Industrial washing machine control optimization method and system based on incremental fuzzy neural network

Abstract

The invention discloses an industrial washing machine control optimization method and system based on an incremental fuzzy neural network, and the method comprises the steps: collecting to-be-washed clothes and washing environment signals through a data collection unit, receiving the data signals through a central processing unit, and displaying the data signals on a display unit; the optimal washing setting is obtained through calculation of the carried fuzzy neural network model, a control signal is sent out, the controller controls the actuators to execute the corresponding washing setting, and the washing machine collects the current washing signal again through the data collection unit after executing the washing operation every time. And through an incremental optimization algorithm, fuzzy neural network sample addition and unfavorable sample elimination are realized, and updating of the fuzzy neural network is completed. According to the intelligent control method and system, the defects that in an existing industrial washing machine control technology, a control method is not flexible, and resource waste is serious are overcome, the intelligent control method and system which can rapidly predict washing information and execute washing operation and have high robustness are provided, and the service life of the washing machine is greatly prolonged.

Description

Technical Field

[0001] The present application relates to the technical field of washing machine control, and in particular to an industrial washing machine control optimization method and system based on an incremental fuzzy neural network. Background Art

[0002] Industrial washing machines have multiple washing functions and are therefore widely used in public places with large-capacity laundry requirements, such as restaurants, hospitals, schools, and factories. In recent years, the demand for industrial washing machines has been driven by multiple factors, including the expansion of commercial and medical facilities, improved hygiene standards, rising labor costs, sustainable development needs, and technological advances.

[0003] As the demand for industrial laundry continues to grow, the control methods of traditional industrial washing machines have shown significant deficiencies in flexibility, efficiency and environmental friendliness. The modern industrial environment requires laundry equipment to not only handle a variety of fabric types and washing loads, but also to have energy-saving, water-saving, high-efficiency and environmentally friendly characteristics. Traditional washing machines provide fixed washing cycles and limited customization options, lacking flexibility, resulting in unsatisfactory washing results when dealing with different laundry loads and fabric types. In addition, traditional washing machines cannot accurately control water temperature, washing time and motor speed, resulting in high energy consumption and serious waste of water resources, which not only increases operating costs, but also has adverse effects on the environment. Frequent mechanical wear and maintenance requirements further reduce equipment reliability and production efficiency.

[0004] Therefore, it is necessary to further solve the above problems. The applicant has proposed an industrial washing machine control optimization method and system based on an incremental fuzzy neural network. Summary of the invention

[0005] In view of the deficiencies of the prior art, the present invention provides an industrial washing machine control optimization method and system based on an incremental fuzzy neural network, which solves the problems of inflexible control methods and serious waste of resources in the prior art industrial washing machine control technology mentioned in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions: An industrial washing machine control system based on an incremental fuzzy neural network, the control system is composed of a data acquisition unit, a central processing unit, a controller unit, a display unit and a server, the data acquisition unit, the controller unit, the display unit and the server are respectively connected to the central processing unit; The data acquisition unit comprises a pressure sensor, a turbidity sensor, a temperature sensor and a camera. The pressure sensor is arranged at the bottom of the laundry compartment, the turbidity sensor is arranged in the laundry compartment, the temperature sensor is arranged in the laundry compartment, and the camera is arranged at the top of the laundry compartment.

[0007] The present invention also provides an optimization method for an industrial washing machine control system based on an incremental fuzzy neural network, the method specifically comprising the following steps: The pressure sensor is used to detect the amount of laundry and output a laundry amount detection signal. ; The turbidity sensor is used to detect the degree of dirtiness of the laundry and output a dirtiness detection signal. ; The temperature sensor is used to detect the water temperature and output a temperature detection signal ; The camera is used to detect the material of the laundry and output a signal of the material of the laundry. ; The central processing unit is used to receive the data signal sent by the data acquisition unit and perform algorithm calculation; The display unit is used to display the washing time , washing intensity and detergent dosage ; The server is used to store washing sample information, including data signals of the data acquisition unit that completes the current washing and control signals output to the control unit; The controller unit is used to receive the washing instruction sent by the central processing unit and output a control signal, and the controller unit receives the control signal to start.

[0008] As a further solution of the present invention: the algorithm calculation is used to convert the data signal output by the data acquisition unit into a control signal output to the controller unit, and the algorithm calculation adopts an incremental fuzzy neural network algorithm, and the incremental fuzzy neural network has a fuzzy neural network and incremental sample update.

