Method and arrangement in a weighing system and a corresponding software product and material handling machine

Abstract

The invention relates to a method in a weighing system, in which method the mass of the bundle is weighed and recorded during both loading % and unloading m.sub.i_p of the bundle, during loading, the total loading mass m.sub.K_kok_j is calculated from the mass m.sub.i_c of one or more bundles weighed during loading and corrected using a correction factor C.sub.j, the total unloading mass m.sub.p_kok_j is calculated from the mass m.sub.i_p of one or more bundles weighed during unloading, with the aid of the said total loading mass m.sub.K_COk_j and total unloading mass m.sub.p_kok_j, a new corrected value Cj+1 is calculated for the correction factor C.sub.j in order to adjust the weighing for the loading of the next load K.sub.j+1. The invention also relates to a corresponding software product, an arrangement, and a material-handling machine.

Claims

1. A method in a weighing system of a crane, the weighing system having a weight sensor, computation unit and a memory, the method comprising: a loading step at a loading site, said loading step comprising: lifting a bundle from a plurality of bundles with the crane equipped with the weighing system into a load space, wherein the plurality of bundles form a load in the load space; weighing, with the weight sensor, and recording, with the computation unit to the memory, a mass of each bundle used to form the load; correcting, via the computation unit, the mass of each bundle weighed during loading in the memory using an initial correction factor; calculating, with the computation unit, a total loading mass of the load by summing the corrected masses of the plurality of bundles weighed during loading; moving the load from the loading site to a remote unloading site; an unloading step at the unloading site, said unloading step comprising: unloading the plurality of bundles of the load with the crane from the load space; weighing, with the weight sensor, and recording, with the computation unit to the memory, a mass of each bundle during unloading of the bundles of the load; calculating, with the computation unit, a total unloading mass of the load by summing the masses of a plurality of bundles of the load weighed during unloading; calculating, with the computation unit, a new correction factor (C.sub.j+1) based on a reference value determined through use of the total loading mass and the total unloading mass of the load according to the following equation: $C_{j+1}=\frac{C_j}{1-BF}$ wherein C.sub.j represents the initial correction factor, B represents the reference value, and F represents a filtering coefficient; and repeating the lifting, weighing and calculating steps above with the new correction factor for a subsequent load.

2. The method of claim 1, further comprising calculating a product of the mass of each bundle weighed during loading and the correction factor in order to give the corrected mass of the bundle.

3. The method of claim 1, wherein the reference value is calculated according to the following equation; $B=\frac{m_{P\_\mathrm{kok}\_j}-m_{K\_\mathrm{kok}\_j}}{m_{P\_\mathrm{kok}\_j}}$ wherein m.sub.P_kok_j represents the total unloading mass and m.sub.K_kok_j represents the total loading mass.

4. The method of claim 1, further comprising adjusting the weighing of loading for each load.

5. The method according to claim 1, further comprising adjusting the mass of each individual loading bundle with the aid of the correction factor.

6. The method according to claim 1, further comprising using the corrected correction factor, calculated with the aid of the load, to adjust the weighing of a subsequent load.

7. The method according to claim 1, wherein the filtering coefficient is based on adaptive filtering that adjusts the parameters on the basis of external information in the calculation of the correction factor.

8. The method of claim 1, wherein the filtering coefficient is a fixed percentage between about 50% and about 90%.

9. The method of claim 1, wherein the filtering coefficient is a fixed percentage between about 65% and about 75%.

10. The method of claim 1, where the filtering coefficient is based on a sliding mean value.

11. The method according to claim 1, further comprising calculating the correction factor condition-sensitively.

12. The method of claim 1, further comprising using a neutral initial correction factor to calculate the correction factor for loading of the load.

13. A software product comprising instructions stored on a non-transitory computer readable medium that perform the method of claim 1.

14. The method of claim 1, further comprising measuring acceleration relative to one or more axes using one or more acceleration sensors to obtain acceleration, where the step of calculating the corrected loading mass further includes correcting the recorded mass of a plurality of bundles weighed during loading in the memory using the correction factor and the acceleration data.

