Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Fuzzy Approach to pH Signal improvement to Reduce Process Pickup of Disturbances Dr.Ramesh Patnaik1, Prof Rajesh kanuganti*2, #1 Department of Instrument Technology, Andhra University A.U.C.E, Visakhapatnam , A.P, India [email protected] #2 Assistant Professor ,Dept of ECE Khammam Institute of Technology & sciences Khammam, A.P,INDIA [email protected] Abstract—pH is a deceptively simple measurement. However, there are many factors that need to be taken into account for a reliable reading. The most important characteristic of pH electrodes is it's very high impedance, of the order of 109 ohms. This is compounded by noisy factory environments and by long distances between the electrode and the controller. A typical pH measuring device would be normally configured to operate in the single ended mode, also known as the asymmetrical mode. This means that the reference electrode would be connected to the ground potential of the amplifier. This configuration works very well as long as the environment is electronically noise-free. This is not the situation in an industrial environment. It is very commonly seen that the readings on a pH controller suddenly fluctuates, even to overrange or under-range condition. In this paper the application of fuzzy logic in combination with smart transmitter network is investigated to reduce the signal wandering due to various disturbing signals. Key words: pH, combined electrode, signal wandering, smart transmitter, fuzzy logic I. INTRODUCTION A. PH Measurement The concept of the pH measurement start as the negative logarithm of the hydrogen ion concentration The glass pH electrode is still the most prevalent online composition measurement in the process industry. The logarithmic relationship offers an incredible rangeability and sensitivity far beyond the capability of any other common field measurements. For the 0-14 pH scale, the glass electrode can measure 14 orders of magnitude change in hydrogen ion concentration. At 7 pH, the glass electrode can detect changes in the hydrogen ion concentration to the eight decimal place. aH = 10−pH pH = − log (aH) aH = γ ∗ cH cH ∗ cOH = 10−pKw (1) (2) (3) (4) Where: aH = hydrogen ion activity (gmmoles per liter) cH = hydrogen ion concentration (gm-moles per liter) cOH = hydroxyl ion concentration (gm-moles per liter) γ = activity coefficient (1 for dilute solutions) pH = negative base 10 power of hydrogen ion activity pKw = negative base 10 power of the water dissociation constant (14.0 at 25oC) More exactly pH is a function of hydrogen ion activity as defined by Equations 1 and 2. Most pH measurements are in extremely dilute water solutions where the hydrogen ion concentration is the same as the hydrogen ion activity. In Equation 3, this one to one relationship corresponds to a unity activity coefficient. In some chemical reactors, the concentration of ions or non-aqueous solvents is high enough to cause the activity coefficient to significantly decrease. In these applications, the pH can change even if the hydrogen ion concentration is constant if the concentration of ions and solvent changes. The product of the hydrogen and hydroxyl ion concentration in a water solution depends upon the water dissociation constant (pKw) per Equation 4. The water dissociation constant changes with temperature on the average about -0.03 per deg C which for a given hydroxyl ion concentration causes the pH to change -0.03 per deg C for strong basic solutions. This change in solution pH with temperature is not corrected by the standard temperature compensator. A 8.3 pH hot waste stream at 50 OC that cools down to 25 OC by the time it reaches the effluent discharge will be above 9 pH, possibly in violation of an environmental constraint. The most popular pH sensor in chemical and waste treatment processes is the combination electrode that consists of a glass measurement electrode surrounded by a reference electrode. A spherical or hemi-spherical glass bulb has a pH sensitive surface that develops a milliVolt potential (E1) that is proportional to pH of the process. The inside surface of the glass electrode also develops a potential that corresponds to a 7 pH buffer. The sensing of pH depends upon a thin hydrated layer on the glass surface. If the surface in contact with process is in the same condition as the surface on the interior of the bulb, the difference in the milliVolt potentials between the interior and exterior glass surfaces can be described by the Equation 5, which is a simplification of the Nernst Equation. The standard temperature compensator corrects for the change in the milliVolts generated per pH unit by the temperature factor in Equation 5. Typically an RTD sensor embedded in the electrode is used to provide automatic compensation that effectively keeps the temperature factor at 298 oK (25 oC) in Equation 5. At 25 oC we have Equation 6 with the often stated potential for the glass measurement electrode of −59.16 milliVolts/pH. If you plot pH versus milliVolts per Equation 6, the intercept is at 7 pH and 0 milliVolts. The slope is −59.16 milliVolts, which corresponds to 100% efficiency. As the pH electrode ages, the slope and efficiency decreases. There must also be a reference electrode whose internal electrolyte is in contact with the process to complete the circuit through the process liquid. For the combination electrode shown in Figure 1 there is concentric porous ring called a “liquid junction” around the glass bulb for the potassium chloride