EP0644376B1 - Method and apparatus for controlling a burner - Google Patents

Description

The invention relates to a method and a device for regulating a burner for a burner-operated furnace, which can be switched in stages or in a modulating manner in accordance with the preamble of claims 1 and 6.

Energy savings and environmental protection are of central importance today in burner-operated combustion plants, since even with optimally set gas, oil or other burner plants, it can never be completely prevented that pollutants are generated and valuable energy is lost during the operating period. Since it is known that the residual oxygen or 02 content in the exhaust gases is a measure of the quality and efficiency of the combustion, it is becoming increasingly common Firing systems equipped with O ₂ controls. The ratio of the amount of air supplied to the amount of fuel supplied is regulated by means of an O ₂ measuring probe in the exhaust duct and a control device such that the concentration of the residual oxygen in the exhaust gases reaches the setpoint. In the ideal case, an almost stoichiometric combustion of the fuel is achieved.

Since it is also known that the sum of the pollutants (CO, NO _x ) in the exhaust gases essentially reaches a minimum where the almost complete combustion of the fuel takes place, the pollutant emissions of a combustion system can also be considerable with an O _{2 control} reduced and kept within prescribed limits.

In combustion plants with burners, which are often operated at different power levels, the problem of non-optimal combustion and the limitation of pollutant emissions arise when switching from a first power level to a second power level. In large plants in particular, pollutant emissions during the switchover phase can be considerably above the normal values that occur during operation at a fixed power level. The reason for such deviations is that the control processes cannot be fully controlled during the transitions.

If O ₂ controls are used, this requires that the control parameters of the control device are generally determined individually for each of the available power levels using a separate method. When switching from a first power level to a second power level, the control parameters of the control device must be switched over, which causes discontinuities during the transition and ultimately means non-optimal combustion during the transitions.
The data sheet 7851D from September 1990 by Landis & Gyr describes an oxygen controller of the type RWF61.190 with signal-dependent control parameters for modulatable burners. It is a PI controller with signal-adaptive behavior for 02 controls in burner systems. Modular means that the burner can be switched continuously in terms of output. Of course, different power levels are also effective with modulating control. The control parameters for this controller are determined individually from measurements of step responses on the open control circuit at two power levels, namely at maximum burner output and at minimum burner output. The control parameters determined in this way are then set on corresponding potentiometers of the controller. The control parameters for a PI controller are the control gain K _R and the reset time T _N. The control gain K _R and the reset time T _N are set for a PI controlled system with dead time using the Ziegler and Nichols method (see, for example, in this regard in the handbook for radio frequency and electrical engineers, volume IV, page 596, formula 134; by the publishing house for Radio-Foto-Kinotechnik GmbH, Berlin-Bosigwalde).

The operation of this PI controller is such that a control variable is calculated from the control deviation, which represents the difference between the O ₂ setpoint and the actual O ₂ value, which contains a proportional component, an integral component and a power-dependent gain factor. The actuating variable controls the actuator, for example an air flap that regulates the ratio of air to fuel.

A PI control of this type has the disadvantage that in the case of a gradual changeover from a first power level to a second power level in the transition stage, the manipulated variable, which in the regulated state essentially corresponds to the integral part of the PI controller, initially in the new operating state on the second Power level is taken over. However, since the integral part of the PI controller contains a part proportional to the respective power level, its size has not yet been adjusted to the new operating conditions of the second power level, so it initially has an incorrect value. Although this incorrect value is subsequently corrected, it causes more pollutants to be emitted in the transition stage due to non-optimal combustion.

A controller is known from JP-A-57-174 618, in which a load-dependent setpoint for the oxygen content is generated and in which the resulting control deviation is processed by a PID controller. This PID controller (in particular the differential component) serves to prevent the blower to be overshooted, which has a high inertia and therefore a not inconsiderable dead time. Here, too, the integral part of the PID controller contains a part proportional to the respective power level, the size of which is not matched to the new operating conditions of the second power level, so that the controller initially has an incorrect value when the power changes. Since the output signal of the PID controller is therefore not power-independent, an overshoot or undershoot or a momentary unfavorable regulation cannot be prevented during the transition phase.