Description

(1) In the following, the invention is described in detail with reference to the accompanying drawings showing some embodiments of the invention, in which

(2) FIG. 1a shows a side view of a forwarder,

(3) FIG. 1b shows a schematic diagram of an as such conventional loader weigher with a central processing unit separate from the loader,

(4) FIG. 2 shows the stages of the method according to the invention,

(5) FIG. 3 shows in greater detail the stages of the method according to the invention for determining the correction factor,

(6) FIG. 4 shows the development of the correction factor and the corrected total loading mass of the method according to the invention in an example graph.

(7) FIG. 1a shows an, as such, known forwarder 10, which includes a loader 11 and a load space 12. FIG. 1b shows a schematic diagram of a loader weigher 13 with its accessories. Here, the forwarder is an example of a material-handling machine, in which there is an arrangement according to the invention. As an example of an embodiment a weighing system 14 is shown, in which the crane 15 is a loader and the weighing system includes a loader weigher 13. The computation unit 16 belonging to the totality is preferably located, for example, in the cab of the forwarder, and has in it a display device 17 for transmitting information and displaying to the operator of the weighing system (FIG. 1b). The display device and the computation unit and the other necessary data processing means can be as such known devices arranged to serve in forestry machine or other vehicle or work machine generally, or particularly arranged for a loader weigher in, for instance, a forestry machine. The computation unit has the necessary processing power to process the measurement data of the loader weigher and display it to the operator. The computation unit can also have the necessary storage capacity to record the data, but this can also be in a memory 28 separate from the computation unit.

(8) The loader weigher is formed of a suspension bracket 18, which is attached between the loader's 11 boom 19 and the rotator 20. In the embodiment of FIG. 1a, attached to the rotator 20 is a grapple 20, which the rotator 20 rotates. The loader weigher can also be installed in other vehicles or machines, which related to material handling and in which reliable weighing is required. The detailed construction of the loader weigher and the placing of the components is the loader can vary considerably within the scope of the present invention. For example, the force and possible acceleration sensors contained in the suspension bracket and referred to in the detailed description of the embodiment could possibly also be located elsewhere than between the end of the boom and the rotator, however in such a way that the desired force and/or weight data and, for example, acceleration data can be measured.

(9) The length of the suspension bracket is typically about 300 mm and its practical weighing range, for example, 70-2000 kg. In the suspension bracket 18 is a weight sensor 21, by means of which a bundle is primarily weighed. Functionally, the question is of a sensor reacting to force. Because the force is caused by a mass being weighed, the term weight sensor is used. The weight sensor can be based, for example, on a strain-gauge sensor, or alternatively, for instance, on a hydraulic operating device, preferably a combination of a hydraulic cylinder and a pressure sensor or transmitter. The suspension bracket also includes one or more acceleration sensors, measuring acceleration relative to one or more axes, such as in this case one two-axis acceleration sensor 22, which can be used to monitor the movements of the loader. In fact, simple loader weighers are known, which are intended to operate reliably only in static situations, in which weighers there is not the said acceleration or other feedback.

(10) Acceleration data can also be used to adjust the loader weigher, as the movement of the bundle naturally affects the weighing result. By means of the information obtained from the acceleration sensors, the mass of the bundle in motion can be corrected. It is then possible to use the whole lifting for weighing the bundle and thus obtain better accuracy. The acceleration sensors' detection axes are arranged at right angles to each other, so that by using two sensors comprehensive information is obtained on the movement and attitude of the suspension bracket. It is naturally also possible to use an arrangement measuring acceleration relative to three axes, in which, the sensors measuring acceleration parallel to each axis can be separate or integrated to form one totality. In the computation unit there are preferably also the necessary power inputs and data-transfer connections for operating different components. In this case, there is a CAN bus 23 between the suspension bracket 18 and the computation unit 16. In a known manner, the data transfer can, if desired, also be implemented completely wirelessly. This can be an advantage when data should be transferred in difficult conditions, such as in connection with the set of booms of a forestry machine, which is easily damaged by external obstacles.