reference electrolyte to move through to reach the process. E1 – E2 = 0.1984*(T + 273)*(7 − pH1) (5) at T = 25 oC: E1 − E2 = −59.16 ∗ (pH1 − 7) (6) Figure 1. Combination pH electrode – The potential (E1) developed at the glass electrode surface is indicative of the process pH. Changes in the hydrated glass surface layer activity and shunt resistance (R11) are corrected by a slope calibration adjustment. Changes in the liquid junction potential (E5) are corrected by a zero (offset) adjustment. B.Fuzzy logic Fuzzy logic has two different meanings. In a narrow sense, fuzzy logic is a logical system, which is an extension of multi-valued logic. But in a wider sense, fuzzy logic is almost synonymous with the theory of fuzzy sets, a theory which relates to classes of objects with un-sharp boundaries in which membership is a matter of degree. The concept of fuzzy logic is very different both in spirit and substance from the concepts of traditional multi valued logical systems. What is added is that the basic concept underlying FL is that of linguistic variable, that is, a variable whose values are words rather than numbers. In effect, much of FL may be viewed as a methodology for computing with words rather than numbers. Although words are inherently less precise than numbers, their use is closer to human intuition. Furthermore, computing with words exploits the tolerance for imprecision and thereby lowers the cost of solution. The traditional deterministic set in a universum can be represented by the characteristic function mapping into two-element set , namely for if , and if . A fuzzy subset membership function unit interval of is defined by a mapping into a closed , where for if , if , and if possibly belongs to but it is not sure. For the last case - the nearer to 1 the value is, the higher is the possibility that . In effect, much of FL may be viewed as a methodology for computing with words rather than numbers. Although words are inherently less precise than numbers, their use is closer to human intuition. Furthermore, computing with words exploits the tolerance for imprecision and thereby lowers the cost of solution. C.Smart transmitters The benefits of using smart transmitter network in the present configuration are of two fold. It automates the entire process at an attractive cost while providing more effective communication within the control system .The smart technology will enable us to utilize its advantages for initial start-up, maintenance work and to change the settings. All parameters for new measuring points or those to be changed are already defined and entered in the office. The stored data only need to be imported to the measuring instrument and the measuring point started up from the control room. The same applies to the operation and maintenance. The status of the measuring instrument can be displayed on-line, or in test mode the current output can for example, be set to a specific value in order to test the whole circuit. Most importantly the communications capabilities of the smart transmitters are utilized in the present system for the convenience of implementation of the logic. II PROBLEMS IN MEASUREMENT: CONVENTIONAL PH A.Electrical Interference –D.C sources pH is a deceptively simple measurement. However, there are many factors that need to be taken into account for a reliable reading. The most important characteristic of pH electrodes is it's very high impedance, of the order of 10 9 ohms. This is compounded by noisy factory environments and by long distances between the electrode and the controller. A typical pH measuring device would be normally configured to operate in the single ended mode, also known as the asymmetrical mode. This means that the reference electrode would be connected to the ground potential of the amplifier. This configuration works very well as long as the environment is electronically noise-free. This is not the situation in an industrial environment. It is very commonly seen that the readings on a pH controller suddenly fluctuates, even to over-range or under-range condition. This situation arises, when for example, the mixing motor is switched on. An old leaky motor might inject some electrical interference of 1 to 2 volts into the liquid whose pH is monitored. This noise being a common signal, is picked up by both the pH and the reference electrodes. Since in the asymmetrical mode, the reference electrode is grounded, the electrical noise is present only on the pH electrode. This noise would be amplified along with the pH signal and thus the fluctuating readings. If the electrical noise was from a DC source, typically like those in an electroplating tank, the problem would not be fluctuating readings mostly stable but incorrect values. B .Effect of mixing water flow rate: The dosing of the chemical would not be regulated based on the deviation of the pH from the set point but at a steady and fixed rate. This would cause overshoot and undershoot of the process and hence the control will not be smooth. In applications where fine control is required like those in food or pharmaceutical applications which usually operate within a narrow band, a limit control would not be acceptable. In addition the delay in ionization and spreading of reading affects the pH reading and by incorporating the mixing water flow rate into fuzzy logic compensation. C.Tank disturbance pick up In certain industries it becomes essential to monitor two parameters simultaneously and carry out corrective action based on one parameter first followed by the