METHOD FOR OPERATING A GAS BURNER AND GAS BURNER FOR CARRYING OUT THE METHOD

Description

[0001] The present invention relates to a method for operating a gas burner and a gas burner, in particular a gas burner for a heating system. The gas burner has a lambda probe arranged in the exhaust gas stream.

[0002] European patent application EP 3 064 937 A1 discloses a generic method which is aimed at detecting faults in the lambda sensor and thus preventing a dangerous operating condition, for example of a gas burner.

[0003] Lambda sensors are used to measure the air / fuel ratio of an exhaust stream from a combustion process. The air / fuel ratio quantifies the residual oxygen content in the exhaust stream. Depending on the measured air / fuel ratio, the combustion process can be controlled to achieve a desired residual oxygen content in the exhaust gas.

[0004] The air / fuel ratio, also known as the combustion air ratio (λ), is a dimensionless quantity calculated as the ratio of the actual available air mass to the minimum required air mass. λ = 1 represents the stoichiometric combustion air ratio at which all fuel molecules can fully react with the atmospheric oxygen without a lack of oxygen required for combustion or leaving unburned fuel. λ > 1 represents an "air excess," also known as a lean mixture. λ < 1 represents an "air deficiency," also known as a rich mixture.

[0005] The functional principle of a broadband lambda sensor is generally known and will only be briefly explained using Fig. 1 summarized. Fig. 1shows a schematic sectional view of a broadband lambda probe 1, which has a pump cell PZ, a measuring chamber MK (or measuring space or measuring gap) and a measuring cell NZ (or reference cell or Nernst cell). The pump cell PZ is placed adjacent to an exhaust gas stream. The measuring chamber MK is located adjacent to the pump cell PZ. The measuring cell NZ for measuring a measuring voltage is arranged next to the measuring chamber MK. Accordingly, the lambda probe 1 is formed in a layered structure from the pump cell PZ, the measuring chamber MK and the measuring cell NZ. The pump cell PZ and the measuring cell NZ each have electrodes 3 that are separated by a solid electrolyte, for example zirconium dioxide. In particular, the lambda probe 1 has a pump cell PZ with a diffusion channel 4 and a regulator VSR. The pump cell PZ is supplied with a pumping current by a power supply, so that the pump cell PZ pumps oxygen from the exhaust gas into or out of the measuring chamber MK.

[0006] The pump cell and the measuring cell are heated to an operating temperature of, for example, approximately 800°C by means of a heater 2. At this temperature, the solid electrolyte of the pump cell PZ and the measuring cell NZ is permeable to oxygen ions.

[0007] The oxygen content of the sample gas in the measuring chamber MK is influenced, on the one hand, by the exhaust gas entering the measuring chamber MK through the diffusion channel 4 and, on the other hand, by the pumping current IP. Depending on the polarity, the pumping current IP pumps oxygen from the exhaust gas through the solid electrolyte into the measuring chamber MK or from the measuring chamber MK to the exhaust gas.

[0008] The pump current IP can be controlled by an external controller so that the air ratio λ in the sample gas precisely balances the oxygen flow through the diffusion channel 4, resulting in a constant value of λ = 1 in the measuring chamber MK. A lambda value of 1 is always present when the voltage VS at the measuring cell NZ is 450 mV. The pump current IP resulting from this control can be used as a sensor signal that is characteristic and meaningful for the oxygen content in the exhaust gas stream.

[0009] Such a broadband lambda sensor is used, for example, to monitor the combustion process in a burner of a gas boiler. The burner can then be controlled based on the measured values ​​of the lambda sensor. Faulty function of the lambda sensor is important for this. If there is a defect in the sensor, the contacts, or the sensor control, the lambda sensor can deliver erroneous signals, which can lead to incorrect burner control. The measured air / fuel ratio then no longer matches the actual air / fuel ratio in the exhaust stream, and undesirable and potentially dangerous concentrations of carbon monoxide or unburned hydrocarbons can be emitted. If the air / fuel ratio in the exhaust stream is very high (λ >> 1), the combustion efficiency can also be dramatically reduced.

