EP0086337B1 - Process for the control of excess air in firing equipments and control equipment for the realization of the process - Google Patents

Abstract

1. A process for controlling the excess air in a furnace in which the air flow and/or fuel flow is corrected on the basis of measurement of the excess air in the flue gas after a prescribed value/actual value comparison of the excess air in a control device, the proportional sensitivity (gain) of the control device being varied according to the load rate (beta), and at least some of the time parameters of the control device are automatically adapted to the load rate (beta) of the furnace, the prescribed excess air value (w, w') fed as command variable (reference variable) to a controller (11, 28) of the control device being varied according to the load rate (beta), characterized in that the prescribed excess air value (w, w') is fed to the controller (11, 28) through a signal filter (10, 26) having at least one time constant, at least one of the time constants (TF ) being varied according to the load rate (beta).

Description

The present invention relates to a method for regulating the excess air in a furnace, in which, based on the measurement of the excess air in the flue gas after a target / actual value comparison of the excess air in a control device, a corrective intervention on the air flow and / or fuel flow takes place, depending on the proportional gain of the control device is varied by the load level, and at least some of the time parameters of the control device are automatically adapted to the load level of the furnace, the excess air setpoint supplied to a controller of the control device being changed as a function of the load level, and a control device for executing the Method with a first input for a load-dependent signal, a second input for an air excess-dependent signal, a controller, the first input being operatively connected to a differential unit via a function generator and the second input n are and at least one circuit is provided for adapting at least some of the time constants of the control device to the degree of load, which circuit has an adaptation input which is acted upon as a function of the signal at the first input.

Compound controls of furnaces are known, primary air or fuel controls, in which the fuel flow and the air flow as a function of the load level β specified by the operator or a fire power controller, defined as the amount of fuel burned per unit time, based on the maximum combustible fuel amount per unit time two interlinked control devices, one in the fuel and one in the air supply, can be adjusted. Other primary controls for air and fuel are also known, e.g. B. Compound regulations in which the amount of air and the fuel are regulated to a value corresponding to the degree of loading β or ratio arrangements in which the degree of loading β is checked and regulated for a constant fuel-air ratio.

Apart from special cases, neither fuel nor air flow can be set and maintained without errors by such primary control or regulating devices, since numerous random interference effects still have an influence in this regard. In the fuel flow, these are e.g. B. fluctuations in temperature, toughness, density, fuel composition etc., in the airflow fluctuations in pressure, temperature and humidity, which cause deviations from the desired value. There are also device-technical imperfections. For this reason, it has become known to eliminate the effects of such interferences by measuring the actual excess air in the flue gas and a correction of the air flow and / or the fuel flow derived therefrom, thus theoretically with an air-excess setpoint value which is adapted as a function of the load optimal combustion conditions can be maintained.

Attempts have been made again and again to implement such an air excess control using commercially available proportional, integral, proportional-integral or PID controllers with fixed time setting values, but in most cases with unsatisfactory results. Above all, whenever the load level was changed over a wide range, there were either stability problems - pendulum control - or sluggish control effects, even with careful control adjustment, combined with large temporary and / or permanent control deviations or even both at the same time. This is due to the dynamic properties of the controlled system, consisting of the combustion system, including the flue gas duct up to the measuring point of the excess air - the static and dynamic parameters of which are subject to major changes during operation.

Another difficulty lies in the fact that even short-term falls below the minimum air supply cannot be tolerated due to the immediate onset of smoke formation and for safety reasons. However, this is unavoidable with an excess air regulator of the usual type, particularly in the case of rapid and large load changes, since the regulation cannot intervene quickly enough.

Furthermore, from DE-OS 2 753 520 and DE-OS 1 526 277 methods for regulating the excess air in furnaces are known, in which a corrective intervention on the air flow takes place on the basis of the measurement of the excess air in the flue gas. A load-dependent adjustment of the proportional gain of the open control loop is carried out. However, it is known that the transmission parameters of a combustion system including the flue gas duct to the measuring point of the excess air in a combustion system are subject to major changes in the operation of the system. This is especially true for the time parameters resp. Time constants of such a route. Solely, load-dependent adjustment of the proportional gain of the open control loop cannot ensure an even optimally optimal dynamic control effect during operation, because the load-dependent variation of the system time constant changes the frequency response of the open control loop as well as its critical frequency and the phase reserve there.

