DOC

temperature-control

By Leroy Fox,2014-04-05 22:10
13 views 0
temperature-control

     1

    A Study On PID Temperature Control For

    ENGR 315, Control Systems

    T. VanDerPuy, Non-IEEE Member

    never erases the steady-state error. ;

    AbstractThis paper focuses on an in depth study of B. Integral industrial temperature controllers, their use in a workplace

    The second system in a PID controller is the integral environment, some terms and options associated with them, and

    a physical testing thermal system. The testing was done in a control. The integral control eliminates the steady-error, but controlled environment using a Watlow Series 935A makes the transient response worse. The integral eliminates temperature controller a 7.5-watt heater, and a half-gallon fish the “droop” caused by the proportional band. Since the power tank. Some other topics included are inputs and outputs of level at set point is zero, and near zero right before it, the temperature controllers, PID, and its function associated with temperature settles at a point slightly below the set point temperature controllers, and some other functions available for

    use with industrial temperature controllers. using just proportional control, this results in a “droop” down

     from the set point. The integral part of PID control Index TermsProportional Control, Temperature Control, eliminates this. The controller output is proportional to the Temperature Measurement, Thermoresistivity, Thermal amount of time the error is present. Variables

    C. Derivative

    I. INTRODUCTION The third system working in a PID controller is the

    derivative control. The derivative control affects the system emperature control in industrial applications is an old

    by increasing stability, and by reducing the overshoot and T science, taking off mostly during the industrial revolution,

    undershoot of the function, and improving transient response. and coming into its own in the United States early in the

    The output under derivative control is proportional to the rate Twentieth Century. This control was very simple,

    of change of the error over time. This part of the control mechanical control that did not go beyond turning a heater or

    system is critical because in some processes, an overshoot in cooling device on or off. PID control, however is a fairly new

    temperature might cause a part or machine to be damaged. concept that was immediately accepted into use for

    temperature control applications, and gave way to an entire

    III. TYPES OF TEMPERATURE CONTROLLERS line of PID temperature controllers, including the entirely

    digital units seen at work in most applications today. Temperature controllers are usually characterized by the

    type of control they provide, and therefore, the type of outputs

    II. PID CONTROL, AND ITS USE WITH TEMPERATURE that are available on each unit.

    PID control stands for, and consists of three distinct A. On/Off feedback and control areas. A block diagram of the control Temperature controller’s outputs generally come in one of system can be found in the appendix. three varieties. The first of these is called an ON/OFF output. A. Proportional The first of these areas is proportional. The output of the proportional controller is relative to the difference between the temperature that is present and the set point. An adjustable proportional band is set up as either a range of temperatures, or a percentage of the set point temperature, and is located below the set point. The Proportional band is good for reducing the rise time of a process, and reduces, but This work was supported in part by Professor Paulo Ribero, Electrical Engineering, Calvin College. Tom VanDerPuy is a Senior Mechanical Engineering Student, Calvin College, Grand Rapids, MI 49507 USA (e-mail: tdv4@calvin.edu)

     2

     FIGURE 1 Time vs. Temperature Graph of ON/OFF Control FIGURE 2 Time vs. Temperature Graph of Time Proportioning Control

     The output on this controller is switched on at intervals, which depend on where the temperature is in the proportional This was the first type to arrive into production, and is the band; it is on more when it just enters the proportional band, simplest of the three. An ON/OFF output controller simply and off more as it reaches the setpoint. Figures 3 and 4 show

    turns the power on or off to a heater or cooling device how the device reacts at different required power loads.

     depending on which side of the setpoint that the temperature

    FIGURE 3 is on. The controllers that use this sort of output are very Time vs. Power Graph of Time Proportioning Control At 50% Power economical, however they have many problems, depending on the system you are controlling with it. The main problem with this is that the ON/OFF control does not provide a steady temperature, and can be off on either side of the setpoint by a certain magnitude depending on the heater or cooling device that is attached.

    One other problem is that this output is usually attached to a relay or other device to transmit the signal to power the heater or cooler. A mechanical relay, or even a transistor-based device only has a certain amount of switches until it needs to be replaced. If the temperature gets around the setpoint, and stays around there, the device will constantly be switching back and forth since it is so close to the setpoint. This puts a lot of stress on the switching device, and shortens its lifespan enormously. B. Time Proportioning The second variety is Time Proportioning Output.

