Preface The tobacco processing line is crucial for ensuring consistent cigarette quality, and the leaf feeding process is one of the most critical and challenging steps in the tobacco processing. The accuracy of leaf feeding significantly impacts the sensory quality of cigarettes. The automatic control method for leaf feeding involves setting the liquid feeding ratio, processing it through an electronic control system, and adjusting the output liquid flow rate according to the amount of leaf material, thus achieving automatic control of leaf feeding. Currently, the key components of the electronic control system used in leaf feeding systems within the domestic cigarette industry are differential pressure transmitters and DR24 loop controllers. Introduction and Problem Analysis of Existing Feeding Systems Cigarette Factory Feeding System: The cigarette factory feeding system is a vital step on the tobacco processing production line. It requires the system to calculate the theoretical flavoring and feeding flow rate based on the actual flow rate of tobacco (leaf) measured by the belt scale on the production line and the required flavoring ratio coefficient, and then control the actual flavoring and feeding flow rate accordingly. A system diagram is shown in Figure 1. Figure 1; Schematic diagram of the flavoring and additive system in a cigarette factory. The original system consists of a measurement system composed of an electronic belt scale, a differential pressure transmitter, and a metering tank; an actuator composed of a pneumatic plunger pump and a pneumatic diaphragm regulating valve; and a loop controller DR24 as the controller. The working process of the additive system is as follows: When the system starts working, the plunger pump reciprocates, pressing the material into the pipeline. The one-way valve opens, and the liquid flows from the metering tank to the additive system. The differential pressure transmitter measures the change in pressure at the bottom of the metering tank per unit time and sends it to the controller DR24 to obtain the actual flavoring and additive flow rate. At the same time, the electronic belt scale inputs the measured tobacco flow signal into DR24. DR24 calculates the theoretical flavoring and additive amount based on the additive ratio and the belt scale signal. The DR24 controller performs PID calculations based on the deviation *e* between the theoretical and actual flow rates. The resulting digital valve position is output as an analog current signal, which is then converted into a pneumatic pressure signal by an electrical converter. This signal controls the opening of the pneumatic diaphragm valve, ensuring the feed rate varies with the tobacco leaf flow rate. This aims to match the actual flow rate as closely as possible to the theoretical flow rate, meeting the process requirements. System Problems In this feeding system, the quality of operation hinges on the accuracy of the differential pressure transmitter and the electronic belt scale. The differential pressure transmitter calculates the instantaneous flow rate based on changes in flow over a period. This method of flow measurement has a certain lag, leading to decreased control accuracy and inaccurate feed rates. Furthermore, due to the high concentration of the feed liquid, scaling over time can reduce accuracy. When the feed liquid flows through the differential pressure transmitter, particulate matter can compress the transmitter diaphragm, affecting accuracy. Additionally, the pipeline installation requirements are stringent; if residual liquid or sediment from the process pipeline flows into the pressure-conducting pipe, pressure measurement errors will occur. Furthermore, the DR24 controller's human-machine interface is simple and rudimentary, unable to intuitively display various operating conditions during the production process, nor can it display various parameters, alarms, and other information. Its network communication requires dedicated hardware and software, which is complex, difficult to maintain, and unsuitable for large-scale data transmission requirements, failing to adapt to current advanced fieldbus technology. New System Design Measurement Mechanism Selection: Considering the varying densities of different flavorings and sugars in cigarette factories, after analysis and research, it was decided to use a mass flow meter to measure flow rate instead of the original system's method of using a differential pressure transmitter and metering tank. The mass flow meter method is unaffected by density, bubbles, temperature, impurities, etc., and its accuracy reaches 0.1%, effectively solving the problem of poor system measurement accuracy. Controller Selection: This system needs to control more than twenty switching quantities and four analog quantities, perform various arithmetic, logical, and control operations, and requires strong communication capabilities to exchange large amounts of data via fieldbus. Furthermore, considering the high temperature, humidity, and dust levels in the industrial environment, a PLC was selected as the controller, with a touch screen as the human-machine interface. The system offers good maintainability, and the programming language is well-suited to the working habits of electrical maintenance personnel. The PLC selected conforms to IEC standards. The original system used a plunger pump as the power source for liquid flow, which resulted in pulsation. This pulsation interfered with the system's measurement and control calculations. Therefore, the improved system uses compressed air as the power source, which is readily available in the workshop and provides a very stable flow, making it easier to achieve the system's control objectives. The actual system uses a sealed metering tank (with a differential pressure transmitter) and an open storage tank. Before starting, the liquid must be pumped into the storage tank. A magnetic float level