A method for detecting laundry eccentricity in a drum washing machine

Abstract

The application discloses a kind of drum washing machine clothes eccentricity detection methods, comprising the following steps: obtaining real-time eccentric load value, and obtaining total load value;According to the size of drum, set balanced load quality;Eccentric load value, total load value and balanced load quality value are input into clothes eccentricity prediction model, and eccentric mass is calculated;Wherein, the training method of the clothes eccentricity prediction model includes: by simulating drum experiment, change drum balanced load quality and eccentric mass, obtain corresponding multiple sets of eccentric load value and total load value;Balanced load quality, eccentric load value and total load value are used as BP neural network input layer, and eccentric mass is used as BP neural network output layer, and the data of input and output are trained.The drum washing machine clothes eccentricity detection method provided by the application can realize cost saving and simplification of detection method steps while ensuring accuracy.

Description

Technical Field

[0001] This invention relates to a method for detecting eccentricity of clothes in a drum washing machine, belonging to the field of washing machine drum motor control. Background Technology

[0002] Because of differences in the material, size, and absorbency of clothing, the distribution of water inside the drum is often uneven. This uneven distribution causes vibration and noise in drum washing machines, and in more serious cases, drum displacement or even disintegration. The even distribution of water inside the drum is also crucial for the motor's ability to operate at high speed during the subsequent spin-drying stage. Therefore, understanding how to detect and reflect this unevenness is an important issue.

[0003] Uneven distribution of clothes inside a drum washing machine causes vibration. When there is no eccentric load or a small eccentric load inside the drum, the load distribution is balanced, and the vibration decreases. However, because the distribution of clothes inside the drum is a complex, non-linear motion that is difficult to model mathematically, completely reducing eccentricity is not easy. Although researchers have been working to improve washing machine performance and reduce vibration, machines sometimes still experience severe vibration and noise. Therefore, technologies to reduce the vibration of drum washing machines still require further research and development.

[0004] Traditional methods to address this issue often start with mechanical structures, which can achieve vibration and noise reduction to some extent. However, the design and analysis are relatively complex, highly dependent on experience, and lack portability, resulting in high costs for vibration reduction design and significant expenses for subsequent maintenance and repair. Some high-precision detection methods utilize high-precision sensors for measurement, such as displacement and acceleration. While this greatly improves accuracy, the challenge lies in converting the measured data of these influencing factors into data that the main control module can process. This data conversion process is not only labor-intensive but also extremely expensive. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to overcome the defects of the prior art and provide a method for detecting the eccentricity of clothes in a drum washing machine, which can achieve cost savings and simplification of the detection method steps while ensuring accuracy.

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0007] A method for detecting eccentricity of clothes in a drum washing machine includes the following steps:

[0008] Get the real-time eccentric load value and the total load value;

[0009] Set the balanced load mass according to the size of the drum;

[0010] The eccentric load value, total load value, and balanced average load mass value are input into the clothing eccentricity prediction model to calculate the eccentricity mass.

[0011] The training method for the clothing eccentricity prediction model includes:

[0012] By simulating a roller experiment, the equalized load mass and eccentric mass of the roller were changed to obtain multiple sets of corresponding eccentric load values ​​and total load values.

[0013] The balanced load mass, eccentric load value, and total load value are used as the input layer of the BP neural network, and the eccentric load is used as the output layer of the BP neural network. The input and output data are then used for training.

[0014] Obtaining eccentric load values ​​includes:

[0015] Control the motor speed to reach the first set speed and maintain it stably for a period of time;

[0016] The real-time speed of the motor is detected, and the difference between the actual speed and the first set speed is calculated to obtain the speed difference value. This speed difference value is then integrated within the detection cycle to obtain the eccentric load value. Obtaining the total load value includes:

[0017] Control the motor to accelerate to the second set speed, and then decelerate to the third set speed;

[0018] The power change of the motor during the deceleration period from the second set speed to the third set speed is detected;

[0019] The working value is calculated based on the power change, and then the total load value is calculated based on the working value. Before obtaining the total load value, it is determined whether the eccentric load value meets the preset range. If the requirement is met, the process of obtaining the total load value is carried out. If the requirement is not met, the clothes in the drum need to be shaken and dispersed before the process of obtaining the eccentric load value is carried out.

