Self-adaptive optimization control method for thermal process of thermal power generating unit

Abstract

The invention relates to the field of optimization control, and discloses an adaptive optimization control method for a thermal process of a thermal power generating unit, which is used for improving the control quality and adaptability of the thermal process of the thermal power generating unit. The method includes calculating deviation between a measured value and a set value of a controlled variable; generating a total disturbance real-time estimated value of the system by using a quasi-continuous disturbance observer, and performing feedforward calculation to generate a feedforward compensation signal according to the total disturbance real-time estimated value; a self-adaptive compensation signal is generated by combining the real-time disturbance estimation value and deviation integral learning; and finally, fusing the output of the basic controller, the feedforward compensation signal and the adaptive compensation signal to form a final control instruction. According to the method, the operation load is monitored to call the parameters, the parameter mapping table is constructed and maintained, and the feed-forward gain coefficient and the adaptive learning rate are optimized, so that the adaptability to working condition change, the anti-disturbance capability and the control effect on a large delay link are effectively improved.

Description

Technical Field

[0001] This invention relates to the field of optimization control, and more particularly to an adaptive optimization control method for the thermal process of a thermal power unit. Background Technology

[0002] In the operation of thermal power units, thermal process control is a crucial link in ensuring the safe, stable, and efficient operation of the units. The thermal processes of thermal power units involve numerous complex physical and chemical processes, and the controlled objects exhibit strong nonlinearity, large delays, and time-varying characteristics. Furthermore, the operating conditions change frequently with variations in unit load. Simultaneously, thermal power units are inevitably affected by various internal and external disturbances during actual operation, such as changes in fuel quality, fluctuations in grid load, and changes in ambient temperature. These factors pose a significant challenge to the precise control of the thermal processes in thermal power units.

[0003] Traditional control methods, such as proportional-integral-derivative (PID) control, while simple in structure and easy to implement, have significant limitations when dealing with complex systems like thermal power units. The parameters of a PID controller are typically tuned for specific operating conditions. When the unit's operating conditions change significantly, the original control parameters may fail to guarantee good control performance, leading to decreased control quality and potentially even system instability. Furthermore, traditional control methods struggle to effectively handle various disturbances in the system and are not ideal for controlling large time-delay components, failing to meet the requirements of modern thermal power units for efficient and stable operation.

[0004] With the continuous development of control theory and computer technology, some advanced control strategies, such as model predictive control and adaptive control, have been gradually applied to the thermal process control of thermal power units. However, these methods also have some problems in practical applications. Model predictive control requires the establishment of an accurate system model, but the complexity and time-varying nature of the thermal process in thermal power units make accurate modeling very difficult, and model errors directly affect the control effect. Although adaptive control can automatically adjust control parameters according to the system operating conditions, its adaptive speed and anti-interference ability still need to be improved when dealing with strong disturbances and large delays.

[0005] Therefore, we propose an adaptive optimization control method for the thermal process of thermal power units to solve the above problems. Summary of the Invention

[0006] This invention provides an adaptive optimization control method for the thermal process of thermal power units, which is used to improve the control quality and adaptability of the thermal process of thermal power units.

[0007] The first aspect of this invention provides an adaptive optimization control method for the thermal process of a thermal power unit. The method includes: calculating the deviation between the measured value and the setpoint of the controlled variable in the target control loop of the thermal process; generating a real-time estimate of the total system disturbance based on the measured value of the controlled variable and the output signal of the basic controller using a quasi-continuous disturbance observer; performing feedforward calculation based on the real-time disturbance estimate to generate a feedforward compensation signal; performing integral learning by combining the real-time disturbance estimate and the deviation to generate an adaptive compensation signal; and fusing the output signal of the basic controller, the feedforward compensation signal, and the adaptive compensation signal to form a final control command.

[0008] Optionally, in a first implementation of the first aspect of the present invention, the controlled variable measurement signal of the target control loop is acquired in real time; the optimized setpoint of the controlled variable is determined according to the current operating load of the thermal power unit; the controlled variable measurement signal is filtered to obtain the filtered measurement value; and the real-time deviation between the filtered measurement value and the optimized setpoint is calculated.