[0009] As a further solution of the present invention: the fuzzy neural network has a five-layer network structure, including an input layer, a fuzzification layer, a fuzzy reasoning layer, a normalization layer and an output layer; The fuzzy neural network has a loss function: in, is the loss function, is the total number of samples, For the The target output of samples is Predict output for the model; The fuzzy neural network reverse update formula: in, are the network parameters in the model, For the The parameter value after iterations; is the learning rate; is the step size, that is, the distance of parameter update in each gradient direction.

[0010] As a further solution of the present invention: the incremental sample update includes sample addition and sample deletion; The sample is appended with an iterative formula: in, To finally meet the washing requirements, is the number of washes, For the Prediction output, including wash time , washing intensity and detergent dosage ; The sample pruning includes time-based sample pruning and verification error-based sample pruning. The time-based sample pruning is used to delete expired data in a fixed time window, and the verification error-based sample pruning is used to delete unfavorable samples in the sample set.

[0011] As a further solution of the present invention: the specific steps of sample deletion based on verification error are as follows: The unfavorable sample calculation formula is: in, is the loss function; is the training sample, ; To test the samples, ; is the optimal solution for the parameters in the model. The model of training samples, is defined as follows: H is the Hessian matrix, and the calculation formula is as follows: .

[0012] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention integrates data acquisition equipment, a server, a controller unit, a display unit and a central processing unit. When working, the data acquisition unit collects information about the laundry and water temperature, the controller unit transmits the collected information as input to the central processing unit, the central processing unit outputs control instructions through the algorithm calculation module, the controller unit receives the control instructions returned by the central processing unit and outputs control signals, the motor, timer and detergent dispensing device receive the control instructions to start, thereby generating washing intensity, washing time and detergent dosage. That is, the system can intelligently generate and execute washing plans, thereby improving the washing efficiency of industrial washing machines and reducing energy consumption.

[0013] 2. The present invention executes the intelligent control algorithm on the central processing unit and interacts with the user through the display screen, thereby reducing costs and achieving more friendly human-computer interaction.

[0014] 3. The present invention uses an incremental fuzzy neural network control algorithm, which can better formulate washing plans, record washing data, extend service life, reduce energy waste, and solve the problems of poor washing effect, energy waste and short service life of industrial washing machines caused by patterned washing.

[0015] 4. This control system uses LabVIEW to design the central processing unit. The display screen and washing machine are connected through a serial port, eliminating the original dedicated controller for the washing machine, reducing costs, and using the display screen to achieve human-computer interaction, which is more efficient and more friendly to the human-computer relationship. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without creative work. Among them: Figure 1 is a system architecture diagram of the present invention; Figure 2 This is the structural diagram of the fuzzy neural network of the present invention; Figure 3 This is a flow chart of sample appending update and fuzzy neural network update for the present invention. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] Example See also Figure 1 As shown, the present invention provides an industrial washing machine control system based on an incremental fuzzy neural network, the control system is composed of a data acquisition unit, a central processing unit, a controller unit, a display unit and a server, and the data acquisition unit, the controller unit, the display unit and the server are respectively connected to the central processing unit; The data acquisition unit comprises a pressure sensor, a turbidity sensor, a temperature sensor and a camera. The pressure sensor is arranged at the bottom of the laundry compartment, the turbidity sensor is arranged in the laundry compartment, the temperature sensor is arranged in the laundry compartment, and the camera is arranged at the top of the laundry compartment.

