Ink jet control system for industrial printer based on negative feedback control

Abstract

The invention discloses an industrial printer ink-jet control system based on negative feedback control, and relates to the technical field of industrial printers. The ink-jet control system for the industrial printer based on negative feedback control comprises a data input processing module which is responsible for receiving printing data input by a user, converting the printing data into a format which can be identified by the printer, and carrying out preliminary optimization and arrangement on the data at the same time; the data input processing module is connected with a high-speed data transmission module, and the high-speed data transmission module is connected with a sensor monitoring module, a feedback signal processing module, an optimization algorithm control module and a distributed local control module. The calculated amount can be reduced and the time required by deviation calculation can be shortened by simplifying the deviation calculation algorithm, and the deviation can be predicted and adjusted in advance by the predictive control algorithm, so that the system can quickly cope with the change of the ink jet state, and the printing efficiency and quality are improved.

Description

Technical Field

[0001] The invention relates to the technical field of industrial printers, and in particular to an industrial printer inkjet control system based on negative feedback control. Background Art

[0002] An industrial printer is a professional printing device designed for large-scale, high-efficiency, and high-precision printing in industrial production environments. It is very different from ordinary home or office printers in terms of function, performance, and application scenarios.

[0003] In the field of industrial printing, the performance of the inkjet control system directly affects the printing quality and efficiency. Although the traditional industrial printer inkjet control system based on negative feedback control can achieve inkjet control to a certain extent, it has many shortcomings. First, the hardware cost is high. The various high-precision sensors, high-performance nozzle drive circuits and control hosts equipped to achieve precise control have greatly increased the system cost, limiting its application in some cost-sensitive industries. Secondly, the system response speed is limited. There is a time delay in the process of inkjet monitoring, feedback signal processing, and deviation calculation and adjustment. In high-speed printing scenarios, it is difficult to meet the strict requirements for printing accuracy. In addition, the adaptability to complex environments is insufficient. In extreme environments such as high temperature, high humidity or strong electromagnetic interference, the sensor accuracy decreases, the signal transmission is disturbed, and the control accuracy is reduced. Moreover, the system maintenance and debugging are difficult. Multiple complex modules and a large number of sensors make troubleshooting and debugging difficult, and the maintenance cost is high.

[0004] Therefore, it is of great practical significance to develop an industrial printer inkjet control system based on negative feedback control with fast response speed, reasonable cost, strong environmental adaptability and easy maintenance. Summary of the invention

[0005] In view of the defects in the prior art, the present invention provides an inkjet control system for an industrial printer based on negative feedback control, including a data input processing module, the data input processing module is responsible for receiving the printing data input by the user, and converting it into a format recognizable by the printer, and performing preliminary optimization and sorting on the data, the data input processing module is connected to a high-speed data transmission module, the high-speed data transmission module is connected to a sensor monitoring module, a feedback signal processing module, an optimization algorithm control module and a distributed local control module, the optimization algorithm control module and the distributed local control module are connected to a nozzle driving module;

[0006] The nozzle driving module is connected to the optimization algorithm control module and the distributed local control module, receives control signals from the two modules, drives the nozzle to perform inkjet operation according to the signals, and feeds back the working status information of the nozzle to the feedback signal processing module;

[0007] The sensor monitoring module is installed at the key position of the nozzle and the ink supply pipeline to collect inkjet related data in real time, and the collected data is transmitted to the feedback signal processing module through the high-speed data transmission module;

[0008] The high-speed data transmission module acts as a bridge for data transmission, connecting various modules that require data interaction, ensuring fast and accurate data transmission between different modules, including sensor data, control signals and print data;

[0009] The feedback signal processing module receives data from the sensor monitoring module, processes it, and sends effective feedback information to the optimization algorithm control module and the distributed local control module;