     3

     Another way that this output reaches the device to be FIGURE 4 controlled is by an SSR or SCR control device that rapidly Time vs. Power Graph of Time Proportioning Control At 75% Power switches the power on and off to the heater, some switch many times a second. The benefits of this output are a completely smooth operation without any of the on/off jumps the other outputs have. This helps the process level out more evenly, reducing overshoot. It also extends heater life because it is getting a more even flow of power than with the other two devices.

    IV. INPUTS

     Another area of options available in choosing the correct PID controller are inputs.

A. Contact Sensors

    Contact Sensors are sensors that are actually touching or immersed in the material whose temperature they are

     measuring. This means that their temperature range is limited because of the degradation of the material that can The benefit of this kind of controller is that the overshoot occur due to high temperatures. The two most distinct types is greatly reduced due to the slowed down operation of contact sensors commercially available are thermocouples approaching the set point. A proportional or full PID and RTDs. controller must be used for this output.

    C. Process Control Output 1. Thermocouples

     The final type of output used in controllers today is the The thermocouple is the most prevalent, consisting of a Process Control output. This output gives a steady signal two-strand wire of dissimilar metals. Some examples of these instead of an on or off command, as shown in figure 5. types of wires are the K type thermocouples, whose wires are nickel and nickel chromium. All metals change their FIGURE 5 electrical EMF potential as temperature increases or decreases. Proportional Control Output (0-5V) In an environment where these two wires are exposed to the same temperature, they experience different changes in electrical potential because they are made of dissimilar elements.

     When joined together, the difference in electrical potential between these two wires can be measured by the temperature controller, and this analog signal in turn can be turned into a digital signal that is sent to the controllers brain to be converted into a number (temperature) based on what type of thermocouple is connected. The controller must be told which type of thermocouple is connected in order to make this computation.

     There are many different kinds of thermocouples for use in different environments, so it is key that the temperature controller is computing results for the right thermocouple. Usually a controller can accept a number of different types, and some can even tell what type is being used as soon as it is plugged in. Proportional Output control is used to turn a heater or 2. Resistance Temperature Detectors cooling device on, but only with partial power depending on how much it determines the heater needs to be actuated. This Resistance Temperature Detectors, or RTDs as they are output is used exclusively on full PID controllers. The output known detect temperature with a single element wire, instead is an analog signal, usually a voltage from 0-5 VDC or a of the two element wire of the thermocouple. They consist of current from 4-20 ma. a length of fine wire wrapped around a glass or ceramic core. As shown in the diagram above, the controller is telling the The temperature controller reads the resistance of the wire, heater to give 25% power. This voltage or current is and correlates it to a temperature listed for that resistance that converted into a percentage of power needed to send to a is programmed into the controller. heater or cooling device, usually by an analog PLC input.

     4

     and sometimes took days on more complicated temperature

     systems.

     This was because a temperature system is one of the more

     complicated systems to model for PID control. There are

     numerous variables that are nearly impossible to simulate

     versus how they behave in the real world. For this reason,

     manufacturers of temperature controllers soon started

     3. Comparing RTDs and Thermocouples producing units that auto-tuned. This was an incredible time

     and labor saving creation because all it took to tune the new

     RTDs are generally better performing than thermocouples controllers was to set it up in the environment and let it run in almost every area, except for a few. The first is that they and decide the right PID variables for the process by itself. are not as robust as thermocouples, and so usually come in a The way it does this is shown in figure 6 below. sheath or jacket to protect them from the environments they

    will be introduced to. Even so, there are many environments FIGURE 6 Temperature vs. Time Auto tuning that they cannot go in, especially ones where they will be jolted or shocked by the machine operation. They are also

    slightly more expensive, but not so much to justify buying

    them over a thermocouple because of their numerous

    advantages.

     The first of these advantages is that they are more accurate

    than thermocouples, and more repeatable as well. A

    thermocouple can get worn out more easily because it is

    usually not jacketed as well as an RTD. Another advantage

    that RTDs have is that they have a higher immunity to

    electrical noise from equipment. This means that they can be

    placed next to generators, transformers, or motors in a

    workplace environment, and give a more accurate

    temperature than a thermocouple.