[0009] Optionally, in a second implementation of the first aspect of the present invention, a nominal dynamic model of the target control loop is constructed; the model output value is calculated based on the output signal of the basic controller and the nominal dynamic model; the output deviation between the measured value of the controlled variable and the model output value is calculated; and the output deviation is subjected to inertial filtering to generate a real-time estimate of the total system disturbance.

[0010] Optionally, in a third implementation of the first aspect of the present invention, the feedforward gain coefficient is determined based on the current operating state of the thermal power unit; the feedforward gain coefficient is multiplied by the real-time disturbance estimate to calculate the feedforward compensation amount; the feedforward compensation amount is subjected to amplitude limiting processing to generate a feedforward compensation signal.

[0011] Optionally, in the fourth implementation of the first aspect of the present invention, a first static feedforward gain is determined by looking up a table based on the correspondence between the current operating load of the thermal power unit and the preset load range; a second dynamic feedforward gain is calculated based on the rate of change of the real-time disturbance estimate; and the first static feedforward gain and the second dynamic feedforward gain are weighted and synthesized to generate the final feedforward gain coefficient.

[0012] Optionally, in a fifth implementation of the first aspect of the present invention, an adaptive learning rate is determined based on the magnitude of the real-time disturbance estimate; an adaptive parameter is generated by integrating the adaptive learning rate with the deviation; the adaptive parameter is combined with the direction information and amplitude information of the real-time disturbance estimate to calculate an adaptive compensation amount; and the adaptive compensation amount is subjected to amplitude limiting processing to generate an adaptive compensation signal.

[0013] Optionally, in a sixth implementation of the first aspect of the present invention, a basic learning rate is obtained from a preset load-rate mapping relationship based on the current operating load of the thermal power unit; the changing trend of the deviation of the controlled variable is monitored, and the basic learning rate is first corrected according to the trend to generate a trend-corrected learning rate; the trend-corrected learning rate is secondly corrected according to the magnitude of the real-time disturbance estimate to generate a final adaptive learning rate.

[0014] Optionally, in a seventh implementation of the first aspect of the present invention, the basic controller output signal, the feedforward compensation signal, and the adaptive compensation signal are received; the basic controller output signal, the feedforward compensation signal, and the adaptive compensation signal are superimposed to obtain a composite control signal; the composite control signal is amplitude-limited according to the safe operating range of the actuator to generate a safety control signal; and the safety control signal is output to the actuator as a final control command.

[0015] Beneficial effects: The introduction of a quasi-continuous disturbance observer breaks away from the limitations of traditional control that relies solely on single feedback. It not only comprehensively considers the measured values ​​of the controlled variable but also deeply integrates the output signal of the basic controller, enabling it to generate an estimate of the total system disturbance in a comprehensive, real-time, and accurate manner. This lays a solid foundation for subsequent precise control, allowing the system to respond to various internal and external disturbances more promptly and effectively. The output signals of the basic controller, feedforward compensation signals, and adaptive compensation signals are deeply integrated to form the final control command. The basic controller ensures the basic control performance of the system under normal operating conditions, providing a stable control benchmark; the feedforward compensation signal can quickly respond to disturbances, compensating in the early stages of disturbance occurrence to reduce their impact on the system; the adaptive compensation signal automatically adjusts the control parameters according to the dynamic changes of the system, keeping the system in an optimal control state at all times. These three complement each other, achieving multi-level and comprehensive optimized control, greatly improving the system's control performance. By monitoring the operating load of thermal power units in real time, the system generates a current operating condition identifier and accurately retrieves the corresponding nominal model parameters, feedforward gain coefficients, and adaptive learning rate parameter sets from a preset parameter mapping table. More advancedly, during actual operation, the system can automatically verify and fine-tune the parameters in the parameter mapping table based on the current control effect. This dynamic adjustment mechanism ensures that the control parameters are always highly matched with the actual operating state of the system, giving the control method extremely strong adaptability and flexibility. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of an embodiment of the adaptive optimization control method for thermal power unit thermal processes in this invention. Detailed Implementation

[0017] This invention provides an adaptive optimization control method for the thermal process of thermal power units, which improves the control quality and adaptability of the thermal process of thermal power units. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or device that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0018] For ease of understanding, the specific process of the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 1 One embodiment of the adaptive optimization control method for thermal power unit thermal processes in this invention includes: 101. Calculate the deviation between the measured value and the set value of the controlled variable in the target control loop of the thermal process.

