FI113696B - Procedure and arrangement for determining the weight of a mining vehicle - Google Patents

Description

113696!

Method and arrangement for determining the weight of a mine vehicle load

A method for determining the weight of a mine vehicle load, if the method determines the weight of the load on the basis of measurement signals obtained by separate measuring means.

The invention further relates to an arrangement for determining the weight of a load carried by a mining vehicle.

Mining machines, such as dumpers and bucket loaders, transport 10 blasted debris from the blasting site to the landfill site. Because of the high speed of this operation and the relatively short transport distances, the weighing must be carried out as the machine moves, so that weighing does not interfere with production. In turn, it is necessary for the Jat co-process to know how much quarry has been transferred to the downstream process. Real-time weighing data enables monitoring of material flows in the mine-15 already inside the mine, facilitating production control and planning. Planning preventive maintenance of machines is also possible by utilizing real-time weighing data.

In one known solution, the weighing is done by measuring the cylinder pressure caused by the load in the lifting cylinder which moves the system formed by the lifting arms 20 and the bucket or pallet. The pressure is measured on both sides of the lift cylinder several times over a given measuring period and t is averaged to calculate the load in the bucket. The effect of the machine tilt and the position of the lift arms or pallet on the differential pressure measured in the lift cylinder · is compensated by

MM

: 25. The calculation method is linear and the load is determined by

• I

· While moving. When calibrating the measuring system, the pressure is first measured with an empty bucket or pallet and then the known load is weighed on the bucket or pallet.

In stable steady state, the measurement values are obtained relatively well ·; ; 30 correctly and the load on the vehicle determined with sufficient accuracy. The problem with On-: is, however, that due to fast and short journeys, weighting has to be done while driving, where the tilt of the vehicle, bumps in the roadway and many other factors affect the weighing result, and in certain types of situations systematic error can easily occur. - • In the direction of 35 teas. Another problem with the solution is the linear method: application to a nonlinear system and the fact that the auxiliary metering set used is not sufficient to compensate for all errors in the pressure signal level due to runtime measurement. Another disadvantage is the use of a fixed measurement time to calculate an average pressure difference from measurement signals that oscillate with varying period lengths.

5 WO99 / 09379 discloses a method in which, using a neural network and a fuzzy system, the weight of a mine vehicle load can be determined on the basis of measurement signals measured by sensors. Measured quantities may include, for example, the cylindrical pressure of the bucket or platform lifting cylinders, the inclination of the vehicle both longitudinally and widthwise, and the position of the lifting arms or platform of the bucket 10. Based on the measured quantities and the dimensions and geometry of the bucket or pallet mechanics, the weight of the payload in the vehicle can be determined. The nonlinear model formed by the neural network and the fuzzy system results in better weighing results than the linear method described above, but the drawbacks of this method are machine calibration, the large amount of training data needed to determine the computation algorithm and the computation algorithm specific to the machine.

US 4,919,222 discloses a method and apparatus for determining the weight of a load vehicle. The determination of the load weight is based on the measurement of the cylinder pressure of the bucket lifting cylinders and the position of the bucket lifting arms during lifting of the bucket. Based on the cylinder pressure of the lifting cylinders and the position of the bucket lifting arms, a signal representing the weight of the load is determined from which the random variations of pressure:: are eliminated using curve fitting and averaging. ta. The resulting load weight curve is interpolated or: extrapolated to the curves determined during equipment calibration. ··· 'to determine the weight of the load in the bucket. However, a disadvantage of the method III of the publication is the dependence of the method on the bucket lifting speed, which must be taken into account in the method. In addition, when the track of the loading vehicle is particularly bulky, causing a considerable slope of the vehicle, · ·: 30 lists cannot be achieved with a sufficiently accurate weighing result.

Fl-94677 discloses deformation of structures. a method based on the measurement of toxins I t. · * ·. to measure the weight of the vehicle, in particular the load on the vehicle. The method is suitable for the calculation of * · · 'static loads practically stationary with respect to structures': 35 but cannot be used to calculate the load of a vehicle in motion.

It is an object of the present invention to provide a novel method and arrangement for weighing a mine vehicle, by which method and arrangement can be carried out with sufficient accuracy even when the vehicle is in motion.