     One real advantage that thermocouples have is that they

    have a good point reading. What this means is that the

    analog signal that is sent to the temperature controller is from

    a single point, where an RTD’s resistance is read along the entire coil. The advantage here is that the thermocouple can be placed in a small area, or placed so that it reads a The temperature controller starts out by putting the heater temperature on a minute area of a part or machine. or cooling device on full power until it reaches 90% of the set Another advantage that thermocouples hold over RTDs is point. It does this to determine how fast the heating or that they can be used over a broader temperature range than cooling device works so that it does not overshoot the set RTDs. RTDs are more sensitive to extreme temperatures as point. As soon as it reaches 90%, it begins to back off the well, and have a more limited range of operation. A full power proportionally to what it has learned about the heating comparison of temperature controllers versus RTDs can be or cooling device, and watches how fast the temperature found in the appendix. drops when it shuts off the device. This is important for it to decide when and how much power to cut when the process

    gets near the set point. B. Non-Contact Sensors After this the auto tuning is complete, and the temperature Most commercially available non-contact sensors are controller decides on reasonable values for the PID. This is devices that detect radiation heat energy from a target that the usually not the end of the process, however, as an operator sensor is pointed towards. These are used in rugged still has to come in and fine-tune the PID values to make sure environments where it is physically impossible or impractical the process is operating at an optimal level. Other situations, due to machine movement or temperature to place a however, are less demanding, and require only the auto thermocouple or RTD. These sensors are usually connected tuning of the controller for safe and optimal operation. to more complicated temperature computers that are not in Most of the time the only problem with the temperature the scope of this paper. controllers auto tuning settings is a slight overshoot in the

    final temperature. This can be adjusted down by simply V. TUNING A TEMPERATURE CONTROLLER changing the value of the integral in the PID control.

     Most contemporary PID controllers come with an easy to Tuning the PID system on a temperature controller was not use interface that can be learned in a couple of hours. The an easy task by any means in the earlier years of PID earlier models were not so easy to use, and tuning a temperature control. It involved setting up the system, temperature controller usually involved bringing in a configuring it to how you best thought it might need, and representative from the company for a day to teach a few guessing what parameters the PID system should use for the maintenance and engineering workers how to set and adjust operation. This was more than tedious for a worker to do, the controllers.

     5

    switched DC to drive the SSR; otherwise I would have needed VI. TYPES OF TEMERATURE CONTROLLERS

    an external power supply to drive it. Temperature controllers come equipped with a number of The controller is very simple to set up, and has a digital different options. Deciding on a temperature controller has a menu system that lets you cycle through all of the variables by lot to do with the inputs and outputs, but there are also other using the three buttons on the front of the controller. The set features that they can utilize that are not necessary for all point can easily be changed without having to go through any operations. of the menus via a quick select button and an up and down Temperature controllers can be used in either a stand-alone arrow. The controller is very compact, and fits in a standard operation, or can be run with a programmable logic controller th DIN size hole. 1/32or PLC. The more complicated operations usually have the The controller is adjustable for Celsius and Fahrenheit temperature controllers hooked up to a main communications temperature ranges, and also is selected for different bus that can be monitored from any part of the installation. temperature ranges. For example, the controller can be set to A step down from this system is a temperature controller record between 0 and 1000 degrees, or 1000 to 2000 degrees. simply hooked up to a stand alone PLC. In this fashion, the It must be adjusted like this because the LED characters limit PLC takes the temperature controller set point as an input, how much data can be presented on-screen. and any alarms that the temperature controller might be The reason that I picked this controller was that it was both programmed with. Through this, the PLC can tell the economical and practical for the application that I was using machine to stop functioning in case of a system overload that it for. It is an entry-level controller from Watlow, some of the temperature controller cannot handle. there more complex controllers that can run up to five For the most basic case, the temperature controller is thousand dollars for a base model. These controllers can run stand-alone, and has no backup system. This is mainly up to sixteen process loops simultaneously, come with incorporated into systems that do not pose a hazard if graphing software, and can also run process loops for flow overloaded, and will not damage any expensive equipment. controls.

     The Watlow Series 935a model that I purchased cost forty VII. PHYSICAL TESTING dollars purchased used on Ebay, the original model would

     As mentioned before, temperature controllers are hard to have cost around two hundred dollars.

    model, or digitize. My initial aspirations were to do a C. Results and Conclusions MATLAB simulation of my proposed design and test it

     The controller took a period of 235 minutes to auto-tune to against a real world model. This proved to be somewhere in

    72 degrees Fahrenheit. The PID values are as follows: between impossible and impractical.