(11) The weight sensor 21 is built into the suspension bracket 18 and can be used to measure forces in both the longitudinal and transverse directions of the suspension bracket 18. In static situations, when the forwarder is on a horizontal surface, the longitudinal direction is essentially parallel to gravity and an essentially straight tension acts on the weight sensor, assuming that the bundle has been gripped centrally relative to its centre of gravity. When the forwarder is on a sloping surface and/or with an unbalanced bundle, slanting forces, which can also be measured using the weight sensor, also act on the suspension bracket. A two-axis acceleration sensor 22 is located on an electronics card 24. At the upper end of the suspension bracket 18 is a hole 25 for a pin, by means of which the suspension bracket 18 is attached to the boom. Correspondingly, at the lower end is second hole 26 for a pin, by means of which the suspension bracket is attached to the rotator. The holes are at right angles to each other, so that it is possible for the grapple to swing in two directions. In other words, the holes are crosswise, in such a way that the upper pin permits movement parallel to the boom and the lower pin permits lateral movement. The same weight sensor can be used to measure the forces caused by an eccentric bundle. In this case, the lower hole is in the transverse direction referred to above. The foregoing describes one embodiment example of the loader weigher. However, the method itself is also suitable for other kinds of weighing system.

(12) FIG. 2 shows schematically the stages 30-52 of the method according to the invention. The method is intended to be used for adjusting a loader weigher, in order to improve the accuracy of the weighing of loading. When still using as an example a forwarder according to FIG. 1, the method is initiated with the loading of tree trunks into the load space of the forwarder in stage 30. The operator uses the grapple to collect a single tree or generally, in the case of energy timber, several thin trees at a time, and lift the trees into the load space. In this connection, reference to a single tree means the same as using the term log. Hereinafter, the name bundle will be used for the trees that are in the grapple at one time and are loaded into the load space. The mass m.sub.i of each bundle is weighed during lifting, in stage 32. In this connection, the subindex i refers to the sequence number of the bundle. In stage 34, the product of the mass m.sub.i of each bundle and the load-specific correction factor C.sub.j is preferably calculated in connection with the weighing of the bundle, which product takes into account the error in the weighing of every bundle lifted during loading. In this connection in turn the subindex j refers to the sequence number of the load. The corrected masses m.sub.i_c of the bundles multiplied by the correction factor are recorded in the memory in stage 36. Multiplication by the correction factor C.sub.j can also take place after the mass m.sub.i of the bundle has been stored in the memory. In stage 38, the corrected values m.sub.i_c of the masses m.sub.i of all the bundles of the loading are summed together to give the total loading mass m.sub.K_kok_j of the load, which is generally sought to be as close as possible to the optimal maximum weight of a single load K.sub.j of each forwarder. More specifically, m.sub.K_kok_j is m.sub.1C.sub.1+m.sub.2C.sub.1+m.sub.3C.sub.1+ . . . +m.sub.nC.sub.1. Summing preferably takes place in real time as loading progresses. In this connection, the term load refers to the total mass, consisting of one or preferably several bundles, in the load space of the forwarder.

(13) When loading is finished, i.e. when the forwarder has be loaded as closely as possible to its optimum load, a move can be made from the loading site to the unloading site, in stage 40. Moves from one loading site to another can also be made during loading, and the number of trunks in the forwarder can even be reduced, in which case the values of the masses of the eliminated trunks are deducted from the total loading mass m.sub.K_kok_j. The weighing system can distinguish between loading and unloading, for example on the basis of the use of the force and acceleration sensors and the crane and rotator, or generally by utilizing two or three data and combining them. In stage 42, after transfer unloading of the load K.sub.j is commenced, for example, to a stack. Preferably, the mass m.sub.i_P of every bundle unloaded from the load K.sub.j is weighed in stage 44 and the value recorded in the memory in stage 46. Because the unloading-direction weighing is very accurate, the weighed masses of unloading are recorded as such in the memory. The Measurement Act in force in Finland forbids the manipulation in any way of the unloading weighings, even to improve accuracy. The values of all the unloading bundles m.sub.i_P are summed in stage 48 to give the total unloading mass m.sub.K_kok_j of the load K.sub.j, which corresponds very well to the real mass of the loaded load.