The object of the invention is to provide a method and a device for regulating a burner for a burner-operated combustion system which can be switched in stages or modulating in terms of output, and which has a more favorable course of combustion during the transition between the output stages.

This object is achieved by the teaching given in the characterizing part of patent claims 1 and 6.

The solution is based on the fact that a performance-independent manipulated variable is calculated.

The advantage of the invention is that when switching between different burner outputs, fewer pollutants are generated and energy losses are reduced.

Fig. 1

eine Brenneranlage mit O₂-Regelung in schematischer Darstellung,

Fig. 2

den O₂-Regelkreis in schematischer Darstellung, und

Fig. 3

die Messung der Sprungantwort am offenen O₂-Regelkreis.

The new method is explained in more detail below with the aid of figures. Show it

Fig. 1

a burner system with O _{2 control} in a schematic representation,

Fig. 2

the O ₂ control loop in a schematic representation, and

Fig. 3

the measurement of the step response on the open O ₂ control loop.

Figure 1 shows a burner system with O _{2 control} in a schematic representation, with a boiler 1, a burner 2, which can be switched between several power levels, and an exhaust gas duct 3. The burner 2 has a fuel supply 4 and an air supply 5, in which Air supply 5 there is an actuator, for example an air flap 6 or a fan, for adapting the quantity of air supplied to the quantity of fuel supplied. Exhaust gases 7 from the combustion are passed on via the exhaust duct 3. There is an O ₂ probe 8 in the exhaust gas duct 3, which measures the residual oxygen content (O ₂ ) in the exhaust gas 7. The actual O ₂ values O _2I measured by the O ₂ probe 8 are fed to a control device 9, where they have an O ₂ target value O _{2S are} compared, whereupon the air flap 6 is controlled by the control device 9 on the basis of the performance of the burner 2 and the difference determined. The optimal air supply or the optimal excess air for combustion is performance-dependent. The control device 9 serves the purpose of controlling the amount of air supplied to the burner 2 such that the residual oxygen content (O ₂ ) measured in the exhaust gas 7 reaches the set O ₂ setpoint O _2S . In the ideal case, an almost stoichiometric combustion of the fuel is achieved, which also means that the pollutant content in the exhaust gas 7 with respect to the gases CO and NO _{x is} minimal.

FIG. 2 shows the O ₂ control loop and its dependence on the power P _{B of} the burner 2 schematically. The O ₂ setpoint O _2S , which is dependent on the power P _{B of} the burner, is compared in a comparator 10 with the actual O ₂ value O _2I , ie a control deviation 11 results from the difference between the two values. Controller 12 supplied, which first calculates a performance-independent manipulated variable Y _R from the control deviation 11. The PID controller 12 also receives control information 14, such as the power P _{B of} the burner 2 and the type of fuel. The power-independent manipulated variable Y _R is converted by a correction element 12a into a manipulated variable 13 dependent on the power P _{B of} the burner. The manipulated variable 13 is fed to the actuator, in this case the air flap 6, the setting value of which, namely an air flap position 15, influences a controlled system 16. The control path 16 is to be understood as the entirety of air supply 5, fuel supply 6, burner 2, boiler 1 and exhaust duct 3 up to the O ₂ probe 8. At the point in the exhaust gas duct 3 at which the O ₂ probe 8 is located, the result of the chain of action of the PID controller 12, the correction element 12a, the air damper 6 and the controlled system 16 results in a specific O ₂ value which is the controlled variable O ₂ represents and the is measured by the O ₂ probe as the actual O ₂ value O _2I and is returned to comparator 10.

The PID controller 12 is used because both permanent and temporary deviations from the O ₂ setpoint can be kept to a minimum.

