Đang tải... (xem toàn văn)
đồ án này chúng mình nghiên cứ dưới sự hướng dẫn của tiến sĩ Nguyễn Thu Hà giúp mô phỏng ổn định đầu ra cho hệ two input two output khử tương tác giữa 2 kênh nhiệt độ và độ ẩm từ đó tối ưu hóa bộ điều khiển cho lồng ấp trẻ sơ sinh. Mình đăng bài này lên để giúp đỡ các bạn hiểu sâu hơn và yêu thích bộ môn lý thuyết điều khiển tuyến tính
PROJECT I REPORT TOPIC: MICROCONTROLLER SYSTEM FOR NEWBORN INCUBATOR Instructor: Dr NGUYEN THU HA Do Huy Hoang 20176978 Nguyen Tuan Hung 20176941 Class: CTTT- HTĐ.TĐH-K62 Preface Firstly, we would like to say thanks to Dr Nguyen Thu Ha for her help in this project She provided us necessary knowledge about control algorithm, and created favorite condition for us to complete this this project In addition, we greatly appreciate her guidance through this project Through carrying out the research requirements, we found automatic control is very interesting Automatic control design technology technology is developing considerably in Vietnam One of the widespread method is PID In this project, we use Matlab & Simulink software to survey and design each block of control algorithm The software allows us to adjust the parameters of every component to achieve the desired result Also, we can create subcircuit block to pack each block with respective function Sincerely! Ha Noi, 17/7/2020 Group of students NGUYEN TUAN HUNG DO HUY HOANG CONTENTS CHAPTER INTRODUCTION 1.1 The important role of newborn incubator 1.2 Aim of project 1.3 Overview CHAPTER 2.HARDWARE DESIGN 2.1 Microcontroller (Arduino Uno R3) 2.2 Temperature Measuring Hardware 10 2.3 Temperature Controlling Hardware 12 2.4 Humidity measuring hardware 13 CHAPTER TRANSFER FUCTION CALCULATION AND CONTOLLER DESIGN 16 3.1 Identify the parameter of transfer function 16 3.2 Design the controller system 21 3.3 A decoupling controller using compensator controller 25 3.4 A new decoupling controller for multi-time delayed TITO processes 27 3.5 Select controller parameter 31 CONCLUSION 39 REFERENCE 40 Table of Figure: Page Figure 1: The newborn incubator……………………………………….7 Figure 2: ARDUINO……………………………………………………9 Figure 3: ARDUINO datasheet…………………………………………10 Figure 4: PT100 sensor………………………………………………….11 Figure 5: MAX 31865…………… ……………………………………11 Figure 6: Connect to measure temperature………………………………12 Figure 7: SSR…………………………………………………………….13 Figure 8: Connection of SSR and heated bar with ARDUINO………….13 Figure 9: Humidity sensor DHT22……………………………………….14 Figure 10: DHT22 connect to ARDUINO……………………………… 14 Figure 11: Servo MG996R……………………………… 14 Figure 12: Servo connect to ARDUINO…………………………… … 15 Figure 13: Structure diagram of microclimate object…………………….16 Figure 14: Temperature response when an input voltage is applied…… 17 Figure 15: Zoom in on the first graph…………………………………….18 Figure 16: Humidity response when voltage is applied………………… 18 Figure 17: Zoom in on the first graph…………………………………….19 Figure 18: When changing the valve opening% we get the characteristic line of the temperature…………………………………………………….19 Figure 19: When