[0020] The power was specifically measured as follows:

[0021] The current is sampled once per mechanical cycle of the washing machine drum to indirectly obtain the actual power value. The actual power values ​​obtained each time are compared, and the maximum and minimum actual power values ​​are selected. This process is repeated multiple times to obtain datasets of maximum and minimum actual power values. The highest and lowest values ​​in the maximum actual power value dataset are removed, and then the average is taken. Similarly, the highest and lowest values ​​in the minimum actual power value dataset are removed, and then the average is taken again. The final actual power value is obtained. The BP neural network training partitions the sample space as follows: 75% training sample set, 15% validation sample set, and 10% test sample set.

[0022] The specific process for calculating the total load value includes:

[0023] Wait for the motor to reach the second set speed, sample the current, perform power filtering, and record the power value P1;

[0024] Wait for the motor to reach the third set speed, sample the current, perform power filtering, and record the power value P2;

[0025] Record the time T of the entire deceleration process, and multiply the difference between the power values ​​P1 and P2 by the time T to calculate the work done in the entire process.

[0026] Repeat the above process multiple times, add up the work values ​​obtained each time to get the total work value, and then divide the total work value by the number of times to get the total load value.

[0027] A device for detecting eccentricity of clothes in a drum washing machine includes:

[0028] The eccentric load value and total load value acquisition module is used to obtain the real-time eccentric load value and total load value;

[0029] The balanced load setting module is used to set the balanced load mass according to the size of the drum.

[0030] A clothing eccentricity prediction model, wherein the training method of the clothing eccentricity prediction model includes:

[0031] By simulating a roller experiment, the equalized load mass and eccentric mass of the roller were changed to obtain multiple sets of corresponding eccentric load values ​​and total load values.

[0032] The balanced load mass, eccentric load value, and total load value are used as the input layer of the BP neural network, and the eccentric load is used as the output layer of the BP neural network. The input and output data are then used for training.

[0033] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a method for detecting eccentricity of clothes in a drum washing machine.

[0034] The beneficial effects of this invention are as follows: This invention provides a method for detecting clothes eccentricity in a drum washing machine. By measuring real-time speed and power, and processing the data, it converts the data into eccentric load values ​​(OOB value) and total load values ​​(LOAD value) that reflect the distribution of clothes inside the washing machine drum. The first stage of the motor's rotational speed is used for OOB value detection, while the second and third stage speeds are used for LOAD value detection. The method is simple, and the detection equipment is low-cost. Furthermore, this invention utilizes a BP neural network to establish a predictive model for clothes eccentricity in a drum washing machine. This model can predict the eccentricity mass inside the drum based on the average load mass, OOB value, and LOAD value. This solves the problem in existing technologies where data processing using algorithms yields a measure of clothes eccentricity, but still requires manual data fitting, significantly improving the accuracy of clothes eccentricity identification. Attached Figure Description

[0035] Figure 1 This is a flowchart illustrating a method for detecting clothing eccentricity provided by the present invention.

[0036] Figure 2 An equivalent mathematical model analysis diagram is provided for this invention;

[0037] Figure 3 A flowchart illustrating the specific algorithm for clothing eccentricity detection provided by this invention;

[0038] Figure 4 The flowchart of the OOB algorithm provided by this invention;

[0039] Figure 5 The OOB numerical processing flowchart provided by this invention;

[0040] Figure 6 The flowchart of the LOAD algorithm provided by this invention;

[0041] Figure 7 The LOAD numerical processing flowchart provided by this invention;

[0042] Figure 8 The diagram illustrates the training process of the BP neural network provided by this invention. Detailed Implementation

[0043] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to illustrate the technical solution of the present invention more clearly, and should not be used to limit the scope of protection of the present invention.

[0044] This invention provides a washing machine eccentricity detection method applicable to drum washing machines before the spin-drying cycle. It measures the distribution of fabric within the drum by detecting Out-of-Body (OOB) and Load (LOAD) values, which differ under varying eccentricity and load conditions. Since the fabric distribution within the drum is uneven, it can be equivalently transformed during actual analysis, such as... Figure 2 The figure shows the mass of clothing equivalent to the sum of eccentric mass and uniformly loaded mass, i.e., M. 衣物 =M LOAD =M OOB +M BLC Among them, M OOB The numerical value represents the eccentric mass, M. BLC The numerical value represents the uniform load mass, M LOAD Indicates the total mass.