[0019] It is understood that the executing entity of this invention can be an adaptive optimization control device for the thermal process of a thermal power unit, or it can be a terminal or a server; no specific limitation is made here. This embodiment of the invention will be described using a server as an example.

[0020] Specifically, the controlled variable measurement signal of the target control loop is acquired in real time; the optimized setpoint of the controlled variable is determined according to the current operating load of the thermal power unit; the controlled variable measurement signal is filtered to obtain the filtered measurement value; and the real-time deviation between the filtered measurement value and the optimized setpoint is calculated.

[0021] It should be noted that, taking main steam temperature control as an example, the server reads the thermocouple signal installed on the steam pipe at the outlet of the boiler's high-temperature superheater in real time through the I / O module of the distributed control system (DCS). This signal is transmitted in the form of a 4-20mA current, representing temperature. Assume that at a certain sampling moment, the current signal read by the server, after range conversion, corresponds to an original main steam temperature measurement of 538.5℃.

[0022] The main steam temperature is not a constant value; its optimal setpoint varies with the unit load. The server stores a pre-optimized "load-setpoint" curve table based on thermal testing. The current actual unit load is 450MW. Looking up the table, at a load of 450MW, the corresponding optimized main steam temperature setpoint is 535.0℃. This setpoint aims to balance unit thermal efficiency with the safety margin of the metal pipe walls.

[0023] The original measurement signal contains high-frequency noise (such as electrical interference). The server uses a first-order inertial filtering algorithm for processing, with a filtering time constant of 3 seconds to smooth out random fluctuations without causing significant delay. Assume the filtered output value at the previous moment was 537.8℃. According to the filtering algorithm, the filtered measurement value at the current moment is calculated to be 538.1℃. This value reflects the actual temperature trend more smoothly and reliably than the original value.

[0024] Subtract the filtered measurement value from the optimized setpoint to obtain the current control deviation: Real-time deviation = Filtered measurement value - Optimized setpoint = 538.1℃ - 535.0℃ = +3.1℃.

[0025] This deviation of +3.1℃ indicates that the current main steam temperature is 3.1 degrees higher than the ideal setpoint, and the control system needs to take action (such as increasing the desuperheating water flow) to reduce the temperature.

[0026] 102. Using a quasi-continuous disturbance observer, a real-time estimate of the total system disturbance is generated based on the measured values ​​of the controlled variable and the output signal of the basic controller.

[0027] Specifically, a nominal dynamic model of the target control loop is constructed; the model output value is calculated based on the output signal of the basic controller and the nominal dynamic model; the output deviation between the measured value of the controlled variable and the model output value is calculated; the output deviation is processed by inertial filtering to generate a real-time estimate of the total system disturbance.

[0028] It should be noted that, based on the current unit load (450MW), the preset nominal model for this operating condition is retrieved from the online parameter library. For the main steam temperature control loop, a simplified first-order inertial model is commonly used to approximate the effect of changes in the opening of the desuperheating water regulating valve on the main steam temperature. Under the current "Operating Condition B" (430-470MW), the model parameters are: static gain K n =0.8℃ / %, time constant T n =120 seconds. This model represents the basic relationship between valve command and temperature response under ideal, undisturbed conditions.

[0029] Based on the deviation (+3.1℃) obtained in step 101, the basic PID controller calculates and outputs a control command, requesting the desuperheating water regulating valve to be opened to 62%. The server inputs this command (62%) into the aforementioned nominal model. Based on the state at the previous moment (assuming its output was 535.6℃) and the current input, the model calculates the theoretical model output value at the current moment to be 535.8℃.