[0010] Since such errors or defects can also be caused by aging processes in the lambda sensor, it is necessary to detect faults in the lambda sensor as early and reliably as possible. The present invention is based on the object of providing a method with which a fault in a lambda sensor can be reliably detected in order to prevent the occurrence of a dangerous operating condition in a gas burner.

[0011] According to a first aspect of the invention, the object is achieved by a method according to claim 1. According to a second aspect of the invention, the object is achieved by a gas burner according to claim 9. According to a third aspect of the invention, the object is achieved by a heating system according to claim 10. Further aspects of the invention are the subject of the dependent claims, the drawings and the following description of exemplary embodiments. Short description of the characters

[0012] Further advantageous embodiments are described in more detail below with reference to an embodiment shown in the drawings, to which the invention is not limited, however.

[0013] They show schematically: Figure 1 Fig. 1 shows a sectional view of an exemplary broadband lambda sensor. Figure 2 Fig. 2 illustrates an example control scheme of a gas burner with broadband lambda probe in control operation. Figure 3 Fig. 3 illustrates an example control scheme of a gas burner with broadband lambda probe in probe test mode. Figure 4 Fig. 4 shows an exemplary flow chart of a method for operating a gas burner. Detailed Description of the invention using exemplary embodiments

[0014] In the following description of a preferred embodiment of the present invention, like reference numerals designate like or comparable components.

[0015] Fig. 4 shows a flowchart of an exemplary method for operating a burner, for example, a gas burner of a heating system, which has a broadband lambda probe 1 for exhaust gas monitoring. The method steps are designated by numbers S1 to S29. Transitions between method steps are represented by solid arrows. Dashed arrows illustrate the storage or reading of values ​​measured and / or calculated in the method, which are designated by letters a to i.

[0016] The process can be divided into six functional blocks as follows: F1 Burner in standby (step S1) F2 Burner start with probe calibration in air (steps S2 to S7) F3Controlled burner operation (steps S8 to S17) F4 Controlled burner operation during active probe test mode (steps S18 to S25) F5 Burner shutdown with probe calibration in air (steps S26 - S28) F6 Safety shutdown in case of error (step S29)

[0017] First, the Fig. 2 and Fig. 3 The control schemes shown provide an overview of the process, particularly of the function blocks F3 and F4. A detailed description of the individual process steps S1 to S29 follows based on the Fig. 4 .

[0018] Function blocks F1, F2, F3, F5, and F6 are essentially implemented similarly in prior art methods. Function block F4, i.e., the probe test mode, is the subject of the present invention. The exemplary method according to the invention described below can enable the use of a commercially available, inexpensive, non-intrinsically safe broadband lambda probe, such as those used in the automotive sector, in a safety environment, such as a gas burner, without the aid of additional sensors.

[0019] The measuring method of the broadband lambda sensor 1 is based on operating two cells (measuring cell VZ and pumping cell PZ) and a measuring chamber MK in a separate closed control circuit, which is Fig. 2 is illustrated.

[0020] Fig. 2shows a control diagram of a grass burner with a broadband lambda probe in normal operation (first operating state). The block labeled A represents a higher-level control circuit R1 of a combustion control unit of the gas burner. This regulates the residual oxygen content in the flue gas O2 flue gas and contains an O2 controller O2R, a controlled system RS and a function IP which supplies a setpoint IP setpoint which is calculated using a sensor characteristic curve f(O2) from a setpoint O2 setpoint for the residual oxygen content in the flue gas. The setpoint of the residual oxygen content in the flue gas can be determined, for example, using a desired air ratio λ. A gas burner is usually operated with excess air, i.e. with an air ratio λ > 1. The setpoint O2 setpoint of the residual oxygen content in the flue gas can be determined as a volume concentration from the desired air ratio λ.

