EP1113102B1 - Method for determining total inertia and unbalanced load in a laundry drum of a washing machine - Google Patents

Description

• The present invention relates to a method for determining the total load and unbalanced load in a washing machine having a motor drive system for rotating the motor at a desired target speed. With the term washing machine we mean all kinds of appliances having a rotating drum whose speed can reach at least a value at which the laundry is applied by centrifugal force to the side wall of the drum. Washers/dryers are included in the above general term.
  The technical problem of determining the amount of laundry loaded in a drum of a washing machine is well known in the art. Known technical solutions are based on amount of water absorbed by dry laundry, time needed for reaching a predetermined target speed of the drum or current absorbed by the motor, EP-A-0143685 discloses a method for determining the mass of laundry from a measured value of the motor torque for driving the drum during an acceleration phase at constant acceleration. This document does not teach how to assess whether and to what extent the load is unbalanced (i.e. not uniformly distributed on the sidewall of the drum).
  EP-A-71308 discloses a method for assessing whether there is an unbalanced distribution of the laundry articles around the drum, by monitoring with a tachogenerator the actual motor speed of the washing machine (if the motor speed changes by more than a predetermined amount, this means that the laundry is unbalanced).
  Both the above methods do not solve the problem of detecting load and unbalanced load, and a machine which is simply using the above two methods in combination would be too expensive since it would need an electronic device for detecting motor torque and an electronic device for detecting motor speed variation.
  In US-A-5507054 a method is disclosed which determines the load by shutting off the motor drive system and by measuring the time needed for reaching a predetermined lower speed. This document discloses also an assessment of unbalanced load by detecting fluctuation of the drum speed, according to the above EP-A-71308. Such assessments are carried out subsequently, therefore the duration of the washing cycle is increased. Moreover the different physical entities measured by the electronic circuit (fluctuation of speed, time for reaching a predetermined speed) render this latter quite expensive and complex, decreasing its reliability.
  It is an object of the present invention to avoid the above mentioned disadvantages, by providing a method and a control system that is cheap, reliable and that can carry out both assessment in a very short time.
  Also document EP 0 903 845 discloses a method and a washing machine according to the preamble of claims 1 and 8, respectively.
  According to a general aspect of the invention, by providing a known signal having a predetermined frequency pattern or spectrum added to the normal periodic control signal at constant speed (for maintaining the drum at the target speed), the system motor - drum - laundry is excited so that by analysing the speed feedback signal (for instance by means of its frequency spectrum analysis) it is possible to assess the total inertia (and therefore mass of laundry) and at the same time out of balance condition of the system. According to one aspect of the invention there is provided a motor drive system that is supplied with an auxiliary signal having a predetermined frequency spectrum, the load and unbalanced load being determined by means of a frequency spectrum analysis at different frequencies.
  Preferably, the auxiliary signal is a periodic signal having a frequency different from the frequency corresponding to the target speed of the motor. More preferably, the unbalanced load and the total load are determined by means of frequency spectrum analysis at a frequency corresponding to the target speed and at a frequency corresponding to said auxiliary signal respectively.
  The principal advantage of the method according to the present invention is that the detection of any out of balance condition of the drum (and the degree thereof) and of the amount of laundry in the drum are both obtained in a very short time by "injecting" in the motor drive system the above auxiliary signal and by using the same frequency spectrum analysis at two different frequencies.
  Another advantage of the present invention is that it provides a method for compensating the load friction so that the load inertia/mass and spinning speed selection, based on moment of inertia load and unbalanced load, result more accurate than with known methods.
  The present method can be used for determining the inertia /mass of the dry load or wet load indifferently. The present method can be used for determining the inertia/mass load and unbalanced load amount aiding the spinning speed selection.
  The dynamic friction is caused by different components, for instance transmission system friction, ball bearings friction and, for horizontal axis washing machines, friction of the laundry against the front door. According to the present invention, using look up table data or Neuro Fuzzy Logic system can compensate data deriving from frequency spectrum analysis. Such look up table data or Neuro Fuzzy Logic system are determined for each kind of washing machine as result of experimental tests carried out on the washing machine with predetermined values of overall friction. From the compensated data of frequency spectrum analysis, the compensated value of total load and unbalanced load can be determined, such values being not affected by dynamic friction.
  An embodiment of the invention will now be described with reference to the accompanying drawings, in which:
• Figure 1 is a schematic mechanical model of a washing machine;
• Figure 2 is a block schematic circuit of the control system used according to the present invention;
• Figures 3a, 3b, 3c are diagrams of system frequency responses used in the method according to the present invention;
• Figures 4a and 4b are portions of a software flow chart control used in the method according to the present invention;
• Figures 5a and 5b are schematic diagrams which show the correlation between frequency responses, compensated frequency responses and dynamic friction; and
• Figure 6 is an example of a fuzzy logic control for the unbalance and inertia detection system according to the present invention.
DETAILED DESCRIPTION OF A PREFERRED EMBODIMENT
• The method according to the invention is used to determine the inertia/mass and unbalance load clothes in a rotating drum for washing machine and dryer machine as well as washer/dryer machine, and it can compensate, if desired, the system dynamic friction of the rotating drum system.
  Referring to Figure 1 a mechanical washer/dryer representation is shown.
  The motor M provides the torque T_m to the drum; motor speed rotation is transmitted to the drum via a motion transmission system (not shown) and the speed is maintained/adjusted by a speed controller system of known type (not shown). Hereafter we will refer to the simple system shown in Figure 1, without taking into account the specific motor controller. According to the second law of dynamics, applied to the shaft motor system, it can be said that the provided total torque equals the sum of the inertia torque and the friction torque. T = J * dω dt + γ * ω where:
• J is total inertial applied to the shaft motor that includes the drum and load torque,
• γ is viscosity friction coefficient,
• T is the total torque given by the difference between the driven torque Tm and the resistance torque Tf,
• ω is the drum rotation speed in radiant/second.
The Laplace transform allows the passage from the time domain to the frequency domain and the above equation can be written as follows:

where s = jω is the complex variable.
The transfer function between the total torque and the drum angular speed is: Ω(s) T(s) = 1γ 1 + s J γ in which the gain G = 1 / γ and the time constant τ = J / γ or the frequency pole ω = γ / J are evidenced.
An additional step can be made by considering the motor torque is proportional to the input signal of the motor drive system α _s such as: T(s) = K_m *α _s (s). The previous equation can be written as the following one in which, simplifying, we assume K_m =1: Ω(s)α(s) = H(s) = 1γ 1 + s J γ If the clothes load, during its drum revolutions, is not well distributed, an unbalanced load, with mass M_u , will generate a speed fluctuation having the amplitude proportional to the M_u.
If we call β the angle between the unbalance and a fixed reference axis, a new torque term has to be added in the torque equilibrium equation. This torque is due to the gravity force g and its value is given by: Mu g r cos(β) in which r represents the distance between the centre of gravity of the unbalanced load and the rotation axis.
This additional torque can be taken into account in a very simple way by considering its effect as added to the speed that would be obtained if M_u were zero. Using this rationale, the measured angular speedω(t) turns out to be: ω(t) = ω0 + Mugr|H(jω0)|cos(ω0 t) where ω₀ is the system output obtained in the same conditions of no unbalanced load and |H(jω₀)| is the module of the transfer function H(s) computed at the frequency ω ₀, namely: |H(jω0)| = 1γ 1 + ω0 2 J 2 γ2 The frequency spectrum of this function is shown in Figure 3a for a given value of the friction coefficientγ (1 Nm sec/rad) and for some value of inertia loads J (0,4 and 0,8). As it can be noted, the amplitude of the frequency spectrum, computed at the frequency ω₀ depends on three variables: unbalanced load U, inertia load J and dynamic friction γ. In particular, the higher is the inertia, the lower is the amplitude of the frequency spectrum.
In figure 3b (where γ = 2 Nm sec/rad), it should be noted the big influence that the friction has on the system under analysis.
Summarising, whether there is an unbalanced load, the speed signal contains fluctuations whose amplitude is related to the unbalance weight, the total inertia of the system and the dynamic friction coefficient.
The applicant discovered that the problem of assessing the amount of unbalance and the total inertia is solved in a surprisingly simple manner by injecting into the system, as shown in Figure 2, a periodic signal with a known frequency ω _inj (for instance via a cosine wave signal) and by analysing the frequency spectrum of the controlled variable.
In Figure 2 the whole controlled system is considered instead of the open loop system schematically shown in Figure 1. All the considerations done above can be repeated in the same way by using, in place of the open loop transfer function H(s), the closed loop one H_c(s) which depends on the actual control law of the speed control system. It should be enlightened that the proposed approach works finely whatever is the controller type since it is based on the structural mechanical properties of the system. Moreover, to simplify the explanation, we consider constant target speedω _t which value has to be above the speed at which the laundry is applied to the side wall of the drum by centrifugal force.
With reference to Figure 2, the motor drive system is energised by two speed components, whose analytical value is given by: α s = α c + A *cos(2πtω inj ) Correspondingly, the measured speed can be split into two different parts ω _t and ω _inj , so that the output ω is the speed resulting by the sum of the two drivers component effects.
The first one (α _c ) is determined by the error speed information and it is driven by the speed controller action. In case of balanced load, the term J* dω / dt is zero, and the torque provided by the motor is the one requested to compensate the dynamic frictions. In case of unbalanced loads, as shown before, the torque has to be able to compensate the gravity force of the unbalanced load during its turning phases and keeping the average target speed. Doing that, a speed fluctuation with amplitude depending on the unbalance itself, is generated. The speed fluctuation signal is generally used to evaluate the unbalance amount but, due to the fact that if the inertial load increases the speed fluctuation amplitude decreases, the known methods provide approximate estimations of the real unbalanced load. The consequence with such known methods is that the spinning speed is reduced to avoid dangerous conditions. With the method according to the present invention, it is possible to assess the actual value of unbalanced load, therefore allowing higher spin speeds if compared with known methods and systems, without any risk of mechanical failure of the washing machine.
The second component (ω _inj ) is in charge of "stressing" the system in order to emphasise the effect of the inertia load. According to the invention, it is proposed and used the analysis of frequency spectrum of the motor speed in a way that, by means of the two frequencies components ω _t and ω _inj , the inertia load (total load) and the unbalanced load are univocally identified. Whereas it is difficult to mathematically solve this issue in the time domain it is, on the contrary, easily solvable in the frequency domain. By using the Laplace transform, the following equation can be written:

Because of the frequency response theorem, the steady-state response of the system to a sinusoidal input is a sine wave with the same frequency (as the input wave) but with a different amplitude and phase shift ϕ. We can so calculate the system output that is obtained after the transient as: ω(t) = ω t + Mugr*|Hc (jω t )|cos(ω t t) + A*|H c (jω inj )|cos(ω inj t + ϕ) The behaviour of the frequency response, shown in Figure 3c (where γ = 1 Nm sec/rad), explains how the system reacts at different inertia loads. It is still well evidenced that the higher is the inertia the higher is the damping effect on the speed fluctuations by looking to the transfer function modules |H_c (jωt)| and |H_c (jω _injt)| (see arrows at target speed of 100rpm, i.e. 1.66Hz, in Figure 3c).
The previous equation can be written as follows:

in which the terms Au and As represents the transfer function modules (multiplied by the M_u g r and A constants) at the frequencies ω _t and ω _inj respectively.
Each term may be computed as follows:


where NSamp are the number of data sampled and ω = ω( T + kΔt) where T is the starting sampling time.
The injected signal type cos(ω_injt) has been selected as an example only and also in view of its simple implementation, but the described method can be adopted with any injected signal type (periodic or not) with a known frequency spectrum.
The inertia load and unbalance load can now be computed on the basis of these two values Au and As.
Experimental tests, done with different inertia and unbalanced loads allow the implementation, in a micro controller, of a Neural Fuzzy Logic system. Alternatively a look up table, in which, from the combination of the system response Au and As, the inertia and unbalance value are obtained, can be implemented as well.
• As mentioned before, the method according to the present invention allows compensating the friction in a simple manner.
  The following embodiment is based on the average of the output value of the speed control. Once the speed control is designed, i.e. the PID parameters, (Proportional, Integrative and Derivative terms of the speed error respectively), a simulation of dynamics friction, caused mainly by the load plus the speed transmission system, can be done by a controlled braking system. When a constant inertia value is set, the value provided by the speed control system α _c or on differently α _s , in steady state condition, is proportional to the friction. The estimation of such variable is so computed by the evaluation of the average of the α _s :

  