[0030] Compare the actual process filter measurement value (538.1℃) obtained in step 101 with the model output value (535.8℃) calculated in the previous step: Output deviation = actual measurement value - model output value = 538.1℃ - 535.8℃ = +2.3℃.

[0031] This +2.3℃ difference is crucial, as it encompasses the sum of all unmodeled dynamics, external disturbances (such as combustion fluctuations and load changes), and model errors—a direct representation of the "total system disturbance."

[0032] To extract a smooth and usable disturbance estimate signal from the output deviation, filtering is required. The server employs an inertial filter loop of a quasi-continuous disturbance observer, with a time constant set to 20 seconds. This value strikes a balance between speed and smoothness. Assume the disturbance estimate at the previous moment was +2.0℃. After recursive calculation using the inertial filtering algorithm, the real-time estimate of the total system disturbance at the current moment is updated to +2.2℃.

[0033] Step 102 successfully estimated the total system disturbance of +2.2℃ in real time through the process of "model calculation of theoretical output -> comparison with actual output to obtain the original deviation -> inertial filtering to obtain smooth estimate".

[0034] 103. Perform feedforward calculations based on real-time disturbance estimates to generate feedforward compensation signals.

[0035] Specifically, the feedforward gain coefficient is determined based on the current operating status of the thermal power unit; the feedforward gain coefficient is multiplied by the real-time disturbance estimate to calculate the feedforward compensation amount; the amplitude of the feedforward compensation amount is limited to generate a feedforward compensation signal. Further, based on the correspondence between the current operating load of the thermal power unit and the preset load range, a first static feedforward gain is determined by looking up a table; a second dynamic feedforward gain is calculated based on the rate of change of the real-time disturbance estimate; the first static feedforward gain and the second dynamic feedforward gain are weighted and synthesized to generate the final feedforward gain coefficient.

[0036] It should be noted that the feedforward gain coefficient determines the strength of the compensation, which is composed of static and dynamic components.

[0037] Determine the first static feedforward gain: Based on the current unit load of 450MW, consult the "Load-Static Feedforward Gain" mapping table. Under "Operating Condition B", the static feedforward gain K... s = -0.40% / ℃. The negative sign indicates that the compensation direction is opposite to the disturbance direction.

[0038] Calculate the second dynamic feedforward gain: The server calculates the rate of change of the disturbance estimate. Assuming the disturbance estimate was +2.0℃ at the previous moment, and is currently +2.2℃, with a sampling interval of 1 second, the rate of change is +0.2℃ / second. The preset dynamic gain coefficient K... d -0.10% / (℃ / second). Dynamic feedforward gain component = rate of change × K d =0.2×(-0.10)=-0.02% / ℃.

[0039] Weighted final gain: static gain weight 0.7, dynamic gain weight 0.3. Final feedforward gain coefficient K. ff =(-0.40×0.7)+(-0.02×0.3)=-0.28-0.006=-0.286% / ℃.

[0040] The feedforward compensation is calculated by multiplying the feedforward gain coefficient by the real-time disturbance estimate: Feedforward compensation = Feedforward gain coefficient × Real-time disturbance estimate = (-0.286% / ℃) × (+2.2℃) = -0.629%. This negative value means that a command to reduce the valve opening by 0.629% needs to be generated on the feedforward channel to directly counteract the disturbance that causes the temperature to rise (+2.2℃).

[0041] To prevent excessive single-step compensation from causing system overload, the calculated feedforward compensation amount is limited. The preset single-step feedforward compensation limit is ±1.0%. The calculated -0.629% is within the limit range, therefore it passes. The final generated feedforward compensation signal is -0.629%.

[0042] Step 103 transforms the real-time disturbance estimate (+2.2℃) into a specific, directional control correction (-0.629%) through a feedforward gain (-0.286% / ℃) determined by both static lookup and dynamic calculation.

[0043] 104. Integral learning is performed by combining real-time disturbance estimates and deviations to generate adaptive compensation signals.

