aligned vertically above the time of the respective sampling moment on the horizontal axis. The calculated mean values of the colour coordinate values and the sum a*+b* are shown in the following table.

[0100] Based on previous experiments, at least the following assumptions had been made for use as possible bases for decision rules:

[0101] - the value of the a* colour coordinate should be between 14.5 and 15.5, the ends included,

[0102] - the value of the b* colour coordinate should be between 20.0 and 22.0, the ends included, and

[0103] - the value of the sum a*+b* should be between 34.5 and 37.5, the ends included, to indicate acceptable severity of the hemihydrolysis reaction.

[0104] Said boundary values for the sum a*+b* are shown with horizontal dashed lines in fig. 18.

[0105] As seen in fig. 18, the mean value of the sum a*+b* was slightly below the desired range at 0:00. At this moment, the small deviation from the desired range was nevertheless deemed temporarily acceptable, as one wanted to stabilize the process. Additionally, at 0:00, the mean value of a* (14.2) was only slightly below the desired minimum (14.5) and the mean value of b* (19.8) was only slightly below the desired minimum (20.0) . Consequently, no changes were made to the values of process parameters. In particular, the setpoint value of the acid dosing in the impregnating fluid, used to impregnate the wood chips before taking them to the hemihydrolysis reactor, was kept at 18 grams per litre. The favourable trend in the sum a*+b* continued until, suddenly, in the sample taken at 16:00, the mean value of the sum a*+b* was found to have dropped to 34.1, which was clearly below the desired range. The mean value of a* was 13.9 and the mean value of b* was 20.1.

[0106] Based on previous experiments a decision rule had been formed according to which a detected drop in the sum a*+b* should lead to decreasing the acid dosing in the impregnating liquid at least in those cases where a simultaneous significant drop was observed in the value of the colour coordinate a* . Based on the measured (mean) colour coordinate (sum) value a*+b* and the measured (mean) colour coordinate value a* at 16:00, a control decision was made to decrease the setpoint of acid dosing in the impregnating liquid to 16.5 grams per litre. Already an hour later, in the sample taken at 17:00, the sum a*+b* was found to have a value 35.4, which was again in the desired range. Also, the value of the colour coordinate a* had recovered to 14.5. The subsequent measurements of the samples taken at 18:00, 19:00, and 20:00 showed that the favourable change persisted as both the sum a*+b* and the value of a* remained within their respective desired ranges.