changing the valve opening% we get the characteristic line of the humidity……………………………………………………….20 Figure 20: PID controller…………………………………………………22 Figure 21: Block diagram of temperature channel……………………… 23 Figure 22: Block diagram of humidity channel………………………… 23 Figure 23: Block diagram of PI controller for TITO processes in MATLAB Simulink………………………………………………………………….24 Figure 24: Channel response of temperature and humidity…………… 24 Figure 25: System simulation diagram using compensator controller… 25 Figure 26: MATLAB simulation of control system with compensator controller…………………………………………………………………26 Figure 27: Channel response of temperature and humidity with compensation…………………………………………………………… 26 Figure 28: MATLAB demonstrated simulation for decoupling control….30 Figure 29: Step response with decoupling controller…………………… 30 Figure 30: Step response with decoupling controller after calibration… 32 Figure 31: The first response graph when not yet calibration…………….32 Figure 32: The first response graph when calibrated…………………… 33 Figure 33: Step response with decoupling controller after calibration… 34 Figure 34: The first response graph when not yet calibration…………….34 Figure 35: The first response graph when calibrated…………………… 35 Figure 36: Step response when changing parameter c21 , T′21 , σ21 ………36 Figure 37: Step response when changing c22 …………………………… 36 Figure 38: Step response when changing T′22 ……………………………37 Figure 39: Step response when adding σ22 ……………………………….37 Figure 40: Step response after select controller parameter……………….38 CHAPTER INTRODUCTION 1.1 The important role of newborn incubator The first idea of the newborn incubator was formed in France in 1878 and was immediately patented before 1890 However until 1891, it was perfectly completed and can be used by a American inventor- Edward Brown The advent of the newborn incubator has brought about the life for million of infants around the world And nowadays, with the development of the science and technology, the newborn incubator are being improved and much more completed The newborn incubator is an essential and neccessary medical equipment, the main purpose of this incubator is to keep the stability of the infant body’s temperature at a suitable one for living and growing Additionally, some of the modern incubators are equiped with system to regulate humidity and oxygen concentration and other treating equipment: ultraviolet lamp for treating jaundice, electronic scale,… Some incubator of Parker healthcare: Figure 1: The newborn incubator 1.2 Aim of project Our problem is to apply the theory and calculation into the real system, which need high accuracy and reliability The quantity of each component needed to be adjust to meet the requirements 1.3 Overview The newborn incubator is designed base on parameters: temperature (oC), humidity (%) And between them, temperature is the most important one, which requires high accuracy Besides, due to the neccessity for safety and reliability, the incubator also has back-up system for error to