[0010] The optimization algorithm control module receives the printing data from the data input and preprocessing module on the one hand, and receives the feedback information from the feedback signal processing module on the other hand, calculates the control signal by running the optimization algorithm, and sends it to the nozzle driving module and the distributed local control module, and at the same time, exchanges data with the distributed local control module to collaboratively complete the control of the inkjet process;

[0011] The distributed local control module is connected to the nozzle driving module, the feedback signal processing module and the optimization algorithm control module, performs partial monitoring and control operations on the nozzle locally, feeds back the processing results to the optimization algorithm control module, and receives instructions from the optimization algorithm control module to collaboratively complete the inkjet control task;

[0012] The high-speed data transmission module includes data compression and priority division, high-speed interface and protocol;

[0013] The optimization algorithm control module includes a simplified deviation calculation algorithm and a predictive control algorithm;

[0014] Preferably, the simplified deviation calculation algorithm includes:

[0015] Set reference values. During the system initialization phase, according to the requirements of the printing task, pre-set the ideal inkjet parameters, such as the ink flow setting value Q set , droplet size standard value S set , nozzle position target value (x set ,y set );

[0016] Simplified calculation: Abandoning the complex high-order calculations and redundant steps in the traditional algorithm, when calculating the ink flow deviation, directly use a simple difference calculation, the formula is: flow deviation ΔQ = Q actual -Q set , where Q actualis the actual flow value. For the nozzle position deviation (two-dimensional plane coordinates), the simplified formula is: D x =x actual -x set , D y =y actual -y set ;

[0017] Weighted processing: Considering the different effects of different parameters on printing quality, the deviation of each parameter will be weighted. Assume that there are m deviation parameters, namely ΔP 1 ,ΔP 2 ,…,ΔP m , the corresponding weight is w 1 ,w 2 ,…,w m ,and The comprehensive deviation ΔP total for:

[0018]

[0019] Preferably, the predictive control algorithm includes historical data processing, prediction model establishment and predictive control implementation, and the prediction model establishment includes:

[0020] Time series analysis model (ARIMA):

[0021] Φ(B)(1-B) d Y t =Θ(B)∈ t

[0022] Among them, Y t is the value of the time series at time t, and B is the backward shift operator, namely BY t =Y t-1 , Φ(B) is the polynomial of the autoregressive part, Φ(B)=1-φ 1 B-φ 2 B 2 -…-φ p B p , p is the autoregressive order; (1-B) d is a difference operator used to stabilize a non-stationary time series, d is the difference order; Θ(B) is the polynomial of the sliding average part, Θ(B) = 1 + θ 1 B+θ 2 B 2 +…+θ q B q , q is the sliding average order; ∈ tIt is a white noise sequence. In the inkjet control system of industrial printers, we regard the changes of parameters such as ink flow and nozzle position over time as time series. By determining the appropriate p, d, and q values, we establish an ARIMA model to predict the future inkjet status.

[0023] Neural Network Model:

[0024] Assume that the input layer has n neurons, corresponding to n input features (such as ink flow and nozzle position at previous moments); the hidden layer has m neurons; the output layer has k neurons, corresponding to k predicted inkjet deviation parameters, and the weight matrix from the input layer to the hidden layer is W 1 , the weight matrix from the hidden layer to the output layer is W 2 , the activation function of the hidden layer is σ (such as ReLU function: σ(x)=max(0,x)),

[0025]

[0026] Among them, X is the input vector, H is the hidden layer output vector, is the predicted output vector, b 1 and b 2 are the bias vectors of the hidden layer and the output layer respectively.

[0027] Preferably, the simplified deviation calculation result and the prediction result of the predictive control algorithm are combined to generate a control signal, and an improved PID control algorithm is used, and the formula is:

[0028]

[0029] Among them, K p is the proportionality coefficient, K i is the integration coefficient, K d is the differential coefficient, K f is the prediction coefficient, e(t) is the current deviation, and f(t) is the predicted deviation.