[0021] The first operating state (control operation) of the gas burner comprises in particular the steps: measuring (step S12) a residual oxygen content (e) in the exhaust gas; comparing (step S14) the measured residual oxygen content (e) with a predetermined target value (f) and determining a deviation; regulating (step S15) an opening degree of the gas control valve depending on a deviation; measuring the measured voltage at the measuring cell NZ; comparing the measured measured voltage with a predetermined target voltage and determining a deviation; and regulating the residual oxygen content in the exhaust gas via the measured value of the lambda sensor 1. In particular, the pump current can be regulated depending on the deviation. The individual steps are described in more detail below.

[0022] A gas burner is usually operated with a substantially constant air flow. Using a gas control valve, the combustion control system can regulate the volume flow of the supplied gas (fuel) via the O2 regulator (O2R) depending on the deviation between the target IP and the actual IP. With a constant air flow, the ratio of fuel (gas) to air can be regulated via the gas control valve.

[0023] The controlled system RS in Fig. 2 Summarizes the burner's transmission behavior from the gas control valve to the measurement of the residual oxygen content in the flue gas (O2). The control system RS can be influenced by, among other things, the flame gauze, combustion, and heat exchanger.

[0024] The broadband lambda sensor forms, as in Fig. 2(Block B) illustrates, its own closed control loop R2 for oxygen measurement. Exhaust gas with the residual oxygen content O2 exhaust gas enters the measuring chamber MK through the diffusion channel 4 and contributes an oxygen quantity O2 in to the oxygen content O2 MK in the measuring chamber MK. The cell voltage VS ist is measured via the measuring cell NZ. Depending on the pump current IP ist, the oxygen quantity O2 pump is also pumped into the measuring chamber MK or out of the measuring chamber MK through the pump cell PZ. A regulator VSR regulates the measured cell voltage VS ist to the setpoint VS soll = 450 mV. If the cell voltage VS ist is 450 mV, the measuring gas in the measuring chamber MK has an air ratio λ = 1.

[0025] If an error occurs within the oxygen measurement control circuit R2, which could be caused, for example, by aging of the broadband lambda sensor, cracking in one of the cells or blockage of the diffusion channel 4, the burner control unit receives an incorrect measured value IP ist and then regulates an incorrect residual oxygen content in the flue gas O2 flue gas out. In this case, an equilibrium can always be established within the oxygen measurement control circuit R2, so that no error can be detected based on the typically present measured variables (IP ist and VS ist ). Both the pump current IP and the cell voltage VS are controlled variables that are always regulated to the setpoint by the two controllers (VS controller VSR and O2 controller O2R).

[0026] To detect probe errors and the associated deviation in the residual oxygen content of the exhaust gas O2, the operating mode of the gas burner is cyclically changed according to the invention so that additional measured values ​​can be derived from the control circuit R2 of the oxygen measurement. The second operating state is referred to as probe test mode. The control scheme of the probe test mode is based on Fig. 3 described.

[0027] In Fig. 3 The O2 regulator O2R is deactivated. This can be achieved, for example, by setting IP and IP to be virtually equal. This ensures that the residual oxygen content in the flue gas (O2) is kept constant. In other words, the opening degree of the gas control valve remains unchanged, and the burner is operated in controlled mode.

[0028] In the probe test mode, the VS controller VSR is also deactivated by setting VS and VS to be virtually equal. Furthermore, the pump cell PZ is deactivated (see step S18 in Fig. 4 ). This is achieved by switching off the voltage supply of the pump cell PZ so that no pump current flows. Consequently, at a pump current IP of 0 mA, the transport of oxygen molecules via the pump cell PZ into the measuring chamber MK (or out of the measuring chamber MK) is stopped (O2 pump = 0). The inactive elements are in Fig. 3 shown hatched.