Claims

CLAIMS1 . A method for controlling a pretreatment part in a manufacturing process of chemical bioproducts in which wood chips are taken through impregnating ( 121 ) , hemihydrolysis reaction ( 122 ) , and steam explosion ( 123 ) , the method comprising :- measuring at least one colour coordinate value of a product that has undergone the hemihydrolysis reaction and steam explosion as parts of said pretreatment part of said manufacturing process ,- applying a decision rule to make a control decision at least partly on the basis of the at least one measured colour coordinate value , and- implementing one or more control actions in one or more process steps of said pretreatment part in accordance with said control decision .2 . A method according to claim 1 , wherein said measuring of at least one colour coordinate value comprises measuring an a* value , a b* value , or an L* value , where said a* value , b* value , and L* value are coordinate values relative to the green-red opponent colours , the yellow-blue opponent colours , and the black-white opponent colours of an CIELAB colour space respectively .3 . A method according to claim 1 or 2 , wherein said measuring of at least one colour coordinate value comprises measuring an R value , a G value , or a B value , where said R, G, and B values are coordinate values of an RGB colour coordinate system indicative of relative amounts of red, green, and blue respectively .4 . A method according to any of claims 1 to 3 , wherein said measuring of at least one colour coordinate value comprises measuring a C value , an Mvalue , a Y value , or a K value , where said C, M, Y, and K values are coordinate values of a CMYK colour coordinate system indicative of relative amounts of cyan, magenta, yellow, and key colours respectively .5 . A method according to any of claims 2 to 4 , wherein said applying of a decision rule to make said control decision involves making a decision intended to maintain at least one said colour coordinate value in, or to direct said at least one colour coordinate value towards , a predetermined target range in subsequent similar measurements .6 . A method according to any of the preceding claims , comprising, in the following order :- taking a sample of said product that has undergone said hemihydrolysis reaction and steam explosion,- processing said sample to standardi ze its measurement characteristics for the measurement , and- performing said measuring of said at least one colour coordinate value .7 . A method according to claim 6 , wherein said processing of said sample comprises preparing a standardi zed surface , such as a flat surface , of said s amp 1 e .8 . A method according to any of claims 6 or7 , wherein said processing of said sample comprises standardi zing a moisture content of said sample .9 . A method according to any of claims 6 to8 , comprising, in the following order :- placing a predetermined amount of said product in an open vessel to form said sample ,- flattening an exposed surface of said sample ,- applying heat to said sample to dry said sample ,- allowing said sample to cool , and- performing said measuring of said at least onecolour coordinate value on the flattened exposed surface of said sample .10 . A method according to any of the preceding claims , wherein :- said impregnating ( 121 ) comprises impregnating wood particles in an acidic impregnating solution before subj ecting the impregnated wood particles to said hemihydrolysis reaction, and- said implementing of one or more control actions comprises at least one of : changing an acid concentration of the impregnating solution, changing a residence time of the wood particles in said impregnating solution, changing a storage time of the impregnated wood particles between said impregnating and said subj ecting of the impregnated wood particles to said hemihydrolysis reaction .11 . A method according to claim 10 , wherein said implementing of one or more control actions comprises changing a selected feedstock composition, wherein said feedstock comprises said wood particles before said impregnating .12 . A method according to any of the preceding claims , wherein said implementing of one or more control actions comprises at least one of : changing a residence time of the process stream in said hemihydrolysis reaction, changing a temperature to which the process stream is subj ected in said hemihydrolysis reaction, changing a pressure to which the process stream is subj ected in said hemihydrolysis reaction .13 . A method according to any of the preceding claims , wherein :- said measuring of at least one colour coordinate value comprises measuring two or more colour coordinate values ,- said applying of a decision rule comprises making said control decision at least partly on the basis of the measured two or more colour coordinate values .14 . A method according to claim 13 , wherein said applying of a decision rule comprises making said control decision on the basis of a further result calculated from the measured colour coordinate values .15 . A method according to claim 14 , wherein said applying of a decision rule to make said control decision involves making a decision intended to maintain said further result in, or to direct said further result towards , a predetermined target range in subsequent similar measurements .16 . A method according to claim 15 , wherein :- said measuring of at least one colour coordinate value comprises measuring an a* value and a b* value , where said a* value and b* value are coordinate values relative to the green-red opponent colours and the yellow-blue opponent colours of an CIELAB colour space respectively, and- said applying of a decision rule comprises making said control decision intended to maintain a sum of the measured a* and b* values in, or to direct said sum towards , a predetermined target range in subsequent similar measurements .17 . A method according to any of the preceding claims , wherein said applying of a decision rule comprises using said measured at least one colour coordinate value as an input to an artif icial-intelli- gence-based decision-making system and allowing said artificial-intelligence-based decision-making system to make said control decision .18 . A method according to claim 17 , wherein said artificial-intelligence-based decision-makingsystem uses also other inputs than said measured at least one colour coordinate value to produce said control decision, said other inputs being indicative of further observed characteristics of the manufacturing process .19 . An apparatus for controlling a pretreatment part in a manufacturing process of chemical bioproducts in which wood chips are taken through impregnating ( 121 ) , hemihydrolysis reaction ( 122 ) , and steam explosion ( 123 ) , the apparatus comprising :- measurement means configured to measure at least one colour coordinate value of a product that has undergone a hemihydrolysis reaction and steam explosion as parts of said pretreatment part of said manufacturing process ,- decision-making means configured to receive the measured at least one colour coordinate value and to apply a decision rule to make a control decision at least partly on the basis of the at least one measured colour coordinate value , and- control means configured to receive , as a result of said control decision, instructions to implement one or more control actions in one or more process steps of said pretreatment part in accordance with said control decision, and configured to act accordingly to implement said one or more control actions .20 . Use of a method according to any of claims 1 to 18 to control a manufacturing process of chemical bioproducts .

Images