happen The below figure demonstrates the working principle: Change temperature Create Humidity The incubator Sensor system Controller Display Alarm system CHAPTER 2.HARDWARE DESIGN 2.1 Microcontroller (Arduino Uno R3) To control the system, we will use KIT Arduino Uno R3 Figure 2: ARDUINO Figure 3: ARDUINO datasheet 2.2 Temperature Measuring Hardware a) Temperature sensor PT100 “Pt” stands for Platinum resistance thermometer Because platinum has property that the resistance of it change accurately according to the change of temperature, so it is used widely in resistance thermometer The working principle of PT100 are very simple, which base on the relationship between metal and temperature When the temperature increase, the resistance of metal increase Platinum has the same property According to the standard, when the temperature is 0oC, the resistance of PT100 is 100Ω 10 Figure 26: MATLAB simulation of control system with compensator controller Figure 27: Channel response of temperature and humidity with compensation Simulation of the system with the controller parameters, we see, the results are of good quality When we change the value, the interaction effect between the channels decreases However, when the humidity channel changes, the temperature channel is still affected Improve controller quality 26 by design a new decoupling control method applied for two input two output processes (TITO) with multi-time delays 3.4 A new decoupling controller for multi-time delayed TITO processes Consider an interacted multi-time delayed TITO process described by the following transfer matrix: W=( 𝑊11 𝑊21 𝑊12 ) 𝑊22 The aim of decoupling problem here is an open loop controller: 𝑅 R=( 11 𝑅21 𝑅12 ) 𝑅22 The whole system becomes unconnectedly: 𝑅 𝑅 𝑊 𝑊 𝑅 𝑊 +𝑅 𝑊 𝑅 𝑊 +𝑅 𝑊 G=RW=(𝑅11 𝑅12 ) (𝑊11 𝑊12 ) = (𝑅11 𝑊11 + 𝑅12 𝑊21 𝑅11 𝑊12 + 𝑅12 𝑊22 ) 21 22 21 22 21 11 22 21 21 12 22 22 𝐺 =( 0 ) 𝐺2 Based on this equation given above the purpose of decoupling control could be exposed that all elements 𝑅𝑖𝑗 of decoupling controller R are to be determined, so that they satisfy: 𝑅 𝑊 +𝑅 𝑊 =0 𝑅 𝑊 +𝑅 𝑊 =𝐺 {𝑅11 𝑊12 + 𝑅12 𝑊22 = 21 11 22 21 {𝑅11 𝑊11 + 𝑅12 𝑊21 = 𝐺1 21 12 22 22 𝑊11 = 𝑘11 𝑒 −𝜏11𝑠 + 𝑇11 𝑠 𝑊21 = 𝑘21 𝑒 −𝜏21𝑠 + 𝑇21 𝑠 𝑊12 = 𝑘12 𝑒 −𝜏12𝑠 + 𝑇12 𝑠 𝑊22 = 𝑘22 𝑒 −𝜏22𝑠 + 𝑇22 𝑠 (1) (2) 27 Suppose that all elements 𝑅𝑖𝑗 of decoupling controller are time delayed lead/lags as follows 𝑅11 = 𝑐11 (1 + 𝑇12 𝑠) −𝜎 𝑠 𝑒 11 + 𝑇 ′11 𝑠 𝑅12 = 𝑐12 (1 + 𝑇22 𝑠) −𝜎 𝑠 𝑒 12 + 𝑇 ′12 𝑠 𝑅21 = 𝑐21 (1 + 𝑇11 𝑠) −𝜎 𝑠 𝑒 21 + 𝑇 ′ 21 𝑠 𝑅22 = 𝑐22 (1 + 𝑇21 𝑠) −𝜎 𝑠 𝑒 22 + 𝑇 ′ 22 𝑠 𝑐12 𝑘22 + 𝑐11 𝑘12 = 𝑘 𝑐12 = −𝑐11 12 𝑐12 𝑘22 𝑇′11 + 𝑐11 𝑘12 𝑇′12 = 𝑘22 (1) ↔ { ↔{ 𝑘 𝑐22 𝑘21 + 𝑐21 𝑘11 = 𝑐21 = −𝑐22 21 { 𝑘11 𝑐22 𝑘21 𝑇′21 + 𝑐21 𝑘11 𝑇′22 = { 𝑐12 = −𝑐1 𝑘12 𝑘 22 ↔ { 𝑘21 𝑐21 = −𝑐2 choose 𝑐1 = 𝑘22 ; 𝑐2 = 𝑘11 ; 𝑐12 = −𝑘12 ; 𝑐21 = 𝑘11 −𝑘21 𝜎 +𝜏 =𝜎 +𝜏 =𝜏 And: {𝜎11 + 𝜏12 = 𝜎12 + 𝜏22 = 𝜏1 21 11 22 21 𝜏1 ≥ 𝑚𝑎𝑥{𝜏12 ; 𝜏22 } Choose: 𝜏2 ≥ 𝑚𝑎𝑥{𝜏11 ; 𝜏21 } (2)↔ 𝐺1 = 𝑅11 𝑊11 + 𝑅12 𝑊21 = −𝑐1 𝑘12 𝑘21 (1+𝑇22 𝑠) 𝑘22 (1+𝑇 ′ 𝑠)(1+𝑇21 𝑠) 𝑘11 (1+𝑇 ′ 𝑠)(1+𝑇12 𝑠) (1+𝑇 ′ 𝑠)(1+𝑇11 𝑠) 𝑒 −(𝜏1−𝜏12+𝜏11) + 𝑒 −(𝜏1−𝜏22+𝜏21) 𝐺2 = 𝑅21 𝑊12 + 𝑅22 𝑊22 = −𝑐2 𝑘21 𝑘12 (1+𝑇11 𝑠) 𝑐1 𝑘11 (1+𝑇12 𝑠) 𝑐2 𝑘22 (1+𝑇21 𝑠) (1+𝑇 ′ 𝑠)(1+𝑇22 𝑠) 𝑒 −(𝜏2−𝜏21+𝜏22) + 𝑒 −(𝜏2−𝜏11+𝜏12) 𝑆𝑜 𝑇′ = 𝑇′12 = 𝑇′1 > 𝑚𝑎𝑥{𝑇12 , 𝑇22 } { 11 𝑇′21 = 𝑇′22 = 𝑇′2 > 𝑚𝑎𝑥{𝑇11 , 𝑇21 } 28 To investigate experimentally the performance of proposed decoupling controller it will be hereafter demonstrated on the incubator as an interacting TITO process has the transfer matrix with following particular elements: 0.58𝑒 −90𝑠 𝑊11 (𝑠) = 1796𝑠 + 0.4𝑒 −92𝑠 𝑊21 (𝑠) = − 1100𝑠 + 5.56 × 10−4 𝑒 −0.18𝑠 𝑊12 (𝑠) = 255𝑠 + 1.67 × 10−3 𝑒 −0.02𝑠 𝑊22 (𝑠) = 45𝑠 + Using the proposed procedure above for designing a decoupling controller described by transfer matrix R, it will be obtained particularly: 𝑅11 = 𝑅12 = 0.00167(1 + 255𝑠) −91.82𝑠 𝑒 + 300𝑠 −0.000556(1 + 45𝑠) −91.98𝑠 𝑒 + 300𝑠 𝑅21 = 0.4(1 + 1796𝑠) −2𝑠 𝑒 + 1800𝑠 𝑅22 = 0.58(1 + 1100𝑠) + 1800𝑠 which are achieved by using the PID tuning tool of MATLAB for 𝐺1 and 𝐺2 separately, where their initial values have been determined first by using Ziegler-Nichols method The PI parameter: PI1: 2400 + 1.38 𝑠 and PI2: 2703 + 56.54 𝑠 29 Figure 28: MATLAB demonstrated simulation for decoupling control Figure 29: Step response with decoupling controller Simulating the system with the controller parameters, the results are good quality When we change the value, there is no interaction between the channels Simulation results without overshoot, the steady stage error is zero, the settling time is fast and there is no interaction effect when using the decoupling controller 30 3.5 Select controller parameter With 𝑅11 = 𝑐11 (1 + 𝑇12 𝑠) −𝜎 𝑠 𝑒 11 ′ + 𝑇 11 𝑠 𝑅12 = 𝑐12 (1 + 𝑇22 𝑠) −𝜎 𝑠 𝑒 12 + 𝑇 ′12 𝑠 𝑅21 = 𝑐21 (1 + 𝑇11 𝑠) −𝜎 𝑠 𝑒 21 + 𝑇 ′ 21 𝑠 𝑅22 = 𝑐22 (1 + 𝑇21 𝑠) −𝜎 𝑠 𝑒 22 + 𝑇 ′ 22 𝑠 And the value previously selected 𝑅11 = 0.00167(1+255𝑠) 1+300𝑠 𝑒 −91.82𝑠 𝑐11 > 0.00167 → over shoot (P.O% ) increase 𝑐11 < 0.00167 → slow response • Choose 𝑐11 = 0.0015 → P.O%= and fast response 𝑇′11 > 300 → high over shoot 𝑇′11 < 300 → slow response • Choose 𝑇′11 = 300 → P.O%= and fast response 𝜎11 > 91.82 → the first of graph is more noise 𝜎11 < 91.82 → decrease noise • Choose 𝜎11 = 0.02 → ignore noise at temperature channel 31 Figure 30: Step response with decoupling controller after calibration Figure 31: The first response graph when not yet calibration 32 Figure 32: The first response graph when calibrated 𝑅12 = −0.000556(1+45𝑠) 1+300𝑠 𝑒 −91.98𝑠 𝑐12 > 0.000556 → over shoot (P.O%) increase, fast response 𝑐12 < 0.000556 → slow response • Choose 𝑐12 = 0.000556 → P.O%= and fast response 𝑇′12 > 300 → high over shoot 𝑇′12 < 300 → slow response • Choose 𝑇′12 = 300 → P.O%= and fast response 𝜎12 > 91.98 → fast response 𝜎12 < 91.98 → slow response and the first of graph is more noise • Choose 𝜎12 = 220 → ignore noise at temperature channel decrease over shoot and fast response 33 Figure 33: Step response with decoupling controller after calibration Figure 34: The first response graph when not yet calibration 34 Figure 35: The first response graph when calibrated 𝑅21 = 0.4(1+1796𝑠) 1+1800𝑠 𝑒 −2𝑠 𝑐21 > 