  At this point all the variables are known; the Inerta Load and Unbalanced Load amount can be estimated by following the next procedure scheme :
• Compute the Alpha_Mean
• Compensate the As* = f_TI (As,Alpha_{Mean )}) by compensating the As term,
• Compute the UnbalanceLoad,InertiaLoad = f_UL (Au,As*) on the basis of the compensated As and Au values.
f_TI and f_UL are look up table data or a Neura Fuzzy Logic system, resulting from the experimental tests done.
Qualitative and simplified examples of functions f_TI and f_UL are shown in figures 5a and 5b. Figure 5a shows the general correlation between values of Alpha_Mean and compensated value of As (i.e. As*), based on different curves of As and k. Once As* is determined from the diagram of Figure 5a, such value is used in the diagram of Figure 5b, based on different curves of Au values, for determining the compensated value of Au. It is clear that the diagrams of figures 5a and 5b are experimental diagrams determined for each kind of washing machine for different values of known dynamic frictions, total loads and unbalanced loads. Look up table data, Neura Fuzzy Logic system or correlation diagrams may be advantageously stored in the central processing unit (microprocessor) that controls the washing machine.
The relation found to compensate the As term (shown in figure 5a) has the following general expression:

where k depends on the motor caracteristics and the resistive torque of the mechanical system.
A flow chart of the control algorithm is shown in Figures 4a and 4b. In Figure 4a the first block 10 is the starting phase in which the sampling counter is reset. When the number of sampling is equal or higher than a predetermined number, then in Figure 4b the blocks 12 and 14 compute the frequency responses for assessing total load and unbalanced load respectively. The values computed in blocks 12 and 14 are fed to block 16, together with the value of Alpha Mean, indicative of dynamic friction, for compensating this latter. The compensated values of inertia load (total load) and unbalanced load are the output of the control algorithm and such values may used by the central processing unit of the washing machine for different purposes, for instance for setting the spin speeds or for setting the amount of fresh water to be used during rinsing.
Figure 6 shows a fuzzy rules scheme adopted in case of a fuzzy control solution. Fuzzy rules are of the following type:

Rule 1	If As* is L and Au is M	then J is H and U is H
Rule 2	If As* is M and Au is L	then J is M and U is VL
Rule 3	If As* is H and Au is VH	then J is VL and U is H

etc.

Claims (10)

1. Method for determining load and unbalance load in a washing machine having a motor drive system for rotating the motor at a target speed, characterised in that the motor drive system is supplied with an auxiliary signal having a predetermined frequency spectrum and in that the total load and unbalanced load are determined by means of frequency spectrum analysis at different frequencies.
2. Method according to claim 1, characterised in that said auxiliary signal is a periodic signal having a frequency different from the frequency corresponding to the target speed.
3. Method according to claim 2, characterised in that the total load and unbalanced load are determined by means of frequency spectrum analysis at a frequency corresponding to said auxiliary signal and at a frequency corresponding to the target speed respectively.
4. Method according to any of the preceding claims, characterised in that it comprises a compensation of values of total load and unbalanced load taking into account the dynamic friction, such compensation being based on experimental tests done at different and predetermined values of dynamic frictions.
5. Method according to claim 4, characterised in that said compensation is carried out by means of look up table data or of fuzzy logic system.
6. Method according to any of the preceding claims, characterised in that said frequency spectrum analysis is carried out by means of Laplace transform.
7. Method according to claim 6, characterised in that the module of the transfer function H(s) computed at the frequency ω₀ is of the following type: H(jω0) = 1γ 1 + ω0 2 J 2 γ2 where: jω₀ is the complex variable; γ is the friction coefficient, J is the inertia load value.
8. Washing machine comprising an electronic control system including a motor drive system for rotating the motor at a target speed, and a rotating drum driven by the motor thorugh a transmission system, characterised in that the electronic control system is apt to supply the motor drive system with an auxiliary signal having a predetermined frequency spectrum, the electronic control system being apt to determine the total load and unbalanced load by means of frequency spectrum analysis at different frequencies.
9. Washing machine according to claim 8, characterised in that the electronic control system is apt to supply the motor drive system with a periodic signal having a frequency different from the frequency corresponding to the target speed of the motor.
10. Washing machine according to claim 9, characterised in that the electronic control system is apt to determine the total load and unbalanced load by means of frequency spectrum analysis at a frequency corresponding to said auxiliary signal and at a frequency corresponding to the target speed respectively.

Images