[0019] The present invention also provides an optimization method for an industrial washing machine control system based on an incremental fuzzy neural network, the method specifically comprising the following steps: Step S1: Construct the Figure 2 The five-layer fuzzy neural network structure shown in the figure includes an input layer, a fuzzification layer, a fuzzy reasoning layer, a normalization layer, and an output layer. The training samples are input for network training to obtain a network model that performs well on the test samples. , specifically: (1.1) Determine the feature values ​​and label values ​​of the model training and test samples, where the feature value is the weight of the clothing , Dirt , washing water temperature , Clothing material , the label value is washing time , washing intensity , Detergent dosage ; (1.2) Divide all samples described in (1.1) into 75% of the samples as training sets for loss calculation and updating network weights, and 25% as test sets for testing the performance of samples on unknown data; (1.3) Input the training samples in (1.2) into the input layer of the constructed 5-layer fuzzy neural network, and use the fuzzification layer of the second layer to fuzzify the membership degree of the input variables of the input layer. The specific formula is as follows: In the formula, is the membership function, is the real value of the input, is the center value of the Gaussian function, is the standard deviation. In the fuzzy reasoning process, Gaussian function is often used to fuzzify the input variables and convert them into fuzzy membership values; (1.4) The fuzzy reasoning layer is used to calculate the matching degree of the fuzzy rules. The specific calculation method is as follows: In the formula, is the matching function, is the corresponding membership function in (1.3) , m is the total number of nodes; (1.5) Through the fourth layer, the normalization layer, the fuzzy value is calculated , and finally the fuzzy variables are clarified through the output layer to obtain the final control variable value , the specific calculation method is as follows: , In the formula, is the fuzzy weight, is the fuzzy value; (1.6) Use the gradient descent algorithm to calculate the center value of the Gaussian function in steps (1.1) and (1.5) and fuzzy weight For optimization, the specific gradient descent uses the following mean square error loss function (MSE): in, is the loss function, is the total number of samples, For the The target output of samples is Predict output for the model; The following formula is used to perform gradient calculation and weight update of fuzzy neural network: in, Represents the network parameters in the model, which can be weights or membership function parameters. is the parameter value after the j+1th iteration; is the learning rate; is the step size, that is, the distance of parameter update in each gradient direction.

[0020] Step S2: The data acquisition unit includes a pressure sensor, a turbidity sensor, a temperature sensor and a camera to collect information about the laundry to be washed and the washing environment. Specifically, the pressure sensor is set at the bottom of the laundry bin to detect the weight of the laundry and output a laundry weight detection signal. The turbidity sensor is set in the laundry compartment to detect the dirtiness of the laundry and output a dirtiness detection signal. The temperature sensor is set in the laundry compartment to detect the water temperature and output a temperature detection signal. The camera is set on the top of the laundry bin to detect the material of the laundry and output the material signal of the laundry. .

[0021] Step S3: The central processing unit receives the signal collected by the sensor in step 2, obtains the best washing setting for this laundry by calculation, and executes the optimal setting through the controller, specifically including: (3.1) The central processing unit receives the sensor data from step 2, including the weight of the clothes , Dirt , washing water temperature And clothing material ; (3.2) The central processor converts the sensor data in (3.1) As input value, use the fuzzy neural network model trained in step 1 , that is, in step 1 (1.1)~(1.5), the forward calculation of the fuzzy neural network is performed to obtain the optimal washing setting under the current input variables .

[0022] (3.3) After the central processing unit completes the calculation, it outputs the corresponding control signal to the controller. The controller is a single-chip microcomputer that receives the washing instructions sent by the central processing unit and outputs the control signal. The actuator receives the control signal to start the corresponding operation, that is, to set the optimal washing time in the execution step (3.2).

[0023] Step S4: Measure the actual data information after each washing to perform iterative prediction, add the predicted value and the initial setting value as the new expanded sample for neural network training, use the new sample set for training to eliminate unfavorable samples, reduce the training time of the fuzzy neural network, and ensure the reliability of the network output results. Figure 3 Shown include: (4.1) After completing a wash cycle in step 3, use the data sensor to recollect the current dirt signal ; (4.2) Determine the degree of dirtiness Whether the requirements are met, if yes, go to (4.3), if not, go to (3.2) to get the washing time again , continue to the next washing process, and judge the degree of dirt again after washing. Whether the requirements are met, if yes, proceed to (4.3), otherwise proceed to (4.2); (4.3) Washing is completed and a new sample is finally obtained ,in , , is the test value of the sensor before the first wash. To ultimately meet the washing requirements, the optimal washing settings, is the number of washes, Represents the i-th prediction output; (4.4) When the number of new samples reaches a certain proportion of the number of original samples, all samples are divided into training sets and test sets, where the training samples are , test sample .

[0024] (4.5) Calculate a single training sample in (4.4) For test samples The error influence is taken into account, and samples with large error influence are removed to obtain a new sample set. The specific calculation formula is: in, is the loss function; is the training sample, ; To test the samples, ; is the optimal solution for the parameters in the model. The model of training samples, is defined as follows: H is the Hessian matrix, and the calculation formula is as follows: ; (4.6) Re-execute step 1 using the new sample set to obtain the optimal fuzzy network model , replace the original prediction model with this model, that is , subsequent processors will adopt new To calculate the optimal wash settings .