• Die Regelverstärkung K_R    (P-Anteil)
• Die Nachstellzeit T_N    (I-Anteil)
• Die Vorhaltzeit T_V    (D-Anteil)

The control parameters are determined individually for the PID controller for each power level. The control parameters of the PID controller 12 are:

• The control gain K _R (P component)
• The reset time T _N (I component)
• The lead time T _V (D component)

$${\text{T}}_{\text{N}}\text{=}\frac{{\text{(T}}_{\text{U}}{\text{+ T}}_{\text{S}}{\text{/2)}}^{\text{2}}}{{\text{(T}}_{\text{U}}{\text{+ T}}_{\text{S}}\text{)}}{\text{≈ T}}_{\text{U}}$$

$${\text{T}}_{\text{V}}\text{=}\frac{{\text{(T}}_{\text{U}}{\text{+ T}}_{\text{S}}\text{)}}{\text{2}}\text{≈}\frac{{\text{T}}_{\text{U}}}{\text{2}}$$

mit $${\text{K}}_{\text{S}}\text{=}\frac{{\text{ΔO}}_{\text{2IST}}}{{\text{Δ}}_{\text{POS}}}$$

worin:

• T_U= Verzugszeit
• T_G= Ausgleichszeit
• T_S = Abtastzeit (beispielsweise 0.2 s)
• ΔO_2IST = Differenz des O2-Wertes im Abgas
• Δ_POS = Differenz der Luftklappenposition
• K_S = Streckenverstärkung

The control gain K _R and the reset time T _N and the lead time T _V are set for the PID controlled system with dead time essentially according to the method of Ziegler and Nichols (see, for example, in the Handbook for High Frequency and Electrical Engineers, Volume IV, page 596 , Formula 136; by the Verlag für Radio-Foto-Kinotechnik GmbH, Berlin-Bosigwalde). Since the regulation is microprocessor-controlled and thus represents a sampling regulation, the formulas also include a sampling time T _S.
thereby: $${\text{K}}_{\text{R}}\text{=}\frac{\text{1}}{{\text{K}}_{\text{S}}}\text{1.2}\frac{{\text{T}}_{\text{G}}}{{\text{(T}}_{\text{U}}{\text{+ T}}_{\text{S}}\text{)}}{\text{≈ K}}_{\text{R}}\text{=}\frac{\text{1}}{{\text{K}}_{\text{S}}}\text{1.2}\frac{{\text{T}}_{\text{G}}}{{\text{T}}_{\text{U}}}$$

$${\text{T}}_{\text{N}}\text{=}\frac{{\text{(T}}_{\text{U}}{\text{+ T}}_{\text{S}}{\text{/ 2)}}^{\text{2nd}}}{{\text{(T}}_{\text{U}}{\text{+ T}}_{\text{S}}\text{)}}{\text{≈ T}}_{\text{U}}$$

$${\text{T}}_{\text{V}}\text{=}\frac{{\text{(T}}_{\text{U}}{\text{+ T}}_{\text{S}}\text{)}}{\text{2nd}}\text{≈}\frac{{\text{T}}_{\text{U}}}{\text{2nd}}$$

With $${\text{K}}_{\text{S}}\text{=}\frac{{\text{ΔO}}_{\text{2IST}}}{{\text{Δ}}_{\text{POS}}}$$

wherein:

• T _U = delay time
• T _G = equalization time
• T _S = sampling time (e.g. 0.2 s)
• ΔO _2IST = difference in the O2 value in the exhaust gas
• Δ _POS = difference in air damper position
• K _S = line reinforcement

The values for TU, TG, ΔO _2IST and Δ _{POS are obtained} from measurements of step responses on the open control loop according to FIG. 3. At the open control loop, ie with the burner running at a certain power level, but with the O _{2 control} switched off, it is initially necessary to wait until the O ₂ value measured by the O ₂ probe 8 is stable. Then the position of the air flap 6 is changed abruptly by the value Δ _POS and maintained until the O ₂ value measured by the O ₂ probe 8 is stable again and has set itself to a sufficiently different value than the starting position. This is the starting position for measuring the step response.