(14) In stage 50, a reference value A, which depicts how accurate the weighing of the loading is, is formed preferably on the basis of the calculated total loading mass m.sub.K_kok_j and total unloading mass m.sub.P_kok_j. The reference value A is preferably relative, i.e it can be calculated, for example, using the following equation

(15) $A=\frac{m_{P\_\mathrm{kok}\_j}-m_{K\_\mathrm{kok}\_j}}{m_{P\_\mathrm{kok}\_j}}$
i.e. by subtracting the total loading mass m.sub.K_kok_j from the total unloading mass m.sub.P_kok_j and dividing this by the total unloading mass m.sub.P_kok_j. The reference value is preferably relative, as an absolute reference value, for example, the difference between the total masses of loading and unloading, is dependent on the size of the load. If the loads remain with always the same mass an absolute reference value can be used. In stage 52, on the basis of the reference value A the value C.sub.j+1 is calculated for the correction factor from C.sub.j, which replaces the value C.sub.j of the correction factor used in connection with the loading of the load K.sub.j in stage 54.

(16) FIG. 3 shows in greater detail the calculation of the correction factor, as a simplified flow diagram. In connection with the first load K.sub.1, or generally when some significant condition affecting the loading has changed, an initial correction factor C.sub.1 is used to correct the masses m.sub.i of the bundles of the loading. In stage 56, some initial value is chosen for the initial correction factor C.sub.1, which can be, for example, a neutral number such as 1 or 0, which keeps the masses m.sub.i of the weighed bundles as they are. According to one embodiment, the initial correction factor can also be some other number, for example 0.7, if it is ascertained that the masses of the bundles weighed in loading are always larger than the masses of the bundles weighed in connection with unloading. As in FIG. 2, the initial correction factor C.sub.1 is used for multiplying the masses m.sub.i of the bundles in loading the first load K.sub.1, in stage 34. After this, the corrected masses m.sub.i_c are stored in the memory in stage 36 and summed to form a total loading mass m.sub.K_kok_j in stage 38. After this, in stage 58 the relative difference value B is calculated with the following equation

(17) $B=\frac{m_{P\_\mathrm{kok}\_j}-m_{K\_\mathrm{kok}\_j}}{m_{P\_\mathrm{kok}\_j}}$

(18) In this example, the relative difference value B is used as the reference value A, which is shown in the embodiment of FIG. 2. In stage 60, using the relative difference value B, it is possible to calculate the corrected correction factor C.sub.j+1, which is obtained from the equation

(19) $C_{j+1}=\frac{C_j}{1-BF}$
in which F is the chosen filtering coefficient and X refers to a conventional multiplication and not a cross product. The corrected correction factor C.sub.j+1 replaces the previous correction factor C.sub.j in stage 62. The initial correction factor C.sub.1 is used to correct the loading of the bundles of the first load K.sub.1. The correction factor C.sub.j+1, corrected with the aid of the preceding load K.sub.j, is used for the subsequent loads K.sub.j+1. Thus, a corrected correction factor is calculated, which takes into account the error arising in the weighing of the loading of the previous load K.sub.j. As a result, the bundles are weighed more accurately in the loading of the next load K.sub.j+1. With the aid of the correction factor C.sub.j dependent on the reference factor A, it is possible to form an adjustment circuit for feedback to the correction factor C.sub.j, which will minimize the error in weighing due to the inaccuracy of loading entirely or nearly entirely within a few loads, or even immediately after a single load.

(20) In the method according to the invention, filtering is preferably used in the weighting of the reference value A, in order to calculate the corrected correction factor C.sub.j+1. Filtering is intended to reduce the effect of individual loadings on the development of the correction factor. In filtering, several different filtering alternatives can be used to determine the filtering coefficient. The filtering coefficient F can be, for example, a fixed percentage, 50-90%, preferably 65-75%, by which the effect of an individual loading on the corrector factor of the next load is reduced. The use of a filtering coefficient F of less than 50% is disadvantageous, as in that case the correction of the error arising in the weighing of loading by adjusting the weighing of the loading will be slow and require several loads to remove the error to a sufficient accuracy. The use of a filtering coefficient of one hundred percent will rapidly correct the weighing error in loading, but may in turn cause noise in the correction factor. Noise will arise, if the error of an individual load deviates, for one reason or another, from the other preceding loads, in which case the error will increase. The correction factor will then change radically according to the preceding load, even though the individual load was an exception. Thus, a large amount of noise will cause an error in determining the mass of an individual load in the loading direction.