Now it is known from Art. «Adaptive controls of controller setting with simple means by W. Peinke, in Control Technology, Volume 6 (1966), pages 274 ff. That the characteristic values of process control systems - using the example of an analysis control system shown, which also includes regulations for excess air in furnaces - often on the flow rate and on the load level depend. Based on this, it is proposed there to adjust the readjustment time of a controller as a function of the flow, that is, in the case of excess air control, as a function of the degree of load. Likewise, in «Regulation of combustion by means of a gas / natural gas process computer, magazine of the DVGW German Association of the Gas and Water Industry, year 1980, issue 4, pages 141 ff the process input size is recommended.

Since in such regulations the excess air setpoint must be changed with the degree of load, it is also proposed in the last-mentioned article to functionally link the excess air setpoint with the burner output or the degree of load.

The present invention now aims to improve a method of the type mentioned in terms of its effect.

This is achieved by the method according to claim 1.

By introducing a signal filter, which is used to adjust at least some of the time constants, depending on the degree of load, in addition to the known provision of a control device, by means of which, depending on the degree of load, the proportional gain and at least some of the time parameters are adjusted, the degree of freedom is basically changed the proposed regulation for optimizing their behavior is increased, from which it is readily apparent that more optimization criteria can be met independently of one another.

The problems associated with rapid load changes are further taken into account in that the control effect is supplemented by the load level being applied as a disturbance variable.

A control device for executing the above-mentioned method is distinguished by the wording of claim 6.

The proposed method and the proposed control device for executing the method take into account the difficult properties of the controlled system and the high demands that must be placed on the control quality for economic, ecological and safety reasons. They can be used above all on burner furnaces for solid, liquid and gaseous fuels, as well as on mixed burner furnaces, but also on grate furnaces.

Further features of the invention will become apparent in connection with the description of the figures,

• Figur 1 ein Blockschema der Regeleinrichtung an einer Feuerung, beispielsweise mit Verbundsteuerung,
• Figur 2 eine bevorzugte Realisierungsform der Regeleinrichtung gemäss Fig. 1.

The invention is subsequently explained, for example, using figures. Show it :

• FIG. 1 shows a block diagram of the control device on a furnace, for example with a composite control,
• FIG. 2 shows a preferred embodiment of the control device according to FIG. 1.

1 shows a furnace 1 with a fuel supply line 2 and an air supply line 3. The furnace shown is controlled by a common control system 4 of the usual type. The representation of the invention in connection with a combustion plant with compound control is not intended to imply any restriction on the usability of the method according to the invention and the control device for this. Rather, the invention can be connected to all variants of primary controls and / or controls, by means of which the air supply and / or fuel supply for the firing system is intervened as a function of the load degree β.

The compound controller 4 acts via actuators 5 and. 6 corresponding to the fuel supply in line 2 respectively. the air supply in line 3. 7 shows an adjusting element, by means of which the load level β for the furnace can be set.

The units on the furnace described so far correspond to the conventional equipment of a furnace.

The control device according to the invention now relates to the rest of the device shown in the block diagram according to FIG. 1, with a dashed border. A measuring element 8 is arranged at a suitable point in the flue gas duct la and continuously measures the excess air. This is preferably done by measuring the residual oxygen content in the flue gas. At its output, the measuring element 8 emits a control signal x which is significant for the excess air. The signal at the output of the adjusting unit 7 corresponding to the set degree of load ß, for the sake of simplicity also designated ß, is fed to an input E _{ß of} the control device. The signal β is fed therein to a first function generator 9, at the output of which the signal w appears as a reference variable of the control device. This signal w is fed as a signal w 'to a controller 11 via a reference variable signal filter 10 with the transfer function G (p), together with the controlled variable x. In the controller 11, the control deviation Δx = xw 'and, based on the selected control algorithm, preferably with a PI behavior, the manipulated variable y are formed. This in turn acts on a motor controller 12, which forms the corresponding signals S for forward or backward running of a correction servomotor 13. The signal S _r fed back from the servomotor 13 to the motor controller 12 ensures a proportional association between changes in the manipulated variable y and those of the motor stroke H on the output side of the servomotor 13. The stroke H, as a mechanical signal, is superimposed on a mechanical superposition unit 14 of known type 14 over the actuating stroke h output by the composite controller 4.

The controller 11 respectively. its transfer function are by a gain K _R and by a or several time constants T _{R are} determined. The controller 11 has adaptation inputs, also referred to below as control inputs E _K , E _T , for controlling these time constants. The load level signal β is transmitted via one or more function generators 15a, b to the control inputs E _K , E _{T provided} as signals A _K, respectively. _AT supplied.