[0030] Working principle: First, various high-speed sensors collect inkjet-related analog signal data such as ink flow, droplet size and speed, and nozzle position in real time. These data are transmitted to the high-speed data acquisition card and converted into digital signals through a high-performance analog-to-digital converter. Then, a fast filtering algorithm, such as an improved fast mean filtering algorithm, is used to remove noise from the digital signal, ensure the accuracy and integrity of the data, and provide a reliable data basis for subsequent processing;

[0031] When the system is initialized, the ink flow setting value Q is set according to the printing task set , droplet size standard value S set , nozzle position target value (x set ,y set) and other ideal inkjet parameters. During operation, the traditional complex high-order calculations and redundant steps are abandoned, and the ink flow deviation ΔQ=Q is calculated by simple difference. actual -Q set (Q actual is the actual flow value), and the coordinate difference D is directly calculated for the nozzle position deviation (two-dimensional plane coordinates) x =x actual -x set , D y =y actual -y set Then, according to the influence of different parameters on the printing quality, weights are assigned and the weighted sum is calculated. (m is the number of deviation parameters, w i is the weight, ΔP i is each deviation parameter) to obtain the comprehensive deviation;

[0032] During this period, a large amount of historical inkjet data was collected and sorted, and data mining technology was used to analyze the changing trends of nozzle working frequency, ink flow, etc. under different printing tasks. According to the data characteristics, a time series analysis model (ARIMA) or a neural network model (such as a multi-layer perceptron MLP) was selected to establish a prediction model. Taking the ARIMA (p, d, q) model as an example, Φ(B)(1-B) d Y t =Θ(B)∈ t By determining the parameters such as the autoregressive order p, the differential order d, and the sliding average order q, the system can predict the future state based on the historical inkjet state. During the operation of the system, each time new feedback data is received, the current deviation is calculated and the prediction model is used to predict the future deviation, so as to adjust the nozzle drive parameters in advance.

[0033] Finally, the simplified deviation calculation results and the prediction results of the predictive control algorithm are combined to generate the control signal using the improved PID control algorithm. The formula is: Where K p is the proportionality coefficient, K i is the integration coefficient, K d is the differential coefficient, K f is the prediction coefficient, e(t) is the current deviation, and f(t) is the prediction deviation. The algorithm is used to adjust the drive voltage, current and other parameters of the nozzle drive circuit to achieve precise control of the inkjet process.

[0034] The beneficial effects of the present invention are embodied in:

[0035] By simplifying the deviation calculation algorithm, the amount of calculation is reduced and the deviation calculation time is shortened; the predictive control algorithm predicts the deviation in advance and makes adjustments, so that the system can respond quickly to changes in the inkjet state and ensure high precision even at high-speed printing, thereby improving printing efficiency and quality.

[0036] The optimization algorithm reduces the dependence on some ultra-high precision hardware, reduces hardware costs and improves the market competitiveness of the product while ensuring performance.

[0037] Fast data processing and precise control capabilities enable the system to resist interference from environmental factors on sensors and signal transmission to a certain extent, and maintain stable printing performance in complex environments.

[0038] The optimized algorithm and system architecture make troubleshooting and parameter adjustment more convenient, reduce maintenance costs and technical barriers, and improve system availability and stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for the specific embodiments or the description of the prior art. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn according to the actual scale.

[0040] Figure 1 It is a schematic diagram of the system flow of the present invention. DETAILED DESCRIPTION

[0041] The following embodiments of the technical solution of the present invention are described in detail in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and are therefore only used as examples, and cannot be used to limit the protection scope of the present invention.

[0042] It should be noted that, unless otherwise specified, the technical terms or scientific terms used in this application should have the common meanings understood by those skilled in the art to which the present invention belongs.