other. D. Effect of temperature variation Most pH measurements are in extremely dilute water solutions where the hydrogen ion concentration is the same as the hydrogen ion activity. IIn these applications, the pH can change even if the hydrogen ion concentration is constant if the concentration of ions and solvent changes.The product of the hydrogen and hydroxyl ion concentration in a water solution depends upon the water dissociation constant (pKw). control room. The transmitters can be configured with the help of PC. However it requires a smart to RS232 converter to access the transmitters. In addition, the driver software has to be installed in the computer system. The data communicated to the PC is used as the raw material for the fuzzy processing algorithm. The logic is resident in the system and implemented with the help of Matlab program. The water dissociation constant changes with temperature on the average about -0.03 per OC which for a given hydroxyl ion concentration causes the pH to change -0.03 per OC for strong basic solutions. This change in solution pH with temperature is not corrected by the standard temperature compensator. III.FUZZY APPROACH A.Basic configuration C. Implementation of communications The interface for the smart system is the current output. Digital information can be transmitted simultaneously both ways via the current output cables. The current output signal (0/4 to 20mA) is not affected because the mean value of the signal containing the digital information is equal to zero. The signals are superimposed by means of frequency shift keying(FSK) based on the standard 202Bell.The digital transmission signal is formed from two frequencies: 2200Hz="0" and 1200Hz="1" B. Hardware scheme The system utilized four smart transmitters for mixing water flow rate, recognised D.C source effects, temperature changes and the tank disturbance pickups. The transmitters work on fsk digital signals superimposed on the 4-20mA current loop. The settings of the transmitters are performed with the help of a smart communicator. The communicator can be fixed anywhere on the loop at field or at the In addition to the Normal point-to-point link, smart system also enables a network to be configured with up to 15 field devices. The cycle time is in the range of 10 seconds, which thus only allows limited application. Also, the analog output can no longer be used. The data transmission layer defines the format of a telegram. Access is to the master/slave method. The master station is in all cases the hand-held controller. Two masters can be connected up simultaneously. C.Fuzzy logic implementation. The behaviour of the effecting disturbances were studied over a period of time and also when the operators experience was combined, it was found that the input functions were most closely matching with those depicted in the table 1. The selected rules are given as per the table 3. All the output member ship functions were taken as trimf for simplicity. The fuzzy algorithm was taken of Mamdani type. Fuzzy model: Surface (for one set of two variables) TABLE I INPUT MEMBERSHIP FUNCTIONS 1 Mixing florate water Low Normal High Gaussmf Gauss2mf gaussmf 2 DC disturbance signals 3 Temperature variations 4 Tank disturbances Low Normal High Low Normal High Low Normal High Gaussmf Gauss2mf gaussmf Gbellmf gaussmf gbellmf Gbellmf gaussmf gbellmf IV.RESULTS AND DISCUSSSION TABLE2 OUT PUT MEMBERSHIP FUNCTIONS 1 Correction for pH signal Rules view Low Normal High Trimmf Trimfmf Trimmf A. Transmitter trimming performance without fuzzy The plot shows some irregular and indefinite spikes and flat portions which are infact erratic readings which are received without fuzzy trimming. error plot (Y-axis time, min, Y-axis pH error) B.System with fuzzy trimming Simultaneous plot was taken with fuzzy trimming which is shown below. The disturbing spikes are reduced by this approach as can be observed from the below plot. The improvements were marked in the plot. V.CONCLUSIONS Fuzzy logic trimming of the pH signal proposes an alternative method of signal improvement where the disturbances of a system cannot be truly defined in mathematical terms. The resultant system was shown marked improvements as shown in the results section. REFERENCES [1] Robert.E.King, Computational intelligence in Control engineering, 2nd ed., Marcer Dekker Inc., New York,1999. [2] M.Ramesh Patnaik,Vara Prasad.P.L.H, Aditya.. Control optimization by fuzzy supervisory approach in the sinter plant of an integrated steel plant, IISN-2010 [3] B.Kosko, Neural Networks and Fuzzy systems, a dynamic systems approach to machine intelligence, Prentice Hall., N.Y., 1992. [4] E.O.Doeblin, Measurement systems application and Design, McGrawHill, N.Y.,1992.. [5] Stuart Litch, pH measurement in concentrated alkaline solutions, Journal of Analytical Chemistry, 1985.. [6] M.Ramesh Patnaik, P.L.H.Varaprasad, A novel fuzzy logic alforithm guided smart transmitter network for thjree-element control optimization for a package boiler with extreme load variations in a challenging environment., International conference on Intelligne systems and Networks(IISN-2007) [7] M.Ramesh Patnaik, D.V.Rama Koti Reddy, P.L.H.Vara Prasad, A Novel Fuzzy Logic based model for real time monitoring of dry end parameters of paper machine automation, National Symposium on Intrumentation, NSI 32, 8-10 Dec 2008. [8] Li Zing Wang, A supervisory controller for Fuzzy Control systems that guarantees stability,IEEE transactions on automatic control, vol 39, no9, September,2001