[0045] like Figure 1 and Figure 3 As shown, a washing machine eccentricity detection method of the present invention mainly includes the following steps:

[0046] Step 1: After initializing the program, the program enters a waiting phase, controlling the motor speed to reach the first set speed and maintaining it stably for a period of time.

[0047] Step two involves determining if the real-time rotational speed differs excessively from the set threshold speed. The algorithm pre-sets a range for this difference, and actual speed detection can only proceed after the actual speed meets the set requirements. If the difference is too large, it indicates that the current actual speed is unstable, and the speed needs to be adjusted to a stable level before detection can proceed. If the difference is within the required range (generally 3 to 4 revolutions per minute), the detection phase begins. The initial set speed needs to be determined based on the specific application scenario, typically ranging from 80 rpm to 110 rpm.

[0048] The OOB value is obtained by integral processing within the detection period, which is 1ms. The algorithm flowchart for the OOB value detection stage is as follows. Figure 4 As shown, the testing phase mainly involves detecting the real-time rotational speed at the highest or lowest position of the inner tub. After the required number of tests is reached, data processing is performed. Specifically, the process of analyzing OOB values ​​is as follows: Figure 5 As shown, the absolute value of the difference between the actual rotational speed and the set first rotational speed is integrated within the detection period to obtain the OOB value, including the following steps: Step a, delay for a period of time and wait for the speed to stabilize; Step b, sample the speed and perform filtering processing; Step c, calculate the difference between the actual rotational speed and the set rotational speed and integrate it within the detection period; Step d, compare the integration result with the number of detections to obtain the OOB value.

[0049] Step 3: Determine if the OOB value meets the requirements. If it does, perform a LOAD check. If it does not meet the requirements, the clothes in the drum need to be shaken out.

[0050] Step four: Once the OOB value meets the requirements, the system will enter the LOAD detection stage, waiting for the motor to reach the second set speed, and then decelerating to the third set speed. The second and third set speeds need to be determined based on the specific application scenario, but it is essential to ensure that the second set speed is greater than the third set speed, and the third set speed is greater than the first set speed.

[0051] Step 5: The LOAD value is obtained after algorithmic data processing. The algorithm flowchart for the LOAD detection stage is shown below. Figure 6 As shown, after initialization, the program enters the LOAD waiting phase, which checks if the real-time speed differs significantly from the set threshold speed. If the difference is too large, it indicates that the actual speed is unstable, and the speed needs to be adjusted to a stable level before detection can proceed. If the difference is within the required range, the program enters the detection phase. If the program fails to enter the LOAD detection phase, it needs to be re-initialized. During the detection phase, the speed is increased to the maximum set speed. If the actual speed differs significantly from the set maximum speed, the error between the running time and the detection time needs to be adjusted. After reaching the maximum set speed, the speed is reduced to the minimum set speed. When reducing to the minimum speed, the speed is increased back to the maximum set speed, and the process of speed decreasing from high speed to low speed to high speed is recorded until the required number of repetitions is met. The power change during this process is processed to obtain the LOAD value. The process of analyzing the LOAD value is as follows. Figure 7 As shown, the process includes the following steps: Step a, wait for the highest detection speed to be reached, perform current sampling, perform power filtering, and record the power value P1; Step b, wait for the lowest detection speed to be reached, perform current sampling, perform power filtering, and record the power value P2. The washing machine drum performs a current sampling once per mechanical cycle, indirectly obtaining the actual power value. The actual power values ​​obtained each time are compared, and the maximum and minimum actual power values ​​are selected from the obtained data. This process is repeated 8 times. To prevent outliers, the TRIMMEAN function is used to process the data, first removing the highest and lowest values ​​from the dataset, and then averaging the results. The resulting data is used as the actual power value. Step c, record the entire process time T and calculate the total work done during the process; Step d, repeat the entire process N times, and compare the total work done with the number of repetitions to obtain the LOAD value. The entire LOAD stage needs to be performed twice, and the power obtained from the two repetitions is averaged to obtain the final LOAD value.