[0044] Specifically, the adaptive learning rate is determined based on the magnitude of the real-time disturbance estimate; an adaptive parameter is generated by integrating the adaptive learning rate with the deviation; the adaptive compensation amount is calculated by combining the adaptive parameter with the direction and amplitude information of the real-time disturbance estimate; and the adaptive compensation amount is limited to generate an adaptive compensation signal. Further, the basic learning rate is obtained from a preset load-rate mapping relationship based on the current operating load of the thermal power unit; the changing trend of the controlled variable deviation is monitored, and the basic learning rate is first corrected based on this trend to generate a trend-corrected learning rate; the trend-corrected learning rate is secondly corrected based on the magnitude of the real-time disturbance estimate to generate the final adaptive learning rate.

[0045] It should be noted that the adaptive learning rate is determined based on the magnitude of the real-time disturbance estimate, and the learning rate is dynamically adjusted according to the operating conditions and system state.

[0046] Obtain the basic learning rate: Based on the current load of 450MW (condition B), consult the parameter table to obtain the basic learning rate μ. base =0.05% / (℃·s).

[0047] First correction (based on deviation trend): Monitor the change in the deviation of the controlled variable (currently +3.1℃). Assume the deviation was +3.4℃ at the previous time step, showing a downward trend, indicating effective control. To prevent overshoot, the learning rate needs to be lowered, with a trend correction coefficient of 0.8. Corrected rate: 0.05 × 0.8 = 0.04% / (℃·s).

[0048] Second correction (based on disturbance amplitude): Analyze the real-time disturbance estimate (+2.2℃). Its large amplitude indicates a significant disturbance. To improve the compensation speed for persistent disturbances, the learning rate should be appropriately increased. The preset amplitude correction coefficient is 1.2 (this coefficient > 1 when the disturbance amplitude is greater than the threshold). The final adaptive learning rate μ = 0.04 × 1.2 = 0.048% / (℃·s).

[0049] The integral operation accumulates the "bias × learning rate". Assume the accumulated adaptive parameter A from the previous time step... prev The value is 0.9%. The current sampling interval is 1 second, and the integral increment = deviation × μ = 3.1℃ × 0.048% / (℃·s) ≈ 0.149%. The updated adaptive parameter A... new =0.9% + 0.149% = 1.049%. This parameter reflects the strength of the compensating trend required to eliminate historical cumulative bias.

[0050] The adaptive compensation should be applied in the opposite direction to the main disturbance. The current disturbance estimate is +2.2℃, therefore the compensation direction is negative. The compensation amount is calculated as the product of the adaptive parameter and the disturbance estimate amplitude, with a negative sign: Adaptive compensation amount = -(A new ×|Disturbance estimate|)=-(1.049%×2.2)≈-2.308%.

[0051] To prevent integral saturation, the compensation amount is limited by ±3.0%. The calculated -2.308% is within the allowable range. The final adaptive compensation signal is then generated with a value of -2.308%.

[0052] Step 104 generates an adaptive compensation signal of -2.308% through an integral learner that is dynamically adjusted by the deviation trend and the disturbance amplitude.

[0053] 105. The output signal of the basic controller, the feedforward compensation signal and the adaptive compensation signal are integrated to form the final control command and output to the actuator.

[0054] Specifically, the system receives the output signal of the basic controller, the feedforward compensation signal, and the adaptive compensation signal; it superimposes the output signal of the basic controller, the feedforward compensation signal, and the adaptive compensation signal to obtain a composite control signal; it performs amplitude limiting processing on the composite control signal according to the safe operating range of the actuator to generate a safety control signal; and it outputs the safety control signal as the final control command to the actuator.

[0055] It should be noted that the server receives signals from three channels: The basic controller output signal is generated by the PID controller based on the real-time deviation (+3.1℃). Its output command is to increase the desuperheating water flow rate, corresponding to a valve opening of 62.0%.

[0056] Feedforward compensation signal: from step 103, with a value of -0.629%, meaning an immediate reduction of the opening by 0.629% to counteract the current disturbance.

[0057] Adaptive compensation signal: from step 104, with a value of -2.308%, meaning that based on historical bias learning, the opening needs to be continuously reduced by 2.308%.