(21) According to one embodiment, a sliding mean filter is used in the filtering of the correction factor. The correction factor can then be calculated on the basis of, for example, the previous load, after which the mean value is taken of the correction factors calculated on the basis of the ten previous loads. This mean value can be used as the correction factor of the next load.

(22) According to one embodiment, the filter used in the method can be a smart adaptive filter, the algorithms of which adjust the parameters of the filter automatically on the basis of, for example, changes taking place in loading, measurement, or measurement accuracy. By means of such an implementation, it is possible to detect at an early stage a change in level or a trend affecting the correction factor and caused by a change in conditions, when the reaction speed of the regulator, and thus also of the error correction can be accelerated for the duration of the change in conditions.

(23) FIG. 4 shows an example graph of the behaviour of the mass 76 of a bundle weighed in the loading of a load and of the correction factor 80, when a load with the same real mass of 10000 kg is loaded and unloaded ten times. The sequence number of the loads is shown by the reference number 74 on the horizontal axis. In this example, 1, which is a neutral number in the multiplication of the masses of the load, is selected as the initial correction factor C.sub.1. Next is shown one example of a way, in which the change in the correction factor can be filtered. 13000 kg is obtained as the total loading mass m.sub.K_kok_j of the bundles of the loading of the load K.sub.1. In the graph, the total loading masses m.sub.K_kok_j of the various loads K.sub.j are marked with the reference number 72 and the values of the correction factors C.sub.j with the reference number 70. Graph 82 shows the development of the total loading masses m.sub.K_kok_j between the loads K.sub.j and graph 78 in turn the development of the correction factor C.sub.j. Because the real mass of load K.sub.1 is 10000 kg, 10000 kg is also obtained as the total unloading mass m.sub.P_kok. On the basis of this, the relative difference value B can be calculated, which in this case is (1000013000)/10000, i.e. 0.3. From this in turn a corrected correction factor C.sub.2 can be calculated for the second load K.sub.2, which is 1/(1[0.3*0.75])=0.8163. When this is used to calculate the product of the corrected correction factor C.sub.2 and the mass m.sub.i of each bundle of the following loading K.sub.2, the value 13000*0.8163=10612.2 is obtained as the total loading mass m.sub.K_kok_j of the second load K.sub.2. In the example graph shown in FIG. 4, the same load is loaded and unloaded 10 times. The example is calculated using a more precise resolution than the four decimals shown in FIG. 4, which leads to a change in mass taking place in loads 8 and 9, though the approximate value of the correction factor remains unchanged.

(24) The filtering coefficient used in the previous example is 75%, so that after each load the error is reduced by 75%. Thus, the total loading mass m.sub.K_kok_j of the load rapidly approaches the total unloading mass m.sub.P_kok and already after five loads achieves an error level of less than one promil. At the same time, the value of the correction factor achieves a specific reading value. Because the error diminishes by 75% after each loading, 25% of the error associated with the previous load always remains to the next load. Thus, the error never disappears completely, but diminishes to become immeasurably small. When examining real loads, the loading conditions can vary considerably, so that multiplying by the correction factor before weighing the load can vary to some extent. However, with the aid of the correction factor C.sub.j, the total loading mass m.sub.K_kok_j is brought relatively close to the total unloading mass m.sub.P_kok_j.