gauge next to the storage tank displays the liquid level; the tank can be filled up to 80% of its scale. When the liquid temperature is lower than the set temperature, the controller automatically opens a pneumatic valve to introduce steam into the storage tank's jacket to heat the liquid. Once the set temperature is reached, the pneumatic valve closes. When the liquid level in the metering tank reaches the set lower limit, the process controller automatically starts the feed pump motor to replenish the metering tank with liquid from the storage tank until the liquid level in the metering tank reaches the set upper limit, or until there is no liquid in the storage tank. A one-way valve allows the liquid to flow in one direction, but the pressure in the metering tank cannot be reversed. A constant air pressure is required in the metering tank, typically set at 0.5 MPa. The actuator that directly controls the liquid flow rate uses a linear pneumatic regulating valve with a parabolic valve core. The controller outputs the valve opening as a 4-20mA DC electrical signal to the electrical converter, which converts the electrical signal into a corresponding 0.02-0.1 MPa air pressure signal to drive the pneumatic regulating valve. Working Principle After the start of operation, the electronic scale sends the instantaneous flow rate of the tobacco leaves to the controller. After a delay, this is multiplied by a pre-set proportioning coefficient to obtain the theoretical instantaneous flow rate value. It is worth noting that the delay of the electronic scale is necessary because there is a time difference between the electronic scale and the nozzle; only with accurate delay can the dynamic response of the control be better achieved. The mass flow meter sends a feedback signal to the controller, representing the actual instantaneous flow rate of the liquid. The instantaneous value is compared with the setpoint, and the difference is used as the deviation for PID control. A 4-20mA current signal is output to the electrical converter, which converts this current signal into a 0.02-0.1MPa air pressure signal. This air pressure is used to activate the pneumatic regulating valve, linearly controlling the valve opening from 0-100%. The outflowing liquid enters the nozzle through the float flow meter and is atomized by compressed air onto the blades. In this system, the differential pressure transmitter is only used to measure the weight of the liquid in the metering tank. The working principle diagram is shown in Figure 2. System Implementation To ensure control accuracy and improve the system's real-time response, this system adds a feedforward stage based on the traditional PID control, taking into account conditions such as the blade (blade) flow rate, the gas pressure in the metering tank, and the liquid level in the metering tank. The system first quickly coarsely adjusts the valve opening to near the correct position based on the feedforward value, and then the PID control reaches the equilibrium position. The feedforward value calculation mimics a self-learning system. The controller memorizes 100 discrete state points, and each time control equilibrium is reached, the memorized value is partially corrected. This method effectively solves a significant problem affecting the accuracy of the flavoring and feeding system: the fluctuation of the material flow from the belt scale. To enable the system to have self-checking capabilities, this system uses a differential pressure transmitter located under the metering tank to periodically detect whether the mass flow meter is malfunctioning. This avoids accidents caused by malfunctions in the measuring device. In addition, relying on the relatively powerful programming capabilities of the PLC, this system, in addition to completing the main control work, also uses multiple photoelectric sensors and in-program logic judgments to monitor possible faults and drifts, greatly improving system reliability and accuracy. Figure 2: Working principle diagram of the new feeding system. Figure 3: Working principle diagram of the stack. Liquid temperature control: Temperature control is a large lag regulation. The control method adopted by this system uses the rate of temperature change for predictive control. That is, it judges the value of the liquid temperature in the tank after a period of time based on the rate of temperature rise or fall, thereby controlling the valve opening in advance and avoiding the temperature overshoot problem in traditional methods. Control difficulty: Delaying the instantaneous flow rate of the electronic scale is a necessary step. This system employs a stack-based approach. The instantaneous flow signal from the electronic scale is sent to the PLC of the feeding system, and then processed by the stack (using a first-in, first-out) method with a delay. The principle is as follows: a stack can hold a maximum of 100 variables. Based on the clock pulse signal, these 100 variables are stored in the stack in sequence. When the stack is full, the first variable stored is output first, following the first-in, first-out principle. The length of the clock pulse can be adjusted as needed; here, a second pulse is used, with a maximum delay of 100 seconds, which is sufficient for this system. After stack processing, the signal becomes the delayed instantaneous flow signal from the electronic scale. The stack working principle is shown in Figure 3. Improvements: The new feeding system uses a mass flow meter to measure the feeding flow, significantly improving accuracy and meeting national standards. The user-friendly interface and communication functions allow the workshop control room to monitor production in real time. Compressed air is used as the power source for supplying the liquid material, replacing the original plunger pump, making it economical and reliable. After more than a year of use, the entire system has been operating stably.