[0029] These process steps lead to the oxygen partial pressure in the measuring chamber MK equalizing the oxygen partial pressure in the exhaust gas (O2 MK = O2 exhaust gas), and the cell voltage VS drops significantly. Since the cell voltage VS depends on the oxygen partial pressure in the measuring chamber O2 MK via a specific relationship with sufficient resolution, conclusions can be drawn about the residual oxygen content in the exhaust gas O2 exhaust gas.

[0030] After a defined steady-state period (step S19), a process value VS test_act (g) is derived from the cell voltage VS actual by averaging (S21). This value is calculated in subsequent process steps (S24 and S25) and subsequently monitored for limits. This performs a plausibility check so that errors in the broadband probe are detected if the process value does not lie within specified limits.

[0031] This method can be used to detect defects in the pump cell PZ, as well as in the diffusion channel 4. However, the method requires a fault-free measuring cell VZ. This is continuously monitored separately by measuring the internal resistance R IVS. The required procedure is specified by the manufacturer of the broadband lambda sensor. If this value lies within a tolerance band (S5), the measuring cell VZ is fault-free.

[0032] To ensure this method is completely error-free, a possible cell voltage drift is also adjusted. This is done by calibrating the pump current IP in air before and after each burner operation (probe calibration S6 and S26). If the deviation between both values ​​is within a defined tolerance range, VS test_actual or VS test from the last burner operation is proportionally included in the setpoint VS test_setpoint (S28), resulting in a new, adapted setpoint for VS test.

[0033] In the following, the process steps S1 to S29 are described using Fig. 4 described in detail.

[0034] In step S1 the burner is error-free and is in standby mode.

[0035] In step S2, a check is made to determine whether a burner request is present from a higher-level temperature controller. If a burner request is present (yes in S2), this provides the start signal for burner operation. If there is no burner request (no in S2), the burner remains in standby mode.

[0036] In step S3, the broadband lambda sensor 1 is activated. Activating the broadband lambda sensor 1 is a standard procedure performed according to the manufacturer's specifications. The broadband lambda sensor 1 is first heated using the heating electrodes 2, and then the cell voltage is adjusted to 450 mV.

[0037] If the broadband lambda sensor 1 is ready for operation, the cell resistance R IVS is measured in step S4. This measurement is also a standard procedure performed according to the manufacturer's specifications of the broadband lambda sensor 1. For this purpose, the measuring cell current I CP is switched on and off, while the cell voltage VS is measured, and the internal resistance of the cell (cell resistance) R IVS is calculated from this. The cell resistance R IVS is output as measured value d and stored.

[0038] In step S5, the cell resistance R IVS measured in step S4 is compared with a limit value, which is set, for example, to 400 ohms. If the cell resistance R IVS is less than the limit value (yes), the method proceeds to step S6. A limit value violation (no), i.e., if a cell resistance R IVS greater than the limit value is measured, leads to a safety shutdown in step S29.

[0039] In step S6, a calibration of lambda sensor 1 is performed in air. Once the pump current IP is regulated and constant, a first calibration factor Cal21 1 (measured value a) is calculated using a formula. Kal 21 1 = 4 ⋅ A ⋅ 21 % − C + B 2 − B 2 ⋅ A ⋅ Ip − Offset

[0040] The variables A, B, C and offset are coefficients of the sensor characteristic curve, which are stored in the combustion control unit.

[0041] In step S7, a check is performed to determine whether the calibration factor Cal21 1 (value a) calculated in step S6 lies within a defined tolerance band, for example, by comparing the deviation between the calibration factor Cal21 1 and a setpoint with a limit value Cal21 obs (value c). If the calculated calibration factor Cal21 1 is not within the tolerance band (no in S7), the burner is shut down safely in step S29. If the calculated calibration factor Cal21 1 lies within the tolerance band Cal21 obs (yes in S7), the method continues to step S8.

[0042] Calibration in air is required to compensate for manufacturing tolerances, sensor aging effects, and environmental influences such as atmospheric pressure. A deviation of the calibration factor Cal21 1 from the nominal value, defined as the tolerance band Cal21 obs , is interpreted as a sensor error, resulting in a safety shutdown (step S29).