[0025] The present invention adopts the system shown in the accompanying drawings to complete the above data processing and washing control process, and the central processing unit is designed by LabVIEW to better realize the monitoring and control functions of the system.

[0026] The implementation principle of the intelligent control method and system of an industrial washing machine of the present invention is as follows: during operation, the weight, degree of dirtiness, water temperature and material information of the clothes are collected through the terminal device, transmitted to the central processing unit for incremental fuzzy neural network algorithm calculation, and a control signal is generated and fed back to the controller to instruct the actuator to perform the corresponding washing operation, and various data and status information of the washing process are displayed in real time, thereby realizing efficient and accurate intelligent washing control.

[0027] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An industrial washing machine control system based on incremental fuzzy neural network, characterized by: The control system is composed of a data acquisition unit, a central processing unit, a controller unit, a display unit and a server, and the data acquisition unit, the controller unit, the display unit and the server are respectively connected to the central processing unit; The data acquisition unit comprises a pressure sensor, a turbidity sensor, a temperature sensor and a camera. The pressure sensor is arranged at the bottom of the laundry compartment, the turbidity sensor is arranged in the laundry compartment, the temperature sensor is arranged in the laundry compartment, and the camera is arranged at the top of the laundry compartment.

2. An optimization method for an industrial washing machine control system based on an incremental fuzzy neural network as claimed in claim 1, characterized in that: The specific steps of the method include: The pressure sensor is used to detect the amount of laundry and output a laundry amount detection signal. ; The turbidity sensor is used to detect the degree of dirtiness of the laundry and output a dirtiness detection signal. ; The temperature sensor is used to detect the water temperature and output a temperature detection signal ; The camera is used to detect the material of the laundry and output a signal of the material of the laundry. ; The central processing unit is used to receive the data signal sent by the data acquisition unit and perform algorithm calculation; The display unit is used to display the washing time , washing intensity and detergent dosage ; The server is used to store washing sample information, including data signals of the data acquisition unit that completes the current washing and control signals output to the control unit; The controller unit is used to receive the washing instruction sent by the central processing unit and output a control signal, and the controller unit receives the control signal to start.

3. The optimization method of an industrial washing machine control system based on an incremental fuzzy neural network according to claim 2 is characterized in that: The algorithm calculation is used to convert the data signal output by the data acquisition unit into a control signal output to the controller unit. The algorithm calculation adopts an incremental fuzzy neural network algorithm. The incremental fuzzy neural network has a fuzzy neural network and an incremental sample update.

4. The optimization method of an industrial washing machine control system based on an incremental fuzzy neural network according to claim 3 is characterized in that: The fuzzy neural network has a five-layer network structure, including an input layer, a fuzzification layer, a fuzzy reasoning layer, a normalization layer and an output layer; The fuzzy neural network has a loss function: in, is the loss function, is the total number of samples, For the The target output of samples is Predict output for the model; The fuzzy neural network reverse update formula: in, are the network parameters in the model, For the The parameter value after iterations; is the learning rate; is the step size, that is, the distance of parameter update in each gradient direction.

5. The optimization method of an industrial washing machine control system based on an incremental fuzzy neural network according to claim 4 is characterized in that: The incremental sample update includes sample addition and sample deletion; The sample is appended with an iterative formula: in, To finally meet the washing requirements, is the number of washes, For the Prediction output, including wash time , washing intensity and detergent dosage ; The sample pruning includes time-based sample pruning and verification error-based sample pruning. The time-based sample pruning is used to delete expired data in a fixed time window, and the verification error-based sample pruning is used to delete unfavorable samples in the sample set.

6. The optimization method of an industrial washing machine control system based on an incremental fuzzy neural network according to claim 5, characterized in that: The specific steps of sample deletion based on test error are as follows: The unfavorable sample calculation formula is: in, is the loss function; is the training sample, ; To test the samples, ; is the optimal solution for the parameters in the model. The model of training samples, is defined as follows: H is the Hessian matrix, and the calculation formula is as follows: 。 7. The optimization method of an industrial washing machine control system based on an incremental fuzzy neural network according to claim 2, characterized in that: The central processing unit is designed by LabVIEW.

Images