0.4 → high over shoot 𝑐21 < 0.4 → unstable • Choose 𝑐21 = 0.4 → stable and over shoot = 𝑇′21 > 1800 → unstable 𝑇′21 < 1800 → high over shoot • Choose 𝑇′21 = 1800 → P.O%= and stable 𝜎21 > → unstable 𝜎21 < → unstable • Choose 𝜎21 = → stable → unchanged 35 Figure 36: Step response when changing parameter c21 , 𝑇′21 , 𝜎21 𝑅22 = 0.58(1+1100𝑠) 1+1800𝑠 𝑐22 > 0.58 → unstable at temperature channel 𝑐22 < 0.58 → very high over shoot at temperature channel Figure 37: Step response when changing 𝑐22 • Choose 𝑐22 = 0.58 → stable and over shoot = 𝑇′22 > 1800 → high over shoot 𝑇′22 < 1800 → high over shoot 36 Figure 38: Step response when changing 𝑇′22 • Choose 𝑇′22 = 1800 → P.O%= and stable 𝜎22 > → overshoot and slow response Figure 39: Step response when adding 𝜎22 • Choose 𝜎22 = → stable → unchanged After calibrated, we have: 𝑅11 = 𝑅12 = 0.0015(1 + 255𝑠) −0.02𝑠 𝑒 + 300𝑠 −0.000556(1 + 45𝑠) −220𝑠 𝑒 + 300𝑠 37 𝑅21 = 0.4(1 + 1796𝑠) −2𝑠 𝑒 + 1800𝑠 𝑅22 = 0.58(1 + 1100𝑠) + 1800𝑠 Using turning to find PI controller parameter, have: PI1: 2400 + 1.38 𝑠 and PI2: 2703.51319080862 + 56.5499328559093 𝑠 So we have the result: Figure 40: Step response after select controller parameter The results are very good quality When we change the value, there is no interaction between the channels Simulation results has small overshoot, the steady stage error is zero, the settling time is fast and there is no interaction effect when using the decoupling controller 38 CONCLUSION Through this topic of making project with the topic: " MICROCONTROLLER SYSTEM FOR NEWBORN INCUBATOR" helped us better understand theoretical and practical issues, in order to consolidate the knowledge learned in the past time and less surprised in the actual process later With the enthusiastic tutorial of Dr NGUYEN THU HA we have obtained the following results: - Know about temperature and humidity measurement and control methods - Understand about control system with ARDUINO - Learn more about the features of MATLAB Simulink and using MATLAB to find PID parameter - Calculation and design the decoupling for multi-time delay TITO process Once again, we would like to sincerely thank to Dr NGUYEN THU HA helped us with the project Due to limited research time and our own level, the project still has many shortcomings We are looking forward to receiving your suggestions to make our project be more complete Thank you sincerely! Ha Noi, 17/7/2020 Group of students NGUYEN TUAN HUNG DO HUY HOANG 39 REFERENCE [1] Ha KHCN-BK 2018.pdf [2] Linear Control_Nguyen_Doan_Phuoc [3] Phuoc-Toi-uu-hoa.pdf [4]Vu Thi Thuy Nga, Lecture EE3288E Control Theory [5] Thiết kế phần cứng lồng ấp trẻ sơ sinh.pdf [6] https://123doc.net/document/5947407-he-thong-dieu-khien-vi-khi-hau-trong-longnuoi-duong-tre-so-sinh.htm?fbclid=IwAR2T0pUSvGUFW9KHp2_SuZexhM_cKuHW9RMPHlCOYXBkKEBHCuiZ2MT57Q 40 ... input two output processes (TITO) with multi-time delays 3.4 A new decoupling controller for multi-time delayed TITO processes Consider an interacted multi-time delayed TITO process described by... unstable • Choose