Multivariable dynamic collaborative optimization control method for deacidification of waste incineration flue gas

Abstract

The embodiment of the invention discloses a multivariable dynamic collaborative optimization control method for waste incineration flue gas deacidification, and relates to the technical field of waste incineration flue gas deacidification intelligent control. The method comprises the steps that real-time data are acquired and preprocessed, and a temperature-pressure coupling coefficient KT-P is calculated; calculating a rule weight Wi by using the preprocessed real-time data and the temperature-pressure coupling coefficient KT-P; the rule weight Wi is utilized, after fuzzy rule matching is conducted, a preliminary control quantity delta Qslurry is output, and delta Qslurry is the slurry flow variable quantity; performing prediction-energy efficiency optimization, solving an optimal control quantity through a target function and constraint conditions and quadratic programming, and outputting an optimization instruction to an execution mechanism; and updating the rule base according to the current system state and a preset reward function. According to the method, the problems of low control precision and high operation cost in the existing desulfurization and deacidification process of the waste incineration flue gas can be solved, and the environmental benefits and economic benefits of a waste incineration plant are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent control of waste incineration flue gas deacidification, and in particular to a multivariable dynamic collaborative optimization control method for waste incineration flue gas deacidification. Background Art

[0002] With the continuous development of society and increasingly stringent environmental protection requirements, waste incineration has become a key means of treating municipal solid waste. However, the flue gas generated during the waste incineration process contains a large amount of acidic and harmful gases such as HCl and SO2. If not effectively purified and treated, it will cause serious environmental pollution.

[0003] The current "Standard for Control of Pollutants from the Incineration of Municipal Waste" (GB18485-2014) is no longer able to meet current environmental protection needs. Zhejiang, Hebei and other places have launched pilot projects to raise standards and issued relevant plans, which have put forward more stringent requirements on the emission indicators and purification facilities of waste incineration plants.

[0004] In the flue gas purification system of waste incineration power plants, the SDA (Spray Dryer Absorber, rotary spray drying) flue gas deacidification process is the core link. However, the process control process has the characteristics of large time lag, nonlinearity and multivariable. At present, most municipal waste incineration power plants use traditional PID (proportional-integral-differential) control algorithms and conventional fuzzy control + predictive control methods, which make it difficult to accurately control the amount of slaked lime slurry added when the concentration of flue gas pollutants fluctuates greatly. This not only leads to a significant increase in slaked lime consumption and electricity consumption, but may also cause emissions to exceed standards, seriously affecting environmental and economic benefits. Existing methods often lack effective control and precise calculation of these factors.

[0005] At present, most municipal solid waste incineration power plants use traditional PID control algorithms, and the flue gas deacidification control program is designed to be executed in the flue gas purification system PLC (programmable logic controller) or DCS (distributed control system). The specific process is to adjust the opening of the lime slurry solution valve through a traditional PID controller based on the real-time detection of the content of acidic components (such as HCL and SO2) in the exhaust gas. When it is detected that the content of acidic components is too high, the injection amount of lime slurry solution is increased through the controller to enhance the acid-base neutralization effect; on the contrary, when the content of acidic components is reduced to a lower level, the amount of lime slurry solution is reduced through the controller. Through this control method, it is ensured that the content of acidic components in the exhaust gas can be maintained within the standard range specified by the environmental protection department.

[0006] Traditional PID control has limitations. Although it is widely used in industrial automation, its linear control characteristics are difficult to adapt to the nonlinear and time-varying systems used in waste incineration flue gas deacidification. The specific defects are as follows:

[0007] Difficulty in parameter tuning: The concentrations of HCl and SO₂ in waste incineration flue gas are affected by waste composition (such as chlorinated plastics and sulfides) and combustion efficiency, exhibiting significant fluctuations (instantaneous variations of up to ±100%). Traditional PID controllers require manual tuning of parameters (Kp, Ki, Kd) based on experience, but fixed parameters cannot guarantee effective control when operating conditions change frequently.

[0008] Overshoot and oscillation risks: The flue gas deacidification system has a large time lag characteristic (it takes 20-50 seconds from slurry adjustment to concentration feedback), and the PID integral term is prone to accumulate errors, causing overshoot or continuous oscillation.

[0009] Multivariable coupling issues: Traditional PID control relies primarily on single-loop and cascade regulation, which cannot address the strong coupling effects of temperature, pressure, and pollutant concentration. For example, increased flue gas temperature reduces the slurry atomization efficiency and deacidification efficiency in the deacidification tower, resulting in a decrease in acid gas removal rate and increased deacidification material consumption.