Then the position of the air flap 6 is changed again suddenly by the value Δ _POS , but in the opposite direction, and is again maintained until the O ₂ value measured by the O ₂ probe 8 is stable. The difference between the measured O ₂ value before the sudden change in the position of the air flap 6 and the measured O ₂ value after the sudden change in the same is ΔO _2IST . From the course of the O ₂ measurement curve, the points in time are determined at which the measured change in the O ₂ value has reached 10% or 63% of the final value of ΔO _2IST . This gives the delay time T _U and the compensation time T _G , where T _U essentially the Response time or dead time of the controlled system 16 represents. For large burner systems, values of 5 to 10s result for T _U.

Of course, the control parameters determined in this way for the PID controller must be determined and stored for each fuel used and for each output level of the burner 2. The current control parameters to be used are communicated to the PID controller via the control information 14 (see FIG. 2).

The process for controlling burner 2 using the PID control algorithm is as follows:

wobei:

• Y_PR = leistungsunabhängiger Proportionalteil
• Y_IR = leistungsunabhängiger Integralanteil
• Y_DR = leistungsunabhängiger Differentialanteil

The PID controller 12 calculates a power-independent manipulated variable Y _R from the control deviation 11 (see also FIG. 2). $${\text{Y}}_{\text{R}}{\text{= Y}}_{\text{PR}}{\text{+ Y}}_{\text{IR}}{\text{+ Y}}_{\text{DR}}$$

in which:

• Y _PR = performance-independent proportional part
• Y _IR = integral part independent of power
• Y _DR = power-independent differential component

   wobei

• e(n) = Regelabweichung 11 zum Zeitpunkt (n)
• T_G= Ausgleichszeit
• T_U = Verzugszeit

The performance-independent proportional part Y _{PR is} calculated as follows: $${\text{Y}}_{\text{PR}}\text{(n) = e (n) 1.2}\frac{{\text{T}}_{\text{G}}}{{\text{T}}_{\text{U}}}$$

in which

• e (n) = control deviation 11 at time (n)
• T _G = equalization time
• T _U = delay time

The control deviation 11 is designated in this and in the following formulas with the letter e. Because the scheme is microprocessor-controlled, the formulas are listed in the discrete form of representation suitable for the calculation with microprocessors, namely with reference to the calculated or measured values at a respective discrete sampling time (s).

   wobei

• Y_IR(n-1) = leistungsunabhängige Integralanteil Y_IR zum Zeitpunkt (n-1)
• e(n) = Regelabweichung 11 zum Zeitpunkt (n)
• T_G= Ausgleichszeit
• T_U = Verzugszeit
• T_S = Abtastzeit
• T_N = Nachstellzeit

The power-independent integral component Y _IR at time (n) is calculated as follows:

in which

• Y _IR (n-1) = power-independent integral component Y _IR at time (n-1)
• e (n) = control deviation 11 at time (n)
• T _G = equalization time
• T _U = delay time
• T _S = sampling time
• T _N = reset time

   wobei

• e(n) = Regelabweichung 11 zum Zeitpunkt (n)
• e(n-1) = Regelabweichung 11 zum Zeitpunkt (n-1)
• T_G= Ausgleichszeit
• T_U = Verzugszeit
• T_V = Vorhaltezeit
• T_S = Abtastzeit

The power-independent differential component Y _{DR is} calculated as follows:

in which

• e (n) = control deviation 11 at time (n)
• e (n-1) = control deviation 11 at time (n-1)
• T _G = equalization time
• T _U = delay time
• T _V = lead time
• T _S = sampling time

It is important for the invention that the proportional component Y _PR , the integral component Y _IR and the differential component Y _DR in the forms set out above are determined as performance-independent components. When switching from a first power level to a second power level, the integral component must be available in a power-independent form.

   wobei:

• Y = Stellgrösse 13
• Y_R = leistungsunabhängige Stellgrösse
• K_S = Streckenverstärkung