[0058] The three signals are superimposed to obtain the composite control signal: Composite control signal = 62.0% + (-0.629%) + (-2.308%) = 59.063%.

[0059] The composite signal is amplitude-limited based on the physical limitations of the actuator (desuperheating water regulating valve) to ensure the safety and feasibility of the command. The safe operating range of this valve is 10% to 90% opening. Inspection showed that 59.063% was within this safe range; therefore, the safe control signal value is 59.063%.

[0060] This safety control signal (59.063% opening command) is used as the final control command and sent to the actuator of the desuperheating water regulating valve through the control network. The valve will then operate according to this command, adjusting the desuperheating water flow rate to control the main steam temperature.

[0061] Step 105 is the "culmination" and "guardian" of control commands. It integrates the three control forces targeting deviations, transient disturbances, and long-term disturbances into a unified command (59.063%). 106. Monitor the operating load of the thermal power unit and generate a current operating condition identifier; based on the current operating condition identifier, retrieve the corresponding nominal model parameters, feedforward gain coefficients, and adaptive learning rate parameter sets from the preset parameter mapping table; update the parameter sets to the quasi-continuous disturbance observer, the feedforward calculation process, and the integral learning process, respectively.

[0062] Furthermore, the preset parameter mapping table is constructed and maintained in the following way: step tests are conducted at multiple typical steady-state operating points of the thermal power unit to obtain the dynamic characteristic data of the object corresponding to each operating point; based on the dynamic characteristic data of the object, the nominal model parameters, feedforward gain coefficient, and adaptive learning rate parameters corresponding to each typical operating point are calculated and determined to form an initial parameter mapping table; in actual operation, when the system is in a quasi-steady state, the parameters corresponding to the current operating point in the parameter mapping table are verified and fine-tuned based on the current control effect to generate verified parameters; the verified parameters are used to update the corresponding records in the parameter mapping table.

[0063] It should be noted that the server continuously monitors the unit load. Assume the load steadily increases from 450MW to 500MW and remains stable. The server compares this current load with a preset "load range - operating condition identifier" mapping table. This table defines a load in the 480-520MW range as "Operating Condition C". The server then generates the current operating condition identifier "Operating Condition C".

[0064] Based on the "Operating Condition C" identifier, the server retrieves a complete set of optimized parameters for that operating condition from a pre-defined parameter mapping table. Parameter representation examples and annotations are shown in Table 1 below: Table 1 The server immediately updates the newly retrieved parameter set to each algorithm module online: Update the nominal model time constant to 110 seconds to the quasi-continuous perturbation observer in step 102.

[0065] Update the static feedforward gain to -0.35% / ℃ in the feedforward calculation module of step 103 as the reference value for the static gain.

[0066] Update the basic learning rate = 0.045% / (℃·s) to the integral learning module in step 104.

[0067] This parameter update operation is performed immediately at the start of each control cycle or when the operating condition switching conditions are met. Subsequently, the calculations in steps 101 to 105 within this control cycle will all be based on this updated set of latest parameters. For example, step 103 will recalculate the feedforward compensation using a static feedforward gain of -0.35% / ℃.

[0068] After the parameters are updated, once the system operates stably under "Condition C" (e.g., minimal load and temperature fluctuations for 10 consecutive minutes), the server initiates parameter verification. It evaluates the control performance indicators (e.g., overshoot, settling time) under the new parameters. If performance optimization is found (e.g., a slightly slower response to a disturbance), the system can automatically fine-tune the parameters within preset small steps (e.g., adjusting the static feedforward gain from -0.35 to -0.346), and update this "verified parameter" back to the parameter mapping table, achieving online self-learning and continuous optimization of the parameters.

[0069] Step 106 enables the control system to be "tailor-made." Through real-time operating condition identification and online parameter switching, it ensures that advanced algorithms such as disturbance observation, feedforward compensation, and adaptive learning always operate under the set of parameters most suitable for the current operating state. Its built-in parameter fine-tuning mechanism further guarantees the long-term optimal control performance, which is key to achieving efficient, stable, and adaptive optimized operation of thermal power units across the entire load range.