The method according to the invention is characterized in that a non-linear spatial pattern is formed, which spatially determines the weight of the mining vehicle platform or bucket and lifting arms, load weight and that the mine vehicle load weight is determined from the measurement signals measured by the measuring means by estimating the states of the non-linear state model with a non-linear Kalman filter.

Further, the arrangement according to the invention is characterized in that the arrangement includes a calculating element comprising a nonlinear state model, | which state model is adapted to describe a weighing system formed by a mining vehicle platform or bucket and lifting arms and lifting cylinders and / or tilting cylinders used to move them, and measuring means adapted to the mining vehicle, and which state model is adapted • mining vehicle load weight based on the measurement • YT miss based on the measured measurement signals by estimating nonlinear t:: ': I borrowed spaces with a nonlinear Kalman filter.

It is an essential idea of the invention that a mining vehicle load: 25 weight is determined using a nonlinear Kalman filter which estimates the weight of the load in the vehicle, which is not directly measurable by measurement signals from the vehicle.

An advantage of the invention is that by using a nonlinear Kalman-t · • · ·: 30 filter, a more accurate estimate of the vehicle load weight can be obtained, since a nonlinear method is used to solve the nonlinear problem that also effects the noise included in the measurements. · ·. optimized load weight can be minimized. It also has the advantage that the method is simpler to calibrate and that the method does not need to be specifically taught to recognize different masses. Further, the determination of the load weight 4113696 is faster and more accurate than in prior art methods.

The invention will be explained in more detail in the accompanying drawings, in which Fig. 1 schematically shows a dump truck for mines to which the method of the invention has been applied; and an apparatus which may be used, for example, to determine the load weight of the dump truck according to FIG. 1, and FIG. 4 schematically illustrates the operating principle of a non-linear Kalman filter.

Fig. 1 is a schematic illustration of a dumper with a wheel movable body 1 and a pallet 3 mounted at its rear end by hinges 2 with pivots 2. Lifting cylinders 4 are connected between it and body 1 and lowered from its front end to the brackets 5 when lowered. . Furthermore, the dumper has gravity sensors 6 which measure the inclination of the body 1 with respect to the horizontal both in the longitudinal and the width direction of the dumper. The inclination of the pallet 3 relative to the frame 1 can be measured, for example, by using angle sensors connected to the joints 2 or by measuring the volume of the pressurized fluid supplied to the lifting cylinders 4 and calculating the inclination angle of the pallet 3 by geometry between the anchors.

·: · Figure 2 is a schematic representation of a bucket loader with. : 25 wheeled frame 1 and articulated linkage 7 via articulated links 8, · * ·! ie a bucket 9 which pivots about the link arms 7 about the joints 10. In order to furnish the bucket 9 with respect to the lift arms 7, it has a separate tilt cylinder 11 and for lifting the bucket 9 there is a lift cylinder 4 between the lift arms 7 and the frame 1. Further, the bucket loader has gravity * * »··; 30 based on the inclination measuring transducers 6, which measure the inclination of the bucket loader in relation to the horizontal, based on the gravitational pull of the bucket loader in both length and width. The position of the bucket 9 in the height of the frame 1. '* *. for example, in the case of the joints 8, angle sensors can be used to measure the geometry of the lift arms 7 it * * 35 on the basis of the measurement data transmitted by using the lifting height of the bucket 9 when rotated by its pivoting cylinder 11. Alternatively, the lifting height can also be determined by measuring the volume of pressure fluid supplied to cylinder 4, whereby the lifting height can be calculated based on that volume and the length of the joints and cylinder 4.

Figure 3 schematically shows an apparatus utilizing a non-linear Kalman filter for determining the load carried by the 5 dumpers of Figure 1, which can be used to measure the dump load while the vehicle is in motion or stationary. bucket filling. For measuring the arms 10, measuring sensors or measuring devices are used, of which two measuring sensors 2a and 2b are, for example, strain gauge sensors, which are mounted at a suitable position on the two joints 2 of the platform 3 on the two edges of the dump body. Further, the apparatus includes sensors 4a and 4b for measuring the pressurized fluid pressures of the lifting cylinders 4 both on the side of the lifting cylinders 4 where the pressurized fluid is fed and on the side where the pressurized fluid flows out. In principle, these sensors enable the weight of the load to be determined with sufficient accuracy on a horizontal surface in a static situation.