[0043] In step S8, a check is made to determine whether the burner request is still active. If the burner request is active (yes in S8), the burner is operated in regular mode (continue to step S9). If the burner request is no longer active (no in S8), the process continues in step S26 with a new probe calibration in air.

[0044] The controlled normal operation (first operating state) of the burner corresponds to steps S8 to S17. First, the burner is ignited. Then, in step S9, power modulation begins according to the specifications of the higher-level temperature controller.

[0045] In S10, the probe test mode is activated according to the configured cycle duration. The cycle duration is defined, for example, by the EN 12067 standard and is derived from the hazard potential of a faulty system condition (gas burner) and its duration. A typical cycle duration for a gas burner, for example, is 120 seconds.

[0046] In method step S11, a query is made as to whether the test mode of lambda sensor 1 is requested according to the cycle duration from the previous step S10. If this is not the case (no in step S11), the method proceeds to step S12, and the burner continues to operate in controlled normal mode. If the query indicates that the test mode is requested (yes in step S11), the burner switches to the active test mode, and the method continues to step S18.

[0047] In step S12, the burner is operated in controlled normal mode, with the regular probe control circuit R2 active and providing a measured pumping current IP actual, which is transmitted to the burner control unit. From the measured pumping current IP actual, the actual value of the residual oxygen content (e) in the flue gas can be calculated using a characteristic curve.

[0048] In step S13, the combustion control calculates the actual value of the residual oxygen content O2 (measured value e) in the flue gas from the measured pump current IP ist and the sensor characteristic curve according to the following formula: O 2 ist = A ⋅ IP − Offset 2 ⋅ Kal 21 2 + B ⋅ IP − Offset ⋅ Kal 21 + C

[0049] The variables A, B, C and offset are coefficients of the sensor characteristic curve, which are stored in the combustion control unit.

[0050] In step S14, the O2 controller in the combustion control unit compares the setpoint O2 set (value f) with the calculated actual value of the residual oxygen content O2 actual (value e) from step S13 and determines a deviation. The setpoint O2 set is specified via the parameterization of the combustion control unit. If the two values ​​match (yes), or if the deviation is smaller than a specified limit, the process continues to step S16. If the values ​​are not equal, or if the deviation is greater than the specified limit, the gas quantity V gas, in particular the opening degree of a gas control valve as a manipulated variable, is adjusted in step S15 depending on the determined deviation. The gas quantity

[0051] In step S16, the internal resistance of the measuring cell R IVS (value d) is continuously measured during normal burner operation. This measurement can be performed using the same method as the measurement in step S4.

[0052] Analogous to step S5, the measured cell resistance R IVS (d) is compared with the limit value in step 17. If the limit value is met (yes), the process returns to step 8. If the measured cell resistance R IVS is greater than the limit value (no), the safety shutdown occurs in step S29.

[0053] In step S18, the O2 sensor test mode is active, but the burner's regular operation (O2 controller) is no longer active. The burner is therefore no longer in regulated, but rather in controlled operation. This state typically lasts approximately five seconds. In the sensor control circuit, the power supply to the pump cell is switched off.

[0054] In step S19, an O2 partial pressure equalization is performed between the combustion chamber and the measuring chamber of lambda probe 1. This process can typically take less than one second. A longer time of two seconds can ensure that partial pressure equalization can be performed even under disturbed conditions (e.g., a blocked diffusion passage).

[0055] In step S20, the cell voltage Vs is measured. Instead of 450 mV, this is now typically 60 mV (depending on the residual oxygen content of the exhaust gas).

[0056] In step S21, the measured cell voltage VS actual is averaged over one second to compensate for possible signal noise. The mean value of the cell voltage in test mode VS test_act (g) is saved as an interim result. This value is not used directly for monitoring, but is weighted (see step S24) to make the process more robust against disturbances. If the residual oxygen content O2 actual in the combustion chamber deviates slightly from the target value O2 target at the time of the probe test due to external conditions such as wind or gas pressure fluctuations, this should not influence the result of the probe test too significantly.