[0043] like Figure 1 As shown, an embodiment of the present invention provides an inkjet control system for an industrial printer based on negative feedback control, including a data input processing module, which is responsible for receiving the printing data input by the user and converting it into a format recognizable by the printer, and performing preliminary optimization and sorting on the data, the data input processing module is connected to a high-speed data transmission module, the high-speed data transmission module is connected to a sensor monitoring module, a feedback signal processing module, an optimization algorithm control module and a distributed local control module, and the optimization algorithm control module and the distributed local control module are connected to a nozzle driving module;

[0044] Specifically, the user inputs the printing data into the data input and preprocessing module through a computer or other device. The module converts the data into a format recognizable by the printer and performs preliminary optimization, and then sends it to the optimization algorithm control module through the high-speed data transmission module. The multi-type high-speed sensor monitoring module collects the data related to the nozzle and ink in real time, and transmits the data to the fast feedback signal processing module with the help of the high-speed data transmission module. The processed effective feedback information is then transmitted to the optimization algorithm control module and the distributed local control module. The optimization algorithm control module integrates the printing data and feedback information, uses the simplified deviation calculation algorithm and the predictive control algorithm, calculates the control signal, and sends it to the high-speed nozzle drive module and the distributed local control module. The distributed local control module performs some monitoring and control operations near the nozzle, and its processing results are fed back to the optimization algorithm control module, while receiving instructions from the module. The high-speed nozzle drive module drives the nozzle to spray ink according to the control signal, and feeds back the nozzle working status information to the fast feedback signal processing module. Finally, through the high-speed data transmission module, the data transmission between modules is fast and accurate, which greatly improves the system response speed; the optimization algorithm control module combines multiple algorithms to accurately calculate the deviation and predict it in advance to ensure accurate inkjet control; the distributed local control module cooperates with the central control to reduce data transmission delays and improve real-time control; multiple types of high-speed sensor monitoring modules and fast feedback signal processing modules work closely together to obtain and process inkjet status information in a timely and accurate manner, providing a basis for precise control. In summary, the coordinated operation of each module has significantly improved the printing accuracy, speed and stability of the entire system, and can better meet the high requirements of industrial printing.

[0045] The nozzle drive module is connected to the optimization algorithm control module and the distributed local control module, receives control signals from the two modules, drives the nozzle to perform inkjet operation according to the signals, and feeds back the working status information of the nozzle to the feedback signal processing module. The sensor monitoring module is installed at the key position of the nozzle and the ink supply pipeline to collect inkjet related data in real time. The collected data is transmitted to the feedback signal processing module through the high-speed data transmission module; the high-speed data transmission module serves as a bridge for data transmission, connecting various modules that require data interaction, ensuring fast and accurate data transmission between different modules, including sensor data, control signals and printing data. The high-speed data transmission module includes data compression and priority division, high-speed interface and protocol;

[0046] Specifically, the USB3.0 high-speed data transmission interface and protocol are adopted. With extremely high data transmission rate, the theoretical transmission speed of USB3.0 can reach 5Gbps, and the transmission speed of Ethernet in a gigabit network environment can reach 1000Mbps. Compared with the traditional low-speed interface, the data transmission speed has been improved by an order of magnitude. In terms of sensor data transmission, it may take tens of milliseconds or even hundreds of milliseconds to transmit a frame of sensor data using a low-speed interface. After adopting a high-speed interface, the transmission time can be shortened to a few milliseconds or even shorter, ensuring that the real-time inkjet status data collected by the sensor can be quickly transmitted to the control host, providing a data basis for timely adjustment of the control strategy. In addition, the transmitted data is reasonably compressed, and an efficient data compression algorithm is used to greatly reduce the size of the original data, thereby reducing the amount of data transmission and further improving the transmission speed. At the same time, the data is prioritized, and the key data directly related to the inkjet control, including nozzle position feedback, ink flow monitoring data, etc., are set to high priority to ensure that these data are sent and received first during the transmission process, avoiding transmission delays caused by data congestion, ensuring the timely transmission of control signals and key feedback information, and effectively improving the response speed of the system.

