DE102013102721A1 - Method for determining a condition of a pH sensor comprising a glass membrane - Google Patents

Abstract

The invention relates to a method for determining a state of a pH sensor comprising a glass membrane, in which a temperature (TProz) of a process is measured for the state analysis, to which the pH sensor (1) is exposed, depending on the temperature (TProz). the state of the sensor (1) is determined. In a method which, despite the simple design of the pH sensor, enables reliable diagnosis of the state of the pH sensor, in addition to the temperature (TProz), an electrical process resistance (GProz) of the glass membrane (4) is determined, from which the state of the pH sensor (1) is closed.

Description

The invention relates to a method for determining a state of a glass membrane comprising a pH sensor, wherein the state analysis, a temperature of a process is measured, to which the sensor is exposed, wherein the condition of the sensor is determined depending on the temperature.

In process engineering plants, a variety of probes for process instrumentation are used to control processes. Transmitters are used to record process variables, such as a pH of a medium. By means of actuators, the process flow can be influenced as a function of the detected process variables in accordance with a strategy predetermined, for example, by a control station. As examples of actuators are control valves, a heater or a pump called.

By long-term use of the sensors contained in the probes, in particular the pH sensor, the state of the sensor may change so that it no longer delivers correct measured values. From the EP 1 550 861 B1 a method for determining the state of a probe is known in which a pH sensor, without removal from the probe, cleaned from time to time. The temperature of the measuring probe or of the medium surrounding the measuring probe is measured by means of a temperature measuring sensor. On the basis of the thus registered course of the measured temperature is closed to the state of the probe.

The invention is therefore based on the object to provide a method for determining a state of a glass membrane comprising a pH sensor, which allows a simple and reliable inference to the state of the pH sensor.

According to the invention the object is achieved in that in addition to the temperature, an electrical process resistance of the glass membrane is determined, from which it is concluded that the state of the pH sensor. This procedure is used when the measuring electrode of the pH sensor is intact. In addition, it is assumed that the electrical resistance of glass electrodes changes as a function of the temperature. The specific resistance of glass increases when the glass is exposed to high temperatures for a long time. This property is contrary to the pure temperature dependence, so that due to the opposing processes depending on the glass material different resistance changes over time can arise. This dynamic is a sign of the aging of the glass membrane of the pH sensor and thus represents a parameter for the sensor analysis. The determination of the electrical resistance takes place in the already connected to the pH sensor transmitter, which is why it can be dispensed with an additional evaluation. Thus, a very inexpensive method for determining the condition of a pH sensor is realized.

The basis for this is that a process resistance of the glass membrane is compared with an initial value of the glass membrane resistance at a defined temperature. The process resistance of the glass membrane is understood here to mean a resistance of a glass membrane exposed to the process to be monitored in particular by the pH sensor.

Advantageously, the electrical resistance of the glass membrane is determined at different temperatures in a first step during a final test in the production of the pH sensor, wherein a functional relationship is determined from the course of a resistance-temperature value pair.

In one embodiment, the functional relationship is determined by means of a mathematical regression. This mathematical regression is based on finitely many resistance-temperature value pairs. The regression method is a statistical evaluation method that investigates a possible correlation between data points with an assumed inner relationship. It thus represents a possible correlation between data points with the assumed internal relationship. The determined functional relationship can then be stored in the pH sensor or a transmitter connected to the pH sensor.

In one variant, the mathematical regression is based on an exponential curve of the electrical resistances of the glass membrane over the temperature measured in the final test, wherein the exponential profile of the electrical resistance is linearized and an estimated resistance value is determined as a function of different process temperatures. In addition to repeatedly measuring the actual temperatures of the process and determining the electrical resistance of the glass membrane of the pH sensor, an estimate of the glass membrane resistance is thus made. This function makes it possible to continuously calculate an estimate of the electrical resistance of the glass membrane over the temperature range of the pH sensor. Thus, there is a linear estimation which shows a relationship between the Process temperature and an associated logarithmic electrical resistance of the glass membrane shows. Such a connection serves as a basis for sensor analysis with different methods. On the one hand, statements about the state of the pH sensor are possible in connection with the respective temperature measurement, and on the other hand conditions are created for the use of the glass membrane resistance as a temperature image.