The measurement signals from the strain gauge sensors 2a and 2b are applied via amplifiers 12 to the lowering position calculator 13 of the pallet 3, which 20 positions of the pallet 3 are defined as previously described, and from which the parameter describing the position of the pallet 3 ·· at the input of block 14 implementing the filter. The calculation of the stage 3 position can also be included as part of the Kalman algorithm itself. Further, in the same block of the non-··· linear Kalman filter 14, the measurement signals from the pressure sensors 25 4a and 4b are supplied from the temperature sensor of the pressure fluid in the cylinder 4. ···. 4c the incoming temperature as well as the vehicle inclination measured by the tilt sensors 6, »which measurement signals are fed to block 14 implementing the nonlinear Kalman filter. Block 14 may be a microprocessor, signal processor or other similar calculating element capable of performing pre-programmed functions.

During weighing, the operator lifts the pallet 3 with the dumper in place or in motion so that it is released from the supports 5 shown in Figure 1. In this case, the pilot light illuminates to indicate that the pallet 3 is merely supported by the cylinders 4 and the joints 2. The driver then presses the weighing button after a period of 35 years. Weighing can begin: also automatically after a certain time after lifting the pallet. Based on the inclination angle of the pallet 3 calculated in the calculating element 13 and the measured cylinder pressures, pressure fluid temperature and vehicle inclination, the vehicle load weight is estimated in block 14 implementing the nonlinear Kalman filter. FIG. 3 also shows a memory unit 15 and other values measured, calculated, or estimated during load weight estimation. The memory unit 15 also stores the initial values required to start the estimation process, which are presented in connection with the description of the operation of the non-linear Kalman filter of Figure 4, which can be read from the memory unit 15 into the the calculating means 14, but for the sake of clarity the memory unit 15 is shown in Figure 3 as a separate component.

Figure 4 illustrates the operation of a non-linear Kalman filter used for estimating the weight of a load 15 to be weighed in principle. The model of the weighing system comprising the mining vehicle platform 3 or bucket 9 and the lift arms 7 and the lifting cylinders 4 and / or the tilting cylinder 11 as well as the measuring instruments described above is dynamic, non-linear and discrete. The dynamics of the system can be described by 20 equations x (£ +1) = f [&, x (£)] + v (&), (1) where x (£ + l) is the actual state of the system at time £ + 1, f () is a nonlinear function corresponding to the state transition matrix of the system, x (fc) is the real system - *: · min state at the earlier time k and vector v (fc) is white, zero-mean process noise representing the real system and systemic form. "; the modeling error between a given model and the expected value :: * £ [v (£)] = o and the variance E v (£) v (/) r = Q (30 where Q (&) is the covariance matrix of process noise or model noise is Kronecker's delta, where SkJ = number k = j and otherwise 0, and T represents the matrix; * transposition operation For example, in the system model, for example, the load weight m of the dumper can take into account the top and bottom pressures py and pa, platform tilt γ, the pressure L. 35, for example hydraulic oil, temperature L. In this case, the non-linear state model of the vehicle weighing system 7 113696, i.e. the model, would consist of six elements x = m, py, pa, y, s, L].

Of these, all except the load weight m itself are measurable quantities. From the above measurements, the measurement of the pressure liquid temperature L can also be omitted without a significant change in the accuracy of the load weight estimate m. The dependence of the load weight m on said measurements is non-linear, i.e. the function f () describing the dynamics of the weighing system shown in equation (1) is non-linear. In addition, the function f () of the load weight m model 10 may take into account other factors that are not directly measurable.

The estimation of the system state and thus the load weight using the nonlinear Kalman filter is performed as follows.

At moment k, the true state of the system is x (£) 20. The actual state 21 of 15 at the next time point k +1 according to formula (1) is x {k +1) = f [/ c, x (ä :)] + \ (k) , corresponding to the measurement at 22 at k +1 is z (k +1) = h [& +1, x {k +1)] + w (k +1), (2) where the measurement function h () is a generally non-linear function , but within the scope of the present invention, the measurement function h () can also be linear, and w (fc) is -i white, zero-mean measurement noise, which describes an error summing up between the measuring devices and the measuring environment. Measurement noise w (&)>; · to the expected value £ [w (£)] = 0 * ·, ··· [25 and variance. · · ·. £ [wU) w (/) 7] = R (ä:) <% where R (&) is the covariance matrix of the measurement noise.