[0057] In step S22, the voltage supplies of the pump cell PZ and the VS regulator are reactivated so that the pump current IP returns to the initial level and thus the cell voltage VS of 450 mV is regulated.

[0058] Regulating the cell voltage to 450 mV takes approximately one second. In step S23, an extended delay of, for example, two seconds is performed to ensure increased robustness, since the control response may be slower with an aged lambda sensor 1.

[0059] In step S24, the intermediate result from step S21 VS test_akt (g) is included in the test result VS test (h) using a moving average with a weighting of, for example, 20%.

[0060] In step S25, the result of the probe test VS test (h) is compared with a dynamic limit value VS test _soll (i). If the test result VS test is below the limit value (yes), normal operation continues in step S8. If the dynamic limit value VS test_soll (i) is violated, an immediate safety shutdown occurs in step S29.

[0061] If the burner request is canceled by the higher-level temperature controller, the process continues after the check in step S8 (no) to step S26. In step S26, the flame is first shut off and the gas burner or lambda sensor 1 is re-ventilated. During the re-ventilation, lambda sensor 1 is recalibrated in air. The result of the calibration is a second calibration factor Cal21 2 (b).

[0062] In step S27, a check is carried out to determine whether the second calibration factor Cal21 2 (b) after burner run deviates from the first calibration factor Cal21 1 (a) from burner start-up (see step S6). If the deviation (i.e., the difference between values ​​a and b) is greater than a permissible limit (no in S27), a safety shutdown occurs in step S29. An excessive deviation between calibration factors a and b indicates a fault in lambda probe 1 or the burner system (e.g., flue gas recirculation), meaning safe operation can no longer be guaranteed.

[0063] If the difference between Cal21 1 (a) and Cal21 2 (b) is smaller than the defined limit (yes in S27), lambda sensor 1 is assumed to be fault-free, so that in step S28 a new target value VS test_soll (i) can be determined for the sensor test. This is necessary because the cell voltage VS can drift over its lifetime. This can, in particular, increase the robustness of the process. The target value VS test_soll (i) is corrected or included by 20% with the last result of the sensor test VS test (h). The result is a weighted average that allows long-term correction of the target value so that short-term disturbances do not distort the target value too much. This can further increase the robustness of the process. The new target value VS test_soll is saved as value i and used in the next sensor test (step S25).

[0064] The features disclosed in the above description, the claims and the drawings may be important both individually and in any combination for the realization of the invention in its various embodiments. List of reference symbols

[0065] λAir ratio IPPump current VSCell voltage at the measuring cell VS ist Actual value of the cell voltage VS soll Setpoint of the cell voltage Kal21 1 First calibration factor (a) Kal21 2 Second calibration factor (b) Kal21 obs Tolerance band (c) R IVS Internal resistance of the measuring cell (d) O2 ist Actual value of the residual oxygen content (e) in the flue gas O2 soll Setpoint of the residual oxygen content (f) in the flue gas VS test_akt Average value of VS is in test mode (g) VS test Result of the probe test (h) VS test_soll Setpoint for the probe test (i) O2 pump Transport of oxygen molecules through the pump cell O2 MK Partial pressure of oxygen in the measuring chamber O2 Abgas Partial pressure of oxygen in the flue gas O2R02 Controller RSControl system VSRVS Controller 1 Broadband lambda sensor 2 Heater 3 Electrode 4 Diffusion channel PZ Pump cell MK Measuring chamber NZ Measuring cell (Nernst cell)

Claims

1. Method for operating a gas burner which has a lambda probe (1), arranged in the exhaust gas flow of the gas burner, for measuring a residual oxygen content in the exhaust gas, wherein the gas burner is operated in a first operating state in normal operation, wherein the residual oxygen content in the exhaust gas is regulated by means of the measured value of the lambda probe (1), characterized in that the gas burner carries out a fault test of the lambda probe (1) in a second operating state, comprising the steps: determining a present test voltage (h) of a measuring cell (NZ) of the lambda probe (1) when the power supply of a pump cell (PZ) of the lambda probe (1) is switched off; comparing (S25) the present test voltage (h) with a predefined test setpoint voltage (i) and determining a deviation; and if the deviation exceeds a predefined threshold, carrying out a safety switch-off (S29); or if the deviation does not exceed the predefined threshold, operating the gas burner in the first operating state.