Only then is the output-independent manipulated variable Y _R multiplied by 1 / K _S. K _S is the route reinforcement and performance-dependent. Since K _{S is} in many cases essentially inversely proportional to the power of the burner 2 at the selected power level, the manipulated variable 13 is essentially proportional to the power of the burner 2 at the selected power level in these cases. $$\text{Y =}\frac{\text{1}}{{\text{K}}_{\text{S}}}{\text{· Y}}_{\text{R}}$$

in which:

• Y = manipulated variable 13
• Y _R = output-independent manipulated variable
• K _S = line reinforcement

The manipulated variable 13 is fed to the air flap 6 and causes a change in the air flap position. As a result, the air supply to the burner 2 changes, which ultimately manifests itself with a delay, the reaction time of the controlled system 16, as a change in the O ₂ content in the exhaust gas 7 in the O ₂ probe 8.

In the regulated state, the conditions are stable, that is to say the O ₂ content in the exhaust gas 7 measured by the O ₂ probe 8 remains constant. Since the actual value of the control corresponds to the setpoint of the control (O _2S = O _2I , see also FIG. 2), the control deviation 11 is zero. Only the integral part Y _IR contributes to the maintenance of this state.

• |e(n)| = Regelabweichung 11 zum Zeitpunkt (n)
• NZO₂ = Neutralbereich

If the control deviation 11 is smaller than the O ₂ difference that results for the smallest possible actuation step, no correction is made, since otherwise the control system tends to oscillate.
if | e (n) | <NZO _2/2 then e (n) = 0 is set
in which

• | e (n) | = Control deviation 11 at time (s)
• NZO ₂ = neutral range

The control deviation 11 included in the calculation differs depending on whether there is a lack of air or excess air:
in case of lack of air:
if e (n) <0 then e (n) = e (n)
with excess air:
if e (n)> 0 then e (n) = e (n) / 2
in which
e (n) = control deviation 11 at time (n)

This fulfills the wish to slow down the control in case of excess air in order to avoid undershoots in the CO range as far as possible.

When the burner 2 is switched from a first power level to a second power level, the power-independent integral part Y _{IR of} the power-independent manipulated variable Y _{R is used} as the initial value for the power-independent manipulated variable Y _R at the second power level. This is illustrated below by considering the manipulated variable 13 before and after the power level changeover:

   wobei

• Y_a = leistungsabhängige Stellgrösse 13 im ausgeregelten Zustand bei Betrieb auf der Leistungsstufe A
• K_Sa = Streckenverstärkung für die Leistungsstufe A
• Y_IR = leistungsunabhängiger Integralteil

Actuating variable 13 before switching the power level: $${\text{Y}}_{\text{a}}\text{=}\frac{\text{1}}{{\text{K}}_{\text{Sat}}}{\text{· Y}}_{\text{IR}}$$

in which

• Y _a = output-dependent manipulated variable 13 in the regulated state when operating at output level A.
• K _Sa = distance reinforcement for performance level A
• Y _IR = integral part independent of power

Since the control is in the regulated state, the actual value of the controlled variable corresponds to the setpoint of the controlled variable (O _2S = O _2I , see also FIG. 2). The control deviation 11 is zero. Only the integral component Y _IR makes a contribution to maintaining this state.

   wobei

• Y_b = leistungsabhängige Stellgrösse 13 im ausgeregelten Zustand bei Betrieb auf der Leistungsstufe B
• K_Sb = Streckenverstärkung für die Leistungsstufe B
• Y_IR = leistungsunabhängiger Integralteil

Actuating variable 13 immediately after switching the power level: $${\text{Y}}_{\text{b}}\text{=}\frac{\text{1}}{{\text{K}}_{\text{Sb}}}{\text{· Y}}_{\text{IR}}$$

in which

• Y _b = output-dependent manipulated variable 13 in the regulated state when operating at output level B.
• K _Sb = line reinforcement for performance level B
• Y _IR = integral part independent of power

Assuming that the O ₂ setpoint (O _2S ) for the control is not changed during the switchover, the control deviation 11 remains initially zero due to the reaction time of the control. Only the power-independent integral component Y _IR , which is retained during the transition provides a contribution to maintaining the state. The route gain K _Sb , which is decisive for operation at the new performance level B, is applied at the beginning of operation at the new performance level B. This measure ensures that the integral part taken over during the transition is not distorted by old control parameter influences, which are no longer valid for operation at the current power level and which would otherwise have to be corrected first. This improves combustion during the transition from one power level to the next during operation, and this also results in lower pollutant emissions.