The transfer function G (p) of the guide signal filter 10 is determined by one or more filter time constants T _F. The filter now has one or more adaptation or control inputs E _TF , at which the characteristic time constants T _F can be adjusted. For this purpose, the load level signal β is passed through one or more function generators 16, the output _signals A _Tl -A _Tx of which are connected to the control inputs E _{TF provided} on the filter 10.

Finally, the load level signal β is connected to a function generator 17, the output signal A _{17 of} which acts directly on the motor control 12, whereby load-dependent precontrol of the function of the correction servomotor 13 in the sense of a feedforward control is achieved.

In order to adapt the control device to a specific firing system, the function generators 9, 15, 16, 17 can preferably be adjusted with regard to their function course, as shown with the correspondingly indicated intervention signals P.

A preferred form of implementation of the control device described is shown in FIG. 2. An oxygen measuring probe 21 placed in the flue gas stream generates an electrical potential difference corresponding to the 0 ₂ content, which is evaluated in a function generator 22 by delogarithming and amplified in an amplifier 23 to the electrical signal of the controlled variable x. This signal acts on one input of a differential amplifier 24, on the other input of which the command variable w 'acts. The command variable signal w 'is generated as a function of the load degree β by the function generator 25 and passed through an electronic filter 26, preferably with a low-pass characteristic, as the step response shown shows. The control deviation signal Δx is fed to a controller 28, in which it is multiplied in an electronic multiplier 29 by a signal A _K derived from the load level signal β by means of a function generator 30. The multiplication unit 29 achieves the load-dependent adaptation of the controller gain K _R by modulating the control deviation signal Δx with the signal A _K generated by the function generator 30. The output of the multiplier 29 is the one hand, fed directly to a summing amplifier 31, on the other hand to a further multiplying unit 32, in which it is multiplied by a, depending on the load signal β in a function generator 33 generated signal _T A. The output signal of the multiplication unit 32 is fed to an integration unit 34 and its output signal in turn is fed to the summing amplifier 31. With the signal A _{T is} the integral time T, the controller 28 shown here as a PI controller is modulated depending on the load level of the multiplier 32nd

The load level signal β is further passed to a function generator 35 and its output signal A ₃₅ in the sense of a disturbance variable feed back to the summing amplifier 31.

To control the time constant T _F on the filter 26 and adapt its transmission behavior, the load level signal β is passed to the function generator 37, the output signals A _{TF of which are passed} to the corresponding control inputs E _TF on the filter 26. Here all time constants of the filter are modulated in the same way.

The output of the summing amplifier 31 with the signal A ₃ influences control relays 38 and 39, which in turn cause the servomotor 40 to run forwards or backwards. The movements of the servomotor 40 are, by the superposition device 41, as already explained with reference to FIG. 1, the air flap actuating stroke h by a conventional control or. Regulation overlaid. The load level signal β is finally set, for example, by a potentiometer 42. Likewise, the position control of the servomotor 40 takes place via a potentiometer 44, to which the mechanical output movement of the motor is transmitted and which feeds an electrical signal S _r back to the summing amplifier 31. The function generators can in the usual way, such. B. be implemented with diode networks, its function course is preferably adjustable via interventions P. However, it goes without saying that the entire control device can be constructed digitally, analogously or hybrid.

The control device according to the invention described above can of course only achieve the desired optimal control effect if it is also set according to the static and dynamic properties of the controlled system. The basic values of the controller parameters corresponding to full load at β = 1 are to be determined according to the known optimization rules of control theory. For the setting of the functions on the function generators 30, 33, 35 and 37 according to FIG. 2 and 9, 15, 16 and 17 according to FIG. 1 with regard to the load dependency of their output signal, however, the control theory does not provide any information.

The decisive factor here is the realization that the static and dynamic properties of a given controlled system of the type in question practically depend only on the load factor β. This means that the optimal controller parameters are also only dependent on β, ie they can be controlled as a function of β. Now it can be shown further on the basis of thermodynamic and fluidic considerations that the time constants T _{s of} the line, which characterize their dynamic behavior, are indirectly proportional to the load degree β with sufficient accuracy. The dependency of the line gain K _s on the degree of load, which is determined by valve, flap, fan and transmission characteristics, cannot be specified on the other hand, but must be in the system from case to case can be determined experimentally.

• Regler-Verstärkung :
  
  wobei K_o = konstant der optimalen Kreisverstärkung entspricht.
• Regler-Zeitkonstanten :
  
  wobei T_R allgemein charakteristische Zeitkonstanten am Regler bezeichnet, wie Nachstellzeit T, oder Vorhaltezeit Ty.