[0010] The application of fuzzy control and predictive control combined with feedforward control also has limitations. Fuzzy control solves nonlinear problems by simulating human experience, but it still faces the following bottlenecks in waste incineration scenarios:

[0011] Rule base relies on manual experience: Existing fuzzy rule bases are mostly designed based on limited working condition data. For example, only considering moderate fluctuations in HCl concentration (±10mg / m 3 ), but does not cover extreme working conditions (such as garbage mixed with a large amount of chlorine-containing waste causing a sudden increase in HCl concentration of 50 mg / m 3 For example, a waste incineration plant in Zhejiang Province had a fuzzy rule base containing 50 rules. However, when processing waste containing over 30% chlorine-containing plastics, the lack of corresponding rules resulted in HCl emissions exceeding the standard three times that day.

[0012] Long-term performance degradation: The fuzzy rule base lacks a self-learning mechanism and cannot adapt to changes such as equipment aging and sensor drift after long-term operation. For example, when the efficiency of a slurry pump decreases, the flow rate instructions output by the original rules may not achieve the expected effect.

[0013] Rule conflicts and redundancy: Under complex operating conditions, multiple rules may trigger conflicting instructions. For example, the simultaneous existence of rules requiring "increase slurry flow at high HCl concentrations" and "reduce slurry flow at high temperatures" can cause frequent system fluctuations. Experimental verification shows that in a scenario simulating simultaneous increases in flue gas temperature and HCl concentration, the slurry flow rate under traditional fuzzy control fluctuates by as much as ±20%, a threefold increase compared to single-variable conditions.

[0014] Difficulties in implementing predictive control. Predictive control, such as model predictive control (MPC), relies on high-precision models and data, but faces the following challenges in waste incineration scenarios:

[0015] Model inaccuracy problem: The components of waste incineration flue gas are complex, and traditional ARIMA (Autoregressive Integrated Moving Average Model) models and state-space models are difficult to accurately describe the dynamic characteristics of HCl / SO2. For example, the concentration of HCl is affected by the reaction rate of Cl- and metal compounds. Its kinetic equation contains multiple nonlinear terms, and existing models are often simplified to linear relationships, resulting in prediction errors. For example, when the calorific value of garbage changes by 20%, the predicted deviation of HCl concentration reaches 8mg / m 3 , triggering misadjustment.

[0016] Data quality bottleneck: Smoke detection environments are harsh (high temperature, high dust levels), and sensor signals are susceptible to noise interference. For example, in an 800°C environment, thermal radiation noise can cause the signal-to-interference plus noise ratio (SNR) of a laser spectrometer to drop below 10dB, resulting in high-frequency jitter in the measured data.

[0017] Computing resource limitations: Real-time prediction requires solving optimization problems online (such as quadratic programming), which places high demands on the controller's computing power. For example, a single iteration of the MPC algorithm requires 50ms, while the waste incineration control cycle requires ≤100ms, resulting in controller overload.

[0018] In general, with increasingly stringent environmental protection requirements and the rapid development of automation technology, these control methods have gradually revealed their limitations, such as slow response speed, limited accuracy, and high resource consumption and costs. Therefore, more advanced and efficient automated control methods are needed to optimize the flue gas deacidification process, improve overall operational efficiency and environmental performance, and compensate for the many shortcomings of previous control methods in the flue gas deacidification process of waste incineration power plants. Summary of the Invention

[0019] In view of this, an embodiment of the present invention provides a multivariable dynamic collaborative optimization control method for waste incineration flue gas deacidification to solve the problems of low control accuracy and high operating costs in the existing waste incineration flue gas desulfurization and deacidification process, and to improve the environmental and economic benefits of waste incineration plants.

[0020] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0021] A multivariable dynamic collaborative optimization control method for waste incineration flue gas deacidification, comprising:

[0022] Step S101: Acquire real-time data for preprocessing and calculate the temperature-pressure coupling coefficient K T-P ;

[0023] Step S102: Using the pre-processed real-time data and the temperature-pressure coupling coefficient K T-P , calculate the rule weight W i ;

[0024] Step S103: Using the rule weight W i After fuzzy rule matching, the output of the preliminary control quantity ΔQ slurry , where ΔQ slurry is the change in slurry flow rate;

[0025] Step S104: perform prediction-energy efficiency optimization, solve the optimal control quantity through the objective function and constraint conditions, and output the optimization instruction to the execution mechanism;

[0026] Step S105: Update the rule base according to the current system state and the pre-set reward function.