[0047] The feedback signal processing module receives data from the sensor monitoring module, processes it, and sends effective feedback information to the optimization algorithm control module and the distributed local control module; the optimization algorithm control module receives the printing data from the data input and preprocessing module on the one hand, and receives feedback information from the feedback signal processing module on the other hand, calculates the control signal by running the optimization algorithm, and sends it to the nozzle driving module and the distributed local control module, and at the same time, exchanges data with the distributed local control module to collaboratively complete the control of the inkjet process. The optimization algorithm control module includes a simplified deviation calculation algorithm and a predictive control algorithm;

[0048] Specifically, in the system initialization stage, according to the requirements of the printing task, the ideal inkjet parameters are pre-set, such as the ink flow setting value Q set , droplet size standard value S set , nozzle position target value (x set ,y set ) etc. These reference values ​​serve as a benchmark for judging whether the actual inkjet state is accurate.

[0049] Simplified calculation: Abandon the complex high-order operations and redundant steps in traditional algorithms. For example, when calculating the ink flow deviation, directly use a simple difference calculation, the formula is: flow deviation ΔQ = Q actual -Q set , where Q actual is the actual flow value. For the nozzle position deviation (two-dimensional plane coordinates), the simplified formula is: D x=x actual -x set , D y =y actual -y set , directly calculate the coordinate difference, no longer perform complex Euclidean distance calculation, greatly reducing the amount of calculation while meeting the printing accuracy requirements. Weighted processing: Considering the different degrees of influence of different parameters on printing quality, the deviation of each parameter will be weighted. Assume that there are m deviation parameters, namely ΔP 1 ,ΔP 2 ,…,ΔP m , the corresponding weight is w 1 ,w 2 ,…,w m ,and The comprehensive deviation ΔP total for:

[0050]

[0051] Predictive control algorithms

[0052] Historical data processing: Collect a large amount of historical inkjet data under different printing tasks, covering parameters such as ink flow, nozzle position, droplet size, etc. Use data cleaning technology to remove outliers and noise interference to ensure the accuracy and reliability of the data. Then normalize the data so that the data of different parameters are at the same order of magnitude, which is convenient for subsequent analysis and modeling.

[0053] Prediction model building

[0054] Time series analysis model (ARIMA):

[0055] Principle: The full name of ARIMA model is Autoregressive Integrated Moving Average model, which is based on the autocorrelation of time series data and predicts future values ​​by analyzing historical data. The model consists of three parts: autoregression (AR), difference (I) and moving average (MA).

[0056] Formula: The mathematical expression of the ARIMA (p, d, q) model is:

[0057] Φ(B)(1-B) d Y t =Θ(B)∈ t

[0058] Among them, Y t is the value of the time series at time t, and B is the backward shift operator, namely BY t =Y t-1 , Φ(B) is the polynomial of the autoregressive part, Φ(B)=1-φ 1 B-φ 2B 2 -…-φ p B p , p is the autoregressive order; (1-B) d is a difference operator used to stabilize a non-stationary time series, d is the difference order; Θ(B) is the polynomial of the sliding average part, Θ(B) = 1 + θ 1 B+θ 2 B 2 +…+θ q B q , q is the sliding average order; ∈ t is a white noise sequence. In the inkjet control system of industrial printers, we regard the changes of parameters such as ink flow and nozzle position over time as time series, and establish an ARIMA model to predict the future inkjet state by determining the appropriate p, d, and q values.

[0059] Neural Network Model:

[0060] Principle: Multilayer Perceptron is a feedforward neural network consisting of input layer, hidden layer and output layer. It learns complex patterns and rules in data through the connection weights between a large number of neurons. In predictive control, historical inkjet data is used as input, and the predicted value of future inkjet deviation is output through nonlinear transformation of hidden layer.