[0052] Step 6: Determine whether the detected OOB and LOAD values ​​meet the requirements for pre-dehydration. If they do, perform the pre-dehydration operation. When the drum is performing the pre-dehydration operation, the drum speed should generally not exceed 400 r / min. The purpose of pre-dehydration is to remove as much water as possible from the clothes in the drum, reduce eccentricity, and make it easier to meet the requirements of high-speed dehydration later.

[0053] Step 7: After pre-dehydration, another round of OOB and LOAD value detection will be performed, and the process is very similar to the first OOB value detection process.

[0054] Step 8: Change the balanced load mass, eccentric mass, and position within the drum, and measure for various different situations, recording a set of data for each situation. Place the eccentric block and balance block inside the washing machine drum for testing, performing multiple tests at different positions, and changing the mass of the eccentric block and balance block accordingly, obtaining multiple sets of OOB and LOAD values. Use the balanced load mass, OOB value, and LOAD value as the input layer, and the eccentric mass (balanced load) as the output layer, and import them into a BP neural network. Divide the sample space into a training sample set (75%), a validation sample set (15%), and a test sample set (10%).

[0055] Step nine: Initialize the BP neural network model, set the BP neural network control parameters, determine the number of hidden neurons, adjust the number of training iterations, training rate, target error threshold, etc., and train on the previously constructed sample space to establish the BP neural network clothing eccentricity prediction model. Adjust the BP neural network control parameters, the number of hidden neurons, error, threshold, and other parameters, and repeat the training until the optimal correlation is achieved, and then output the final BP neural network-based clothing eccentricity prediction model.

[0056] Specifically, the constructed BP neural network has 3 neurons in the input layer and 1 neuron in the output layer. The number of hidden layer nodes is calculated using the following formula, while maintaining the required accuracy: Where m is the number of input neurons, n is the number of output neurons, and a is an integer between 0 and 10.

[0057] The basic parameters include: learning rate μ, and weights W from the input layer to the hidden layer. ij The weights W from the hidden layer to the output layer ik The number of biases a from the input layer to the hidden layer j The number of biases b from the hidden layer to the output layer k and the activation function f(x); where the weights W from the input layer to the hidden layer ij The weights W from the hidden layer to the output layer ik The number of biases a from the input layer to the hidden layer jThe number of biases b from the hidden layer to the output layer k The initial value is a random number within the range (-1, 1). The activation function f(x) is calculated using the following formula: Where x is the input of the hidden layer.

[0058] The training of the BP neural network model is as follows: Figure 8 As shown, the specific steps are as follows: Provide a set of model training samples, and loop for each training sample p; the loop is to calculate the input and output of each node of the network layer by layer in the forward direction, calculate the output error of the qth sample and the total error E of the network, and when E is less than the allowable error ε or the specified number of iterations p is reached, the learning process ends; otherwise, backpropagate the error and calculate the error of each node of the network layer by layer to correct the network connection weights.

[0059] Step 10: Once the required OOB and LOAD values ​​are obtained, the new data is predicted using the clothing eccentricity prediction model, and the predicted eccentricity result is fed back to the main control module for judgment. If the requirements for high-speed dehydration are met, the dehydration operation will proceed; otherwise, the clothing in the drum will be re-dispersed, and the process will return to the OOB and LOAD detection stages.

[0060] This invention also discloses a device for detecting eccentricity of clothes in a drum washing machine, comprising:

[0061] The eccentric load value and total load value acquisition module is used to obtain the real-time eccentric load value and total load value;

[0062] The balanced load setting module is used to set the balanced load mass according to the size of the drum.

[0063] A clothing eccentricity prediction model, wherein the training method of the clothing eccentricity prediction model includes:

[0064] By simulating a roller experiment, the equalized load mass and eccentric mass of the roller were changed to obtain multiple sets of corresponding eccentric load values ​​and total load values.

[0065] The balanced load mass, eccentric load value, and total load value are used as the input layer of the BP neural network, and the eccentric load is used as the output layer of the BP neural network. The input and output data are then used for training.