Λ - ·· The estimate x (£ | fc) of the real state x (k) 20 at time 23 is an approximation of the conditional expectation value of the state: '' ',: *. '. 30 x {k \ k) «£ | x (ä:) | Z * j, formed by the measurements of Z * = {z (l), z (2), ..., z (fc) accumulated up to k }. To allow system space at time k +1. to estimate, the system nonlinearities must be linearized from the model dynamics function f () in the vicinity of the state estimate of time k for time k (k \ k) 23 at 1111696. Linearisation uses Taylor series development and depending on whether the series development uses only first-order terms or whether second-order terms are included, either a first-order or a second-order filter is obtained. Nonlinear function linearization is also used in the calculation of the measurement forecast z (k + \\ k) 25. Taylor's series development gives a second-order filter of the form x (& +) = fk, x {k \ k) + ί * (£) x (k) -x (kjk) 1 », f Λ ~ iT r Λ η ' (3) + - Σ no x (k) ~ XW *) / «(*) x (k) - x (k \ k) + ΚΑΤ + wik)

^, = ι L J L

where ηχ is the number of states, which in this example is six, e, is the i th nx dimensional base vector, with the i th component being the number one and the other 10 components being zero, KAT describes the higher order terms that can be ignored in this case excluding and * x (k) = [v, f (*, x) 7 '] 7' | x = (4) is the Jacobian 29 of the vector f at x (k \ k) 23 and / »(*) = [ νΧ / '(Μ)] χ = χ (* Ι *) (5) 15 is the part 29 of the Hessian matrix calculated from the f / th component of the vector.

; : After linearizing, state prediction x (k +1 | &) 24 x (& - f lj &) = e \ f &, x (fc | &) i + iijfj (fc) x (fc) - x (# | ä) i:; i 'il », r Λ rr λ ij (8): * ·.: + Σei x (k) - x (k \ k) fL x (k) - x (k \ k) | • · · "* 20 at time k for time point £ + 1 is given by the conditional expectation value of equation (3) formed by Z · base · · of the measurements accumulated up to time k, with the exception of terms higher than a degree however, the accuracy of the calculation can be increased by taking into account terms that are higher than the second degree Since the term of the first. · .25 degree is an average of zero based on x (yfc | £) «£ [x (ife) | z4], ...: get the state prediction of χ (& + ί | &) for 24 moments k +1 113696 9 x (fc + l | 4 = f &, χ (* | έ) + \ ^ eitr \ f ^ x (k) v {k \ k) \ 1 (7) LJ 2, = 1 where tr is the sum of the diagonal elements of the square matrix and P (£ | &) 28 is the covariance of the state at k. The state prediction error is obtained by subtracting equation (3). (7) By multiplying the prediction error thus obtained by its own transposal equilibrium and taking its conditional expectation value for the measurements Z *, state covariance? (k + \\ k) 30 p (* + i | *) = f, (*) P (* l *) f, (*) r 1 nx r 1 (8) + 9 Σ Σ eie] ^ fL {k) V (k \ k) fi {k) v {l ^ + Q (4 Z / = 17 7 = 1