Based on theoretical considerations, the following relationships result for the load dependency of the controller setting values as a basis for the setting of the function generators:

• Controller gain:
  
  where K _o = constant corresponds to the optimal loop gain.
• Controller time constants:
  
  where T _R denotes generally characteristic time constants on the controller, such as reset time T, or derivative time Ty.

The above time constant relationship can also be used analogously for the load-dependent adaptation of the filter time constant T _F.

• Funktionsgenerator 30 resp.15a :
  
• Funktionsgenerator 33 resp.15b :
  
• Funktionsgenerator 37 resp.16 :
  

This results in the following setting rules, which can be implemented by the corresponding function generators:

• Function generator 30 or 15a:
  
• Function generator 33 or 15b:
  
• Function generator 37 or 16:
  

The setting of the function generator 25, respectively. 9 has to be carried out on the basis of attempts to optimize the furnace, which have to be carried out on a case-by-case basis.

The setting of the disturbance variable function generator 35 and. 17 results from the load-dependent steady-state positions of the correction servo motor 40, respectively. 13, which must also be determined experimentally on a case-by-case basis.

The introduction of the signal filter 10 or 26, by means of which the relevant time constants are adjusted as a function of the load level β, increases the degree of freedom for optimizing the control effect of the control described, since the disturbance and the control behavior of the control can now be optimized separately.

Claims (9)

1. A process for controlling the excess air in a furnace in which the air flow and/or fuel flow is corrected on the basis of measurement of the excess air in the flue gas after a prescribed value/actual value comparison of the excess air in a control device, the proportional sensitivity (gain) of the control device being varied according to the load rate (ß), and at least some of the time parameters of the control device are automatically adapted to the load rate (ß) of the furnace, the prescribed excess air value (w, w') fed as command variable (reference variable) to a controller (11, 28) of the control device being varied according to the load rate (β), characterized in that the prescribed excess air value (w, w') is fed to the controller (11, 28) through a signal filter (10, 26) having at least one time constant, at least one of the time constants (T_F) being varied according to the load rate (ß).

2. A process as claimed in Claim 1, characterized in that the control effect is supplemented by superimprosing the load rate as disturbance variable (17, 35).

3. A process as claimed in one or both of Claims 1 and 2, characterized in that the proportional sensitivity (K_R) (gain) of the control device (11, 28) is automatically varied so that the proportional sensitivity of the open control circuit (K_o) remains constant independent of load.

4. A process as claimed in one or more of Claims 1 to 3, characterized in that at least one controller time constant (T_R) is automatically varied so that the product of load rate (β) and time constant (T_R) remains constant, i. e. independent of load.

5. A process as claimed in one or more of Claims 1 to 4, characterized in that at least one time constant (T_F) of the signal filter (10, 26) is varied so that the product of load rate (ß) and time constant (T_F) remains constant, i. e. independent of load.

6. A control device for carrying out the process claimed in Claims 1 comprising

a first input (Ep) for a load-rate-dependent signal (ß),

a second input (E_x) for a signal (x) dependent on the excess air,

a controller (11, 28),

the first input (E_β) being connected via a function generator (9, 25) and the second input (E_x) being operatively connected to a differential unit (24) and at least one circuit (16, 10, 15b, 11) being provided for adapting at least some of the time constants (T_F, T_R) of the control device to the load rate (ß), said circuit comprising an adaption input activated in dependence upon the signal at the first input (E_β), characterized in that the first input (E_β) is operatively connected via filter (10, 26) to the differential unit (24) and said input (E_o) acting via at least one function generator (16, 37) on at least one adaption input (E_TF) on the filter (10, 26) to vary at least one filter time constant (T_F) in dependence upon the load rate (β).

7. A control device as claimed in Claim 6, characterized in that the first input (E_β) acts on at least one adaption input (E_T) to influence a controller time constant (T_R).

8. A control device as claimed in one or both of Claims 6 and 7, characterized in that the first input (E_o) is connected via at least one function generator (15a, b, 30, 33) with a preferably adjustable (P) function shape (A_K = A_K(P, P_15a); AT = A_T(β, P_15b)...) to the adaption input on the controller (11, 28).

9. A control device as claimed in one or more of Claims 6 to 8, characterized in that the first input (E_β) is operatively connected, preferably via a function generator (17, 35) with a preferably adjustable (P₁₇, P₃₅) function shape (A₁₇ = A₁₇ (β, P₁₇); A35 = A₃₅(β, P₃₅)), to a superimposing unit (12, 31) at the output of the controller (11, 28).

Images