In one embodiment, a functional pH sensor is concluded when the electrical resistance of the glass membrane, determined at a process temperature, is within a resistance interval. Due to this simple method, the state analysis can be carried out very quickly and reliably. In the simplest case, predefined interval limits are sufficient here. This state analysis can thus be repeated in the field during the use of the sensor.

When using the pH sensor in the process, a process temperature is preferably measured and an electrical process resistance of the glass membrane is determined and a difference between the determined, logarithmic electrical process resistance of the glass membrane and the logarithmic resistance value estimated for the measured process temperature is formed and with a resistance limit value or a resistance limit Resistance interval compared. For this method, besides the calculation of the estimated value of the electrical resistance of the glass membrane, only a limit defined in advance is necessary. This provision is a robust method.

Alternatively, a resistance interval is determined by calculation using statistical methods. By means of such statistical methods it is possible to obtain a well-founded evaluation of whether the measured process resistance corresponds to the originally determined functional relationship between the estimated glass membrane resistance and the respective temperature.

In one embodiment, a warning signal is emitted by the determined electrical process resistance of the glass membrane once the resistance interval has been undershot or exceeded. This should alert the service personnel that there is an initial suspicion of malfunction of the pH sensor.

In an alternative, when the resistance interval is exceeded or fallen short of several times, the pH value of the sensor is changed by the electrical resistance of the glass membrane. The repeated exceeding or falling below the interval limits of the resistance interval by the determined process resistance can reliably conclude that the pH sensor is defective and must be replaced by a new pH sensor.

The invention allows numerous embodiments. One of them will be explained in more detail with reference to the figures shown in the drawing.

It shows:

1 : schematic representation of a pH sensor,

2 : schematic representation of the functioning of the pH sensor

3 FIG. 2 shows an exemplary embodiment for the dependence of the electrical resistance of the glass membrane of the pH sensor on the temperature during a final production test, FIG.

4 : Representation of the dependence of the electrical resistance of the glass membrane on the temperature determined by linear regression 3

In 1 is a pH sensor 1 shown schematically, which is an inner tube 2 in which there is a measuring electrode 3 extends. The inner tube 2 is over a glass membrane 4 opposite the measuring medium 8th completed. A the inner tube 2 surrounding outer tube 5 includes a reference electrode 6 , where the outer tube 5 by means of a diaphragm 7 opposite the measuring medium 8th is disconnected.

How out 2 can be seen during the measurement process, both the glass membrane 4 as well as the diaphragm 7 into the measuring medium 8th immersed. Inside the inner tube 2 is an internal buffer containing the conductive connection between the inside of the glass membrane 4 and the measuring electrode 3 manufactures. The voltage potential during the measurement at the measuring electrode 3 arises, is connected to the voltage potential at the reference electrode 6 compared that within the outer tube 5 immersed in an electrolyte, preferably KCl, passing through a porous partition in the form of the diaphragm 7 slowly into the medium 8th diffused. The voltage potential of the measuring electrode 3 and the reference electrode 6 be connected to a transmitter 9 evaluated.

In addition to this actual measurement function, the transmitter takes over 9 also monitoring the condition of the pH sensor 1 , This is in the transmitter 9 a state analysis software stored, which will be explained in more detail below. For diagnosis of the pH sensor 1 based on the temperature dependence of the glass resistance of the glass membrane 4 will be in a first Step in the final production test of the pH sensor 1 the glass membrane resistance _measuring G determined at at least four different temperatures. Such a dependence is in 3 represented, wherein the electrical resistance G _{Mess of} the glass membrane 4 has an approximately exponential dependence on the temperature T. y = a × e ^bx (1)

This relationship is treated by means of a mathematical regression. Thus the mathematical regression with less effort in the transmitter 9 implementable, the exponential function is converted into a linear function: ln (y) = ln (a + b _x) (2)