Based on the state prediction from equation (7), we can compute the prediction z (k +1 | £) for k at time k for 25 times k +1 10 z {k + \\ k) = h & + 1, χ (& + 1 | £) + “^ Βίί ** [^ (Α:) 4 · Ρ (Α: + ΐ (Λ) |, (9) LJ 2 / = 1 where is not the i th nz dimensional base vector and in this example n2 is Five, that is, the number of measurements, based on the actual measurement z (k + \) 22 and the measurement prediction z (k +1 | £) 25, we can calculate the measurement residual or innovation u {k +1) 26 at k +1 15 v (k +1) = z (k +1) - z {k + 4), (10) and the associated innovation covariance S (& + l) 31 is% 'S (k + 4) = h, (£ + l) p (fc + l \ k) hx (k + l) r. * 1 »z» z (11).:. + T Σ Σ tr [h'xx (k + 0P (* + l \ k) hi (k + 0P (* + P)] + R (4.: ^ I = i 7 = 1 *: where: ) - (5) respectively '·· * h ^ (^: + l) = + χ) Γ x = x (& + lj £) (12) »» 20 and t;> +1) = + l, x)] jx = i (* + 1 | A). (13): The filter gain W (k +1) 32 can be calculated from formula V. W (it + l) = i: [x (Ä: + l) u (Ä: + l) r | Zi], (14) where x (£ + l) is the prediction error for state x (£ + l) 21 based on information available at time 25 ί. Updated estimate of the state, ie the state of the filter. the set value x (k +1 | k +1) 27 at k +1 based on information available at k +1 is 10 113696 x (£ + l | Jt +1) = i (jfc + l | Jt) + W (it + l) u (Jfc +1) (15) and the updated covariance P of the state P (& + \\ k +1) 33 at k +1 based on information available at k +1 is P (* +1 | Jfc +1) = P (jb +1 | Jk) - W (Jfc + l) S (Jk +1) W (Jk + l) r. (16) 5 The estimation of the state of the system, i.e., according to the present invention, also the weight of the load m, with a Kalman filter can in principle be divided into three parts; state prediction, calculation of filter gain and measurement residual to calculate an estimate of the system condition, and in this case especially the load weight m. The uncertainties of the model and measuring equipment of the weighing-10 influence the filter cavity through the state covariance, that the state estimate update takes into account, in an appropriate proportion, the information provided by the measurements about the state of the system and the state calculated from the model of the system, since neither of them alone is completely reliable, that is, ιοί 5. The updated values obtained are still used for the next time point estimate. These calculation rounds are repeated until the space provided by the filter at the outlet, i.e. in this case especially the load m of the vehicle, has reached a certain level, which thus corresponds to the estimate of the load m of the vehicle. The estimation 20 can be terminated, for example, when the variance of the load weight estimate falls below a predetermined threshold value, which limit can be changed, i.e., it is a parameter of the algorithm. The initial state estimate ί (θ | θ), the state covariance Ρ (θ | θ) corresponding to the initial state, and the uncertainty of the weighing system model and measuring equipment, all of which are stored, are required to begin the calculation. ·: 25 to the memory units 15, from which they are read to the non-linear Kalman filter implementing block 14 upon initiation of weighing. The factory default values for this vehicle can be used as starting values. The first measurement can also be used as the initial values of the states to be measured, whereby the actual calculation is started from the second measurement. That's why the weight of the load is estimated. The value of 30 m at the very first calculation round of the Kalman filter does not give a • correct result because the calculation starts from the initial value of the state, which is not necessarily correct. In addition, the measurement signals, especially at the beginning of the measurement, also have various interferences which the Kalman filter must first filter: off.

: 35 To calibrate the weighing system, the vehicle shall be loaded with a known test load. 11 113696 For Multiple Vehicles Weighing an Empty Bucket and Weighing With One Known TEST LOAD, but also several TEST LOADS of different weights can be used. Calibration is done on a Machine-by-Machine basis. Further, the Calibration may be renewed during the Operation of the Machine to compensate for the effects of changes in Machine rotation or changes in Components. Non-linear Kalman filtering can also be used during calibration. Also used for estimating non-linear model parameters of the weighing system.

Similarly, as described above, weighing can be done with the Bucket-10 Loader, whereby the position of the Bucket and other issues can be easily taken into account. In the case of a bucket loader, the measurement diagram shown in Figure 3 can in principle be used, whereby the position of the bucket 9 in the height direction of the frame 1 and / or its slope arms 7 is taken into account in the design of the weighing system. Thus, the model space factor x and the system dynamics k-15 kaa are represented by the functions described above, while the principle of load weight estimation remains the same.

The drawings and the description related thereto are intended only to illustrate the idea of the invention. DETAILED DESCRIPTION The disclosure may vary within the scope of the claims. Thus, the structure of the mining vehicle need not be exactly like that shown in Figures 1 and 2, but it is essential that the load weight estimation is based on the estimation of states of a nonlinear model formed by a weighing system with a nonlinear Kalman filter. Also, different removable sections of the Kalman filter, such as the Wiener filter or other similar methods, can be used in a similar manner to the Mining Vehicle Load Weight.