Claims (6)

1. A method of regulating a burner (2) for a firing installation, which burner can be switched over in respect of output in a staged or modulating mode, comprising:

   - an O₂-probe (8) for measuring the residual oxygen content in the flue gases, wherein the measured residual oxygen content is used as an O₂-actual value (O_2I) in an O₂-regulation system,

   - an adjusting member (6) for regulating the air excess to the burner (2), and

   - a regulator (12, 12a) for calculating an adjusting parameter (13) for control of the adjusting member (6) on the basis of a regulation deviation (11) which is calculated from the difference between the output-dependent O₂-reference value (O_2S ) and the O₂-actual value (O₂I), wherein regulating parameters for the existing output stages of the burner (2) are individually ascertained from measurements of jump responses at the open regulating circuit and are used in the regulator (12, 12a) in dependence on the output stage at which the burner (2) is operated,

      characterised in that

   a PID-regulator (12) is used,

   the PID-regulator (12) calculates an output-independent adjusting parameter (Y_R) from the regulation deviation (11),

   the output-independent adjusting parameter (Y_R) comprises an output-independent proportional component (Y_PR), an output-independent differential component (Y_DR) and an output-independent integral component (Y_IR), and

   the product of the output-independent adjusting parameter (Y_R) and an output-dependent system amplification gives the adjusting parameter (13).
2. A method according to claim 1 characterised in that when the burner (2) is switched over from a first output stage to a second output stage the output-independent integral component (Y_IR) of the output-independent adjusting parameter (Y_R) is used as the initial value for the output-independent adjusting parameter (Y_R) in the second output stage.
3. A method according to claim 1 or claim 2 characterised in that when there is an air excess regulation is slowed down by a procedure whereby when there is an air excess the regulation deviation (11) is multiplied by a factor <1 and is then passed to the PID-regulator (12).
4. A method according to one of the preceding claims characterised in that the output-dependent system amplification is substantially inversely proportional to the output of the burner (2) at the selected output stage.
5. A method according to one of the preceding claims characterised in that the adjusting member (6) is an air flap.
6. Apparatus for regulating a burner (2) for a firing installation, which burner can be switched over in respect of output in a staged or modulating mode, comprising:

   - an O₂-probe (8) for measuring the residual oxygen content in the flue gases, wherein the measured residual oxygen content can be used as an O₂-actual value (O_2I) in an O₂-regulation system,

   - an adjusting member (6) for regulating the air excess to the burner (2), and

   - regulator (12, 12a) for calculating an adjusting parameter (13) for control of the adjusting member (6) on the basis of a regulation deviation (11) which can be calculated from the difference between the output-dependent O₂-reference value (O_2S) and the O₂-actual value (O_2I) wherein regulating parameters for the existing output stages of the burner (2) can be individually ascertained from measurements of jump responses at the open regulating circuit and can be used in the regulator (12, 12a) in dependence on the output stage at which the burner (12) is operated.

      characterised in that

   the regulator (12) is a PID-regulator,

   the PID-regulator (12) calculates an output-independent adjusting parameter (Y_R) from the regulation deviation (11),

   the output-independent adjusting parameter (Y_R) comprises an output-independent proportional component (Y_PR), an output-independent differential component (YDR) and an output-independent integral component (Y_IR), and

   the adjusting parameter (13) comprises the product of the output-independent adjusting parameter (Y_R) and an output-dependent system amplification.

Images