[0061] Structure and formula: Assume that the input layer has n neurons, corresponding to n input features (such as ink flow, nozzle position, etc. at the previous moments); the hidden layer has m neurons; the output layer has k neurons, corresponding to k predicted inkjet deviation parameters. The weight matrix from the input layer to the hidden layer is W 1 , the weight matrix from the hidden layer to the output layer is W 2 The activation function of the hidden layer is σ (such as ReLU function: σ(x) = max(0,x)).

[0062]

[0063] Among them, X is the input vector, H is the hidden layer output vector, is the predicted output vector, b 1 and b 2 are the bias vectors of the hidden layer and the output layer respectively. Through training with a large amount of historical data, the weight matrix W is continuously adjusted. 1 and W 2 and the bias vector b 1 and b 2 , enabling the model to accurately predict inkjet deviation.

[0064] Predictive control implementation: Integrate the established prediction model into the inkjet control system. Every time new feedback data is received, not only the current deviation is calculated, but also the prediction model is used to predict the deviation that may occur in the future. According to the prediction results, the drive parameters of the nozzle are adjusted in advance to achieve advance control of the inkjet process and effectively shorten the response time of the system.

[0065] The distributed local control module is connected to the nozzle driving module, the feedback signal processing module and the optimization algorithm control module, performs partial monitoring and control operations on the nozzle locally, feeds back the processing results to the optimization algorithm control module, and receives instructions from the optimization algorithm control module to collaboratively complete the inkjet control task;

[0066] Specifically, various high-speed sensors collect inkjet-related data in real time, high-speed flow sensors collect ink flow data at a high frequency, high-speed photoelectric sensors quickly capture droplet size and speed information, and high-speed position sensors accurately obtain the nozzle position coordinates. These sensors output data in the form of analog signals, which are transmitted to a high-speed data acquisition card and converted into digital signals through a high-performance analog-to-digital converter (ADC) for processing by the control host. During the conversion process, the appropriate number of sampling bits and sampling frequency are set according to the accuracy of the sensor and system requirements to ensure the accuracy and integrity of the data. The collected digital signals contain noise and are processed using a fast filtering algorithm. Finally, the control signal is generated by combining the simplified deviation calculation results and the prediction results of the predictive control algorithm. The improved PID control algorithm is adopted, and the formula is:

[0067]

[0068] Among them, K p is the proportionality coefficient, K i is the integration coefficient, K d is the differential coefficient, K f is the prediction coefficient, e(t) is the current deviation, and f(t) is the prediction deviation. By introducing the prediction coefficient K f and predicted deviation f(t), so that the control signal can respond to possible deviations in the future more promptly, further improving the control accuracy and response speed of the system. In practical applications, the integral term is usually discretized and approximated by accumulating past deviations; the differential term is approximated by the difference between the current deviation and the deviation at the previous moment.

[0069] Finally, 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 aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. These modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and specification of the present invention.

Claims

1. An inkjet control system for an industrial printer based on negative feedback control, comprising a data input processing module, which is responsible for receiving the print data input by the user and converting it into a format recognizable by the printer, while performing preliminary optimization and sorting of the data, characterized in that: The data input processing module is connected to a high-speed data transmission module, the high-speed data transmission module is connected to a sensor monitoring module, a feedback signal processing module, an optimization algorithm control module and a distributed local control module, and the optimization algorithm control module and the distributed local control module are connected to a nozzle driving module; The nozzle driving module is connected to the optimization algorithm control module and the distributed local control module, receives control signals from the two modules, drives the nozzle to perform inkjet operation according to the signals, and feeds back the working status information of the nozzle to the feedback signal processing module; The sensor monitoring module is installed at the key position of the nozzle and the ink supply pipeline to collect inkjet related data in real time, and the collected data is transmitted to the feedback signal processing module through the high-speed data transmission module; The high-speed data transmission module acts as a bridge for data transmission, connecting various modules that require data interaction, ensuring fast and accurate data transmission between different modules, including sensor data, control signals and print data; The feedback signal processing module receives data from the sensor monitoring module, processes it, and sends effective feedback information to the optimization algorithm control module and the distributed local control module; The optimization algorithm control module receives the printing data from the data input and preprocessing module on the one hand, and receives the feedback information from the feedback signal processing module on the other hand, calculates the control signal by running the optimization algorithm, and sends it to the nozzle driving module and the distributed local control module, and at the same time, exchanges data with the distributed local control module to collaboratively complete the control of the inkjet process; The distributed local control module is connected to the nozzle driving module, the feedback signal processing module and the optimization algorithm control module, performs partial monitoring and control operations on the nozzle locally, feeds back the processing results to the optimization algorithm control module, and receives instructions from the optimization algorithm control module to collaboratively complete the inkjet control task; The high-speed data transmission module includes data compression and priority division, high-speed interface and protocol; The optimization algorithm control module includes a simplified deviation calculation algorithm and a predictive control algorithm.