* »· • ·» · • • •% V I ·

I I

• »»

Claims (19)

A method for defining the weight of the load in a mining vehicle, characterized in that a nonlinear condition model is formed, which condition model describes a weighing system formed by the mine vehicle (3) or bucket (9) and lifting arms (7) and lifting cylinders (4) and / or a gradient cylinder (11) to be used for moving them and measuring devices provided in connection with the mining vehicle, and in which state of the state at least one condition describes the weight (m) of the load in the mining vehicle to be defined and that the weight ( m) the load in the mining vehicle is defined on the basis of measurement signals measured with the measuring instruments by estimating the condition of the nonlinear state model with a nonlinear Kalman filter. Method according to claim 1, characterized in that at least one condition in the non-linear model of the weighing system comprises the pressure (py, pa) in the pressure fluid of the lifting cylinder (4). Method according to claim 1 or 2, characterized in that at least one condition in the non-linear model of the weighing system comprises the temperature (L) of the pressurized fluid in the lifting cylinder (4). Method according to claim 3, characterized in that the at least one condition in the non-linear model of the weighing system comprises that of the mining vehicle. ·,: Slope (f) in relation to the horizontal plane. • * f Method according to claim 4, characterized in that at least one condition in the non-linear model of the weighing system comprises the slope of the load surface (3) in relation to the position of the mining vehicle or the bucket (9) of the mining vehicle and / or the lifting arms (7). ) slope in the chassis of the mining vehicle (1). 6. A method according to any of the preceding claims, characterized in that, as values for the initial state of the non-linear Kalman filter, preset values are used. Method according to claim 6, characterized in that as the values for the initial state of the nonlinear Kalman filter, factory-set values are used. _ 8. A method according to any of the preceding claims, characterized in that the estimation of the weight of the load is completed when the estimate of: · ·: the weight of the load has stabilized at a certain level. I »*» • »113696 9. A method according to claim 8, characterized in that the estimation of the weight of the load is terminated when the value of the variance for the estimation of the weight of the load falls below a pre-set limit value. 10. A method according to any of the preceding claims, characterized by the non-linear model of the weighing system! is calibrated by using one or more test loads of known weight. 11. Arrangement for defining the weight of the load that a mining vehicle carries, characterized in that the arrangement comprises a calculator (14) comprising a non-linear state model, the state model being adapted to describe a weighing system formed by the surface of the mining vehicle (3). ) or bucket (9) and lifting arms (7) and lifting cylinders (4) and / or an inclined cylinder (11) to be used for moving them and measuring devices arranged in connection with the mining vehicle, and in which state of the model a condition is arranged to describe the weight (m) of the load in the mining vehicle 15 to be defined and that the calculator (14) is arranged to define the weight (m) of the load in the mining vehicle on the basis of measurement signals measured with the measuring instruments by estimating the non-linear state of the state model with a nonlinear Kalman filter. Arrangement according to claim 11, characterized in that the arrangement comprises measuring means (4a, 4b) for measuring the pressure (py, pa) in the pressure fluid of the lifting cylinder (4). ,! Arrangement according to claim 12, characterized in that the arrangement comprises measuring means (4c) for measuring the temperature (L) of the pressure fluid in the lifting cylinder (4). ,. : 25 Arrangement according to claim 13, characterized in that, '. The arrangement comprises measuring means (6) for measuring the slope (?) of the mining vehicle in »· ;;; relating to the horizontal plane. Arrangement according to claim 14, characterized in that the arrangement comprises measuring means for measuring the slope of the flake (3) in relation to the position of the mining vehicle or the bucket (9) of the mining vehicle and / or the slope of the lift arms (7) in the chassis of the mining vehicle (1). ) elevation. ; V. 16. Arrangement according to any of claims 11-15, characterized by •. The arrangement comprises the memory unit (15) for storing the estimated state of the state model and / or the initial values required at estimation. 112696 Arrangement according to claim 16, characterized in that the calculator (14) comprises a memory unit (15). Arrangement according to any of claims 15 - 17, characterized in that the arrangement comprises a calculator (13) for defining the position of the mining vehicle (3) or the position of the bucket (9) and / or the slope of the lift arms (7). Arrangement according to claim 18, characterized in that the calculator (14), which with the non-linear Kalman filter estimates the state of the nonlinear state in the weighing system of the mining vehicle, comprises a calculator (13) which defines the position of the mining vehicle's flange (3). ) or bucket (9) and / or the slope of the lift arms (7). * · • »·» »» · * · ♦ • 1 · * · * 1 ♦ • · • · »· * 1 I t» · • »t