From this formula (2) is a linear relationship between an estimated resistance value G logarithmic _estimated and the measured temperature T _measurement results. G ' _estimate = ln (G _estimate ) = b1 * T _measurement + b2 (3)

Such a connection is in 4 illustrates in which the curve A the logarithm of the measured resistances ln (G _Mess ) according to 1 represents, while the straight line B shows the variation of the estimated logarithmic resistance value G _'estimation over temperature. Point C represents a process resistance ln (G _Proz ) or G ' _Proz actually measured during use in the field at a process temperature T _Proz .

During the use of the pH sensor 1 In the process a comparison of the measuring points P ' _Proz = f (G' _Proz , T _Proz ) and P ' _Schätz = f (G' _estimation, T _proz ) takes place at the point T = T _Proz. In this case, the actually determined, logarithmic process resistance G _{'per cent} with the estimated logarithmic resistance value G' _estimation compared T _{per cent} at the predetermined measuring temperature. If the determined process resistance G ' _Proz lies within a simple resistance _{interval spanned} by a previously defined limit value, then the pH sensor becomes 1 classified as functional. However, if the determined process resistance G ' _Proz lies outside the simple resistance interval, then this is an indication that either the glass membrane resistance of the pH sensor 1 is faulty or the temperature sensor is defective, which measures the temperature of the process medium to be examined.

G’_Schätz = b1·t + b2

lineare Regression (Methode kleinster Fehlerquadrate)

s_F

Standardfehler des individuellen Prognosewertes

t_v,α/2

Testgröße nach Student-Verteilung für definierte Irrtumswahrscheinlichkeit 1 – α und Freiheitsgrad ν

ν = n – 2

Freiheitsgrad

n

Anzahl der Messpunkte

Alternatively, you can use the appropriate software in the transmitter 9 but also calculate the resistance interval statistically. For the resistance interval, which is preferably called the forecast interval, the following applies: G ' _estimate - t _{ν; α / 2} s _F ≤ G' _Prog ≤ G ' _Estimate + t _{ν; α / 2} s _F (4) With

G ' _estimate = b1 * t + b2

linear regression (least squares method)

s _F

Standard error of the individual forecast value

t _{v, α / 2}

Test size according to student distribution for defined error probability 1 - α and degree of freedom ν

ν = n - 2

degree of freedom

n

Number of measuring points

Example for determining a forecast interval:

Starting from the resistance values determined in a final production test for different temperatures, as described in the 3 are shown, the coefficients are determined for a regression function and in the pH sensor 1 or in the transmitter 9 stored. About these parameters and in the pH sensor 1 or in the transmitter 9 deposited regression function allows the measuring point to determine an estimated resistance value G _estimated or G _'estimate for a process temperature T _Perc. After the calculation of a standard error of an individual prognosis value s _F , taking into account a test variable t _{v, α / 2} , that in the pH sensor 1 or the transmitter 9 is stored, an estimation interval for the individual forecast value G _Prog or G ' _{Prog are} determined from which results the forecast interval.

If you take z. B. 16 resistance values at different temperatures ( 3 and 4 ) And performs a linear regression on their logarithmic values, we obtain the coefficients b1 and b2 of the equations 1 and 2. For a given process temperature is thus T _{per cent} receives an estimated logarithmic resistance value G _'estimation. After calculating the standard error of the individual forecast value s _F and a defined test variable t _{v, α / 2} , which results from the number of degrees of freedom ν = n - 2 = 14 and the error probability 1 - α = 0.999, the result is a forecast interval with an upper one and a lower limit for G ' _prog . With an error probability of 1 - α = 0.999, the logarithmic individual process resistance G ' _becomes the glass membrane 4 at the given process temperature T _Proz between the lower and the upper limit. A logarithmic process resistance below the lower limit value is present Initial suspicion of a systematic deviation from the sensor-specific, functional relationship between glass membrane 4 and temperature reading. The same applies to a logarithmic process resistance above the upper limit. If the interval limits are exceeded or exceeded several times, this may be due to a faulty status of the pH sensor 1 be assumed.