(25) According to another embodiment, sliding mean-value filtering, differing from the filtering described in the paragraph above, is used for filtering. In this case, an individual reference value is not multiplied by the filtering coefficient, but rather a new corrected correction factor, calculated using 100-% filtering, is multiplied by a weighting factor. The unfiltered correction factor C.sub.j unfiltered, calculated on the basis of the most recent load, is calculated using the formula C.sub.j/(1B). A new filtered correction factor C.sub.j+1 is calculated using a mean value of the weighted correction factor, calculated on the basis of the previous correction factor and the most recent load, using the following equation (pk_previous*C.sub.j+pk_unfiltered*C.sub.j_unfiltered)/(pk_previous+pk_unfiltered), in which pk_previous and pk_unfiltered are weighting factors. The value 0.25 can then be used as the weighting factor pk_previous and the value 0.75 can be used as the weighting factor pk_unfiltered. The larger the relative weighting factor than is given to the correction factor calculated on the basis of the most recent load, the more rapidly the new correction factor will conform to changes in conditions. On the other hand, a very large relative weighting value of the correction factor of the last load may cause correction-factor noise, if, for some reason conditions vary significantly between loads. If the weighting factor of the correction factor of the most recent load is divided by the weighting factor of the previous correction factor, the ratio thus obtained can be, for example, in the range 0.1-10, preferably in the range 1-3.

(26) The weighing system preferably operates condition-sensitively. This means that the weighing system takes into account changes in the weighing conditions, i.e. for example, who the driver of the machine is and what timber grade is being loaded at the time. In addition, the weighing system preferably takes into account whether changes have occurred since the previous weighing in the machine being used for loading, i.e. in the lifting boom or grapple and what kind of timber is being loaded. For example, in loading energy timber the shape of the bundle being loaded can be very different to that when loading logs. It is then possible to use, as the initial correction factor of the loading, the most recent correction factor, calculated on the basis of the loadings made by the operator in question with the timber and goods grades and recorded in the weighing system. When conditions change, the calculation of the correction factor can always be started either from an initial correction factor, or from the most recent correction factor of the previous weighing.

(27) Though the loading mass is very quickly made accurate by means of the method according to the invention, the reliability of the loading weighing can be checked by using an as such known test weight as the bundle. The test weight can be, for example, a three-metre long steel pipe, which has been cast full of concrete or other material to achieve the desired mass. As such the size and mass of the test weight is of little significance, as long as the mass is known precisely and it can be assumed to be distributed evenly inside the test weight. The check weighing is then carried out only statically, so that the loader weigher is brought theoretically to the correct measurement range and, for example, errors due to device faults can be detected. Precision greater than the precision required of the weighing system is preferably used and the check weighing is also made when the test weight is tilting. A tilt is achieved by gripping the test weight eccentrically, in such a way that the test weight settles in a slanting attitude with one end of the test weight lower than the other. In other words, the suspension bracket's upper and lower pins permit the tilting of the bundle to an attitude corresponding to a state of equilibrium. In static check weighing, an accuracy of, for example, 2% is demanded. From time to time, for example, once a week, the values of the check weighing made are recorded and exploited in the method according to the invention.

(28) Using the weighing system according to the invention, the saving in time over check weighing according to the prior art is easily more than two hours a week, if previously the check weight was used once a day and check weighing took about half an hour. At the same time, the accuracy of the loader weigher when loading improves substantially.

(29) Using the weighing system according to the invention several different factors causing errors can be eliminated by comparing the differences between the loading and unloading of an individual load. Errors arise in the loading stage, in which there is often considerable variation relative to different variables. The load unloading stage, however, is often very constant and identical between different loads, so that the accuracy of unloading remains good. Generally, the accuracy required of unloading-direction weighing is in the order of 4%.

(30) Technically, the weighing system is able to determine the unloading direction without action by the operator, as sensor means monitoring the rotation angle are installed in the loader weigher, for example, in the loader or when installed in a loader or crane, for example in its rotation device. Thus, the sensoring of the loader weigher notifies the weighing system of the rotation angle at any time. On the other hand, the movements of the loader can be monitored sufficiently accurately for the purpose also without separate sensoring, if the operations of the loader are monitored, particularly the control of its rotation device during working.

(31) The arrangement according to the invention can also be used to monitor the operation of forwarders. The computation unit can record the total masses of all the individual loads in its memory, to that they can be examined later. This permits, for example, overloads to be monitored also afterwards, if the forwarder develops a fault.

(32) In this connection, the material-handling machine according to the invention can be any device whatever suitable for applying the idea of the invention, which can be used to weigh bundles both in loading and unloading. The material-handling machine can be, for example, a scrapyard crane or similar.