2. According to claim 1, an industrial printer inkjet control system based on negative feedback control is characterized in that: The simplified deviation calculation algorithm includes: Set reference values. During the system initialization phase, according to the requirements of the printing task, pre-set the ideal inkjet parameters, such as the ink flow setting value Q set , droplet size standard value S set , nozzle position target value (x set ,y set ); Simplified calculation: Abandon the complex high-order calculations and redundant steps in the traditional algorithm, and directly use simple difference calculation when calculating the ink flow deviation. The formula is: flow deviation ΔQ = Q actual -Q set , where Q actual is the actual flow value. For the nozzle position deviation, the simplified formula is: D x =x actual -x set , D y =y actual -y set ; Weighted processing: Considering the different effects of different parameters on the printing quality, the deviation of each parameter will be weighted. Assume that there are m deviation parameters, namely ΔP1, ΔP2, …, ΔP m , the corresponding weights are w1,w2,…,w m ,and The comprehensive deviation ΔP total for:

3. The inkjet control system for industrial printers based on negative feedback control according to claim 1, characterized in that: The predictive control algorithm includes historical data processing, prediction model establishment and predictive control implementation, and the prediction model establishment includes: Time Series Analysis Model: Φ(B)(1-B) d Y t =Θ(B)∈ t Among them, Y t is the value of the time series at time t, and B is the backward shift operator, namely BY t =Y t-1 , Φ(B) is the polynomial of the autoregressive part, Φ(B)=1-φ1B-φ2B 2 -…-φ p B p , p is the autoregressive order; (1-B) d is a difference operator used to stabilize a non-stationary time series, d is the difference order; Θ(B) is the polynomial of the sliding average part, Θ(B) = 1 + θ1B + θ2B 2 +…+θ q B q , q is the sliding average order; ∈ t It is a white noise sequence. In the inkjet control system of industrial printers, we regard the changes of parameters such as ink flow and nozzle position over time as time series. By determining the appropriate p, d, and q values, we establish an ARIMA model to predict the future inkjet status. Neural Network Model: Assume that the input layer has n neurons, corresponding to n input features; the hidden layer has m neurons; the output layer has k neurons, corresponding to k predicted inkjet deviation parameters, the weight matrix from the input layer to the hidden layer is W1, the weight matrix from the hidden layer to the output layer is W2, and the activation function of the hidden layer is σ, Among them, X is the input vector, H is the hidden layer output vector, is the predicted output vector, b1 and b2 are the bias vectors of the hidden layer and the output layer respectively.

4. The inkjet control system for industrial printers based on negative feedback control according to claim 1, characterized in that: Combine the simplified deviation calculation results and the prediction results of the predictive control algorithm to generate a control signal, using an improved PID control algorithm, the formula is: Among them, K p is the proportionality coefficient, K i is the integration coefficient, K d is the differential coefficient, K f is the prediction coefficient, e(t) is the current deviation, and f(t) is the predicted deviation.

Images