[0066] Finally, this invention discloses a computer-readable storage medium storing a computer program that, when executed by a processor, implements a method for detecting eccentricity of clothes in a drum washing machine.

[0067] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method of detecting the eccentricity of laundry in a drum washing machine, characterized in that: The method comprises the following steps: obtaining a real-time eccentric load value and a total load value, wherein the total load value is obtained by: controlling the motor to accelerate to a second set speed and then to decelerate to a third set speed; detecting the power change of the motor during the deceleration from the second set speed to the third set speed; calculating the work value according to the power change, and then calculating the total load value according to the work value; setting a balance load mass according to the size of the drum; inputting the eccentric load value, the total load value and the balance load mass value into a clothes eccentricity prediction model to calculate an eccentric mass; wherein the training method of the clothes eccentricity prediction model comprises: changing the balance load mass and the eccentric mass through a simulation drum experiment to obtain a plurality of sets of corresponding eccentric load values and total load values; taking the balance load mass, the eccentric load value and the total load value as the input layer of the BP neural network, and taking the eccentric mass as the output layer of the BP neural network, and training the input and output data.

2. The laundry eccentricity detection method of a drum-type washing machine according to claim 1, characterized in that: The method for obtaining the eccentric load value comprises: controlling the motor speed to reach a first set speed and maintaining the stability for a period of time; detecting the real-time speed of the motor, obtaining the speed difference value by subtracting the actual speed from the first set speed, and integrating the speed difference value in the detection period to obtain the eccentric load value.

3. The laundry eccentricity detection method of a drum-type washing machine according to claim 2, characterized in that: Before obtaining the total load value, it is judged whether the eccentric load value meets the preset range, if the requirement is met, the total load value is obtained, and if the requirement is not met, the clothes in the drum need to be shaken and the eccentric load value is obtained again.

4. The laundry eccentricity detecting method of the drum-type washing machine according to claim 2, characterized in that: The power is determined as follows: The current is sampled once every mechanical cycle of the drum to indirectly obtain the actual power value, the actual power values obtained each time are compared, the maximum actual power value and the minimum actual power value in the obtained data are screened out, the maximum actual power value data set and the minimum actual power value data set are obtained by repeating multiple times, the maximum value and the minimum value in the maximum actual power value data set are removed and then averaged, and the maximum value and the minimum value in the minimum actual power value data set are removed and then averaged, and finally the actual power value is obtained.

5. The method of claim 1, wherein the method comprises: determining whether the laundry is present in the drum; and determining whether the laundry is present in the drum based on the rotational frequency of the drum. The training of the BP neural network divides the sample space into a training sample set of 75%, a verification sample set of 15% and a test sample set of 10%.

6. The method of claim 1, wherein the method comprises: determining whether the laundry is present in the drum; and determining whether the laundry is present in the drum based on the rotational frequency of the drum. The specific process for calculating the total load value comprises: waiting for the motor to reach the second set speed, sampling the current, performing power filtering processing, and recording the power value P1; waiting for the motor to reach the third set speed, sampling the current, performing power filtering processing, and recording the power value P2; recording the time T of the entire deceleration process, multiplying the power value P1 and the power value P2 after the subtraction, and calculating the work value of the entire process; repeating the above process multiple times, adding the work values obtained each time to obtain the total work value, and then dividing the total work value by the number of times to obtain the total load value.

7. A drum laundry machine laundry eccentricity detection device, characterized by, The method comprises: an eccentric load value and a total load value obtaining module for obtaining a real-time eccentric load value and a total load value; a balance load mass setting module for setting a balance load mass according to the size of the drum; The clothes eccentricity prediction model, and a training method of the clothes eccentricity prediction model comprises: Through a simulation drum experiment, a balanced uniform load mass and an eccentric mass of a drum are changed, and corresponding multiple sets of eccentric load values and total load values are obtained; The balanced uniform load mass, the eccentric load value and the total load value are taken as a BP neural network input layer, and the eccentric mass is taken as a BP neural network output layer, and data of input and output are trained. 8.A computer readable storage medium, having stored thereon a computer program, which, when executed by a processor, implements the clothes eccentricity detection method of the drum washing machine according to any one of claims 1-6.

Images