[0070] The present invention also provides an adaptive optimization control device for thermal processes of thermal power units. The adaptive optimization control device for thermal processes of thermal power units includes a memory and a processor. The memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the processor performs the steps of the adaptive optimization control method for thermal processes of thermal power units in the above embodiments.

[0071] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when the instructions are executed on a computer, cause the computer to perform the steps of the adaptive optimization control method for the thermal process of the thermal power unit.

[0072] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0073] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0074] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An adaptive optimization control method for the thermal process of a thermal power unit, characterized in that, include: Calculate the deviation between the measured value and the set value of the controlled variable in the target control loop of the thermal process; A real-time estimate of the total system disturbance is generated using a quasi-continuous disturbance observer based on the measured values ​​of the controlled variable and the output signal of the basic controller. Based on the real-time disturbance estimate, feedforward calculation is performed to generate a feedforward compensation signal; By combining the real-time disturbance estimate with the deviation, an adaptive compensation signal is generated through integral learning. The final control command is formed by integrating the output signal of the basic controller, the feedforward compensation signal, and the adaptive compensation signal.

2. The adaptive optimization control method for thermal processes of thermal power units according to claim 1, characterized in that, The controlled variable measurement signal of the target control loop is acquired in real time; the optimized setpoint of the controlled variable is determined according to the current operating load of the thermal power unit; the controlled variable measurement signal is filtered to obtain the filtered measurement value; Calculate the real-time deviation between the filtered measured value and the optimized set value.

3. The adaptive optimization control method for thermal power unit thermal processes according to claim 2, characterized in that, Construct a nominal dynamic model of the target control loop; calculate the model output value based on the output signal of the basic controller and the nominal dynamic model; Calculate the output deviation between the measured value of the controlled variable and the output value of the model; perform inertial filtering on the output deviation to generate a real-time estimate of the total system disturbance.

4. The adaptive optimization control method for thermal processes of thermal power units according to claim 3, characterized in that, The feedforward gain coefficient is determined based on the current operating status of the thermal power unit; the feedforward gain coefficient is multiplied by the estimated real-time disturbance value to calculate the feedforward compensation amount; the amplitude of the feedforward compensation amount is limited to generate a feedforward compensation signal.

5. The adaptive optimization control method for thermal processes of thermal power units according to claim 4, characterized in that, Based on the correspondence between the current operating load of the thermal power unit and the preset load range, a first static feedforward gain is determined by looking up a table; based on the rate of change of the real-time disturbance estimate, a second dynamic feedforward gain is calculated; the first static feedforward gain and the second dynamic feedforward gain are weighted and synthesized to generate the final feedforward gain coefficient.

6. The adaptive optimization control method for thermal processes of thermal power units according to claim 1, characterized in that, The adaptive learning rate is determined based on the magnitude of the real-time disturbance estimate; an adaptive parameter is generated by integrating the adaptive learning rate with the deviation; the adaptive parameter is combined with the direction and amplitude information of the real-time disturbance estimate to calculate the adaptive compensation amount. The adaptive compensation amount is subjected to amplitude limiting processing to generate an adaptive compensation signal.

7. The adaptive optimization control method for thermal processes of thermal power units according to claim 6, characterized in that, The basic learning rate is obtained from the preset load-rate mapping relationship based on the current operating load of the thermal power unit. Monitor the changing trend of the deviation of the controlled variable, and make a first correction to the basic learning rate based on the trend to generate a trend-corrected learning rate; The trend-corrected learning rate is then further adjusted based on the magnitude of the real-time perturbation estimate to generate the final adaptive learning rate.

8. The adaptive optimization control method for thermal processes of thermal power units according to claim 1, characterized in that, The system receives the output signal of the basic controller, the feedforward compensation signal, and the adaptive compensation signal; it superimposes the output signal of the basic controller, the feedforward compensation signal, and the adaptive compensation signal to obtain a composite control signal; it performs amplitude limiting processing on the composite control signal according to the safe operating range of the actuator to generate a safety control signal; and it outputs the safety control signal as the final control command to the actuator.

Images