2. Method according to claim 1, wherein: the lambda probe (1) is a broadband lambda probe; the gas burner further has an automatic furnace for regulating or controlling an amount of gas supplied to the gas burner via a gas regulating valve; and the automatic furnace sets the opening degree of the gas regulating valve to a predetermined value in the second operating state.

3. Method according to claim 1 or 2, wherein the gas burner is switched from the first operating state to the second operating state cyclically after a predetermined time interval has elapsed (S10).

4. Method according to at least one of the preceding claims, wherein determining the present test voltage (h) comprises: measuring (S20) a measuring voltage at the measuring cell (NZ); forming (S21) a temporal mean value (g) of the measured measuring voltage over a defined period of time; calculating (S24) the present test voltage (h) by forming a weighted mean value from the temporal mean value of the measuring voltage (g) and a stored test voltage of a previous fault test of the lambda probe (1); and storing the present test voltage (h).

5. Method according to claim 4, wherein in the second operating state, after forming the temporal mean value (g) of the measured measuring voltage, the method further comprises: switching on (S22) the power supply of the pump cell (PZ); measuring the measuring voltage at the measuring cell (NZ); comparing the measured measuring voltage with a predefined setpoint voltage and determining a deviation; and regulating the residual oxygen content in the exhaust gas as a function of the deviation.

6. Method according to at least one of the preceding claims, wherein the gas burner carries out a probe calibration (S26) in air in a third operating state if there is no active burner request, comprising the steps: switching off the flame; re-aerating the lambda probe (1); calculating a first calibration factor (b); comparing (S27) the first calibration factor (b) with a stored second calibration factor (a) and determining a deviation; and if the deviation exceeds a predefined threshold, carrying out the safety switch-off (S29) of the gas burner; or if the deviation does not exceed the predefined threshold, calculating (S28) the test setpoint voltage (i) by forming a weighted mean value from a stored test setpoint voltage and the present test voltage (h) determined in the second operating state; and storing the test setpoint voltage (i).

7. Method according to at least one of the preceding claims, further comprising: measuring (S4, S16) an internal resistance (d) of the measuring cell (NZ); comparing (S5, S17) the measured internal resistance (d) with a predefined setpoint value and determining a deviation; and if the deviation exceeds a predefined threshold, carrying out a safety switch-off (S29); or if the deviation does not exceed the predefined threshold, operating the gas burner in the first operating state.

8. Method according to at least one of the preceding claims, further comprising: carrying out a probe calibration (S6) in air and determining a second calibration factor (a); storing the second calibration factor (a); comparing (S7) the second calibration factor (a) with a predefined setpoint value (c) and determining a deviation; and if the deviation exceeds a predefined threshold, carrying out a safety switch-off (S29); or if the deviation does not exceed the predefined threshold, operating the gas burner in the first operating state.

9. Gas burner for a heating system, comprising: a lambda probe (1), arranged in the exhaust gas flow of the gas burner, for measuring a residual oxygen content in an exhaust gas of the gas burner, wherein the lambda probe (1) has a measuring cell (NZ) and a pump cell (PZ); and an automatic furnace for regulating a volume flow of a supplied fuel by means of a gas regulating valve, characterized in that the gas burner is configured to carry out a method according to at least one of claims 1 to 8.

10. Heating system for a building, characterized in that the heating system comprises a gas burner according to claim 9.