This detailed statistical method in this case represents only one option, the sensor diagnostics. The above-explained, very robust way for direct comparison of estimated resistance value G _estimated and the determined process resistance G _{per cent} at temperature T _{per cent} is equally possible. Furthermore, the detour via a linearization of the temperature dependence of the glass membrane resistance also represents only one option.

The illustrated statistical method is used on the one hand to analyze the current state of a pH sensor 1 on the verification of the functional relationship between the electrical resistance of the glass membrane and the process temperature. If the resistance values of the glass membrane determined in the process lie outside the interval limits of the resistance interval determined as the prognosis interval, ie if the relationship between glass resistance and process temperature is no longer present, either the temperature sensor of the pH sensor is present 1 Defective or the glass membrane of the pH sensor 1 has changed due to aging or contamination. On the other hand, assuming the validity of the functional relationship between the electrical resistance of the glass membrane and the process temperature, the electrical resistance of the glass membrane can be used as an alternative to determining the state of a measuring probe based on the temperature. In this case, a continuous analysis of this functional relationship between glass membrane resistance and temperature, based on the initial state of the pH sensor, is recommended 1 ,

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited patent literature

• EP 1550861 B1 [0003]

Claims (10)

Method for determining a state of a glass membrane-comprising pH sensor, in which for the state analysis a temperature (T _Proz ) of a process is measured, to which the pH sensor ( 1 ) is exposed, wherein, depending on the temperature (T _Proz ), the state of the sensor ( 1 ), characterized in that in addition to the temperature (T _Mess ) an electrical process resistance (G _Proz ) of the glass membrane ( 4 ) is determined from which the state of the pH sensor ( 1 ) is closed. A method according to claim 1, characterized in that a process resistance of the glass membrane is compared with an initial value of the resistance of the glass membrane at a defined temperature. Method according to claim 1 or 2, characterized in that during a final test of the pH sensor ( 1 ) an electrical resistance (G _measurement ) of the glass membrane ( 4 ) is measured at different temperatures (T _measurement ) and a functional relationship (G _measurement = f (T _measurement )) is determined from the course of a resistance-temperature value pair (G _measurement , T _measurement ). Method according to Claim 3, characterized in that the functional relationship (G _Mess = f (T _Mess )) is determined by means of a mathematical regression. Method according to Claim 4, characterized in that the mathematical regression has an exponential profile of the electrical resistances (G _Mess ) of the glass membrane measured in the final test ( 4 Wherein the exponential curve of the electrical resistance is linearized) above the temperature _{(measurement)} is taken as a basis, and an estimated resistance value (G _'estimation) (as a function of different process temperatures T _{per cent)} is determined. Method according to one of the preceding claims, characterized in that a functional pH sensor ( 1 ) Is closed when a (at a process temperature T _{per cent)} of certain electrical resistance process (G _{per cent)} of the glass membrane ( 4 ) is within a resistance interval. A method according to claim 6, characterized in that when using the pH sensor ( 1 ) in the process, the process temperature (T _proz ) measured and the electrical process _resistance (G _Proz ) of the glass membrane ( 4 ) Are determined and a difference of the determined, logarithmic electrical process resistor (ln (G _Perc)) of the glass membrane and the (for the measured process temperature T is formed _{per cent)} estimated logarithmic resistance value (G _'estimation) and compared with a limit resistance value. A method according to claim 6, characterized in that the resistance interval is determined by means of statistical methods by calculation. Method according to at least one of the preceding claims 6 to 8, characterized in that, if the resistance interval is exceeded once or exceeded by the determined electrical process resistance (G _Proz ) of the glass membrane ( 4 ) a warning signal is issued. Method according to at least one of the preceding claims, characterized in that, if the resistance interval is exceeded or _{fallen below} several times by the determined electrical process resistance (G _Proz ) of the glass membrane ( 4 ) the pH sensor ( 1 ) is replaced.

Images