     A. Impractical To Model Proportional: 25

     The system to be modeled, as shown in the photograph in

    the appendix turned out to be simply too complicated, even Integral: 10

    though it is just a fish tank. This is the reason that

    manufacturers of temperature controllers started Derivative: 2.01

    incorporating auto tuning functions integral in the controllers.

    The industrial systems, most of them more complicated than A graph of the auto tuning from points taken every 5 my fish tank set up, are just impractical to model because of minutes over those 235 minutes is shown in the appendix. all of the variables inherent in a physical temperature system. The experiment was a success, the auto-tuning went as

     Conduction, convection, and radiation heat losses would described in the text on the controller by Watlow. The have to be considered, along with the actual density of the temperature rose to about 90% of the set point, and then water (tap water at our house is more dense than a control proceeded to fall, and rose back again a couple of times before water because of the high mineral content). The actual power settling at the set point.

    output of the heater would also have to be considered along In conclusion, PID temperature control systems have with how much of that heat got transmitted to the water developed over time to become incredible savers of manpower through its casing. and resources. They have revolutionized the way that

     Along these lines, the placement of the heater and industry controls temperature and made many manufacturing temperature sensor, and a number of other variables that environments safer. One of the major issues that became would contribute to making modeling the system impractical. apparent while studying PID temperature controllers was how

    difficult it is to model a physical environment, and how the B. The Physical System newest batch of PID controllers takes care of that through The option that I chose to go to was a physical system features like auto-tuning, and data logging. The world of modeling. The environment is a half-gallon fish tank with a temperature control has become more tangible and real to MarkeTech International Model 10975 15-watt immersion even the lowliest worker through these devices. heater. The heater is hooked up to a Watlow solid-state relay (SSR) that runs off of the switched 12-volt output of the temperature controller. The controller itself is a 2003 model Watlow Series 935a full PID temperature controller. The specs of the controller can be seen in the appendix. The controller has an option between a relay output and switched DC output. I chose the

     6

    VIII. APPENDIX

    Table 1:

    Thermocouple VS. RTD Data

    [3]

    Figure 1:

    Block Diagram of Temperature Control

     7

    [2]

    Figure 2:

    Auto Tuning Graph

    Time vs Temperature73

    72

    71

    70

    69

    68

    Temperature (F)67

    66

    050100150200250

    Time (min)

    Table 2:

    Data Set For Auto-Tuning

    Time (min) Temperature (F) Time (min) Temperature (F) Time (min) Temperature (F) 0 66.9 80 67.7 160 69.7 5 67.8 85 67.6 165 69.9 10 69.1 90 67.4 170 70.3 15 69.5 95 67.6 175 70.3 20 69.3 100 67.7 180 70.4 25 69.1 105 67.8 185 70.7 30 69 110 67.8 190 70.9 35 69 115 67.9 195 71.1

     8 40 68.9 120 68.6 200 71.2 45 68.8 125 68.6 205 71.5 50 68.6 130 68.8 210 71.4 55 68.7 135 69 215 71.7 60 68.6 140 69 220 72 65 68 145 69 225 72 70 68.1 150 69.3 230 72 75 67.8 155 69.5 235 72

    Figure 3:

    Watlow Series 935a Specifications

9

[4]

[1] Johnson, Phil, MC Shane INC, Understanding a PID Controller, General information, Available: http://www.mcshaneinc.com/html/Library_UnderstandingPID.html VIII. REFERENCES [2] Temperatures.com, Types of Temperature Sensors, 2004, Available: http://www.temperatures.com/sensors.html

10

[3] Watlow Corp, Watlow Educational Series, Temperature Control, the Watlow Educational Series Book 5, pp. 11-32, 1995 [4] Watlow Corp, Literature, Series 953a User Manual, p 52, General information also, January 2002 [5] Weed Instrument Company, Weed Information Central, RTD vs. Thermocouple Comparison, 2004, Available: http://www.weedinstrument.com/info_central/rtd.html [6] Williams, Charles D., School of Physics, Feedback and Temperature Control, University of Exeter, General information, Available: http://newton.ex.ac.uk/teaching/CDHW/Feedback/#Preface

Report this document

For any questions or suggestions please email
cust-service@docsford.com