[0027] The present invention has the following beneficial effects:

[0028] Improved control accuracy: Comprehensive application of multi-variable dynamic collaborative optimization control achieves high-precision control of the flue gas desulfurization and deacidification process, and the system response time is improved by 28.1%;

[0029] Energy saving and consumption reduction: Through multi-variable coordinated control and precise parameter adjustment, the use of slaked lime slurry, process water and energy is optimized, reducing operating costs and energy consumption. The operating cost of slaked lime is reduced by 7.7%;

[0030] Enhanced system stability: Advanced control algorithms and real-time monitoring mechanisms improve the system's adaptability to changes in operating conditions and interference, enhance system stability and reliability, and increase the number of control system overshoots by 16.7%. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0032] Figure 1 This is a schematic flow chart of the multivariable dynamic collaborative optimization control method for waste incineration flue gas deacidification according to the present invention;

[0033] Figure 2 This is a schematic diagram of the multivariable dynamic collaborative optimization control method for waste incineration flue gas deacidification according to the present invention;

[0034] Figure 3 This is a schematic diagram of the dynamic weight control principle in the present invention;

[0035] Figure 4 This is a schematic diagram of the prediction-energy efficiency integrated optimization principle in the present invention;

[0036] Figure 5 This is a schematic diagram of the reinforcement learning rule base in the present invention. DETAILED DESCRIPTION

[0037] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0038] It should be understood that the embodiments described are only a portion of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without creative work are within the scope of protection of the present invention.

[0039] The core flaws of existing technologies can be summarized as follows: linear control cannot handle nonlinear systems, fuzzy rules lack dynamic optimization capabilities, and predictive models are limited by data quality and computing power. These issues directly lead to the triple dilemma of large emission fluctuations, high operating costs, and difficult retrofitting for waste incineration plants.

[0040] The present invention proposes a systematic solution to the aforementioned pain points, addressing the problems of low control accuracy and high operating costs caused by high volatility (instantaneous changes in HCl / SO2 concentrations of ±100%) and strong coupling (interactions between temperature, pressure, and pollutant concentrations) in the deacidification process of waste incineration flue gas. The present invention proposes a multivariable dynamic collaborative optimization control method, aiming to provide a high-precision, adaptive, and low-cost intelligent control method for desulfurization and deacidification of waste incineration flue gas, achieving all-round, high-precision control of the flue gas deacidification process, ensuring stable and economical HCl and SO2 emissions, reducing operating costs, and minimizing environmental impact.

[0041] The embodiment of the present invention provides a multivariable dynamic collaborative optimization control method for waste incineration flue gas deacidification, such as Figure 1-2 As shown, including:

[0042] Step S101: Acquire (flue gas) real-time data for preprocessing and calculate the temperature-pressure coupling coefficient K T-P ;

[0043] As an optional embodiment, in step S101, the real-time data includes HCl concentration, SO2 concentration, temperature and pressure. Pre-processing can be various conventional processing methods in the art such as filtering and noise reduction, which will not be described here.

[0044] like Figure 3 As shown, in specific implementation, the input variables may also include:

[0045] e HCl: HCl concentration deviation (set value - measured value);

[0046] Δe HCl : HCl concentration deviation change rate.

[0047] As another optional embodiment, the temperature-pressure coupling coefficient K T-P The calculation formula is:

[0048]

[0049] (Temperature range: T min =120℃, T max =180℃; Pressure range: P min =-500Pa, P max =0Pa)

[0050] Parameter definition:

[0051] T: real-time temperature value (unit: °C);

[0052] T min : lower temperature limit (120℃); T max : Temperature upper limit (180℃);

[0053] P: real-time pressure value (unit: Pa);

[0054] P min : Pressure lower limit (-500Pa); P max : Upper pressure limit (0Pa).

[0055] Temperature item Normalize the temperature T to the interval [0,1] to reflect the relative position of T within its extreme value range. min When T=T max When , the temperature item value is 1.

[0056] Pressure Item Similarly, the pressure P is normalized to [0,1].

[0057] Weight distribution: The weights of T and P are 0.3 and 0.7, respectively, indicating that in the waste incineration process, the fluctuation of pressure P has a greater impact on the deacidification efficiency.

[0058] Comprehensive results: K T-P The value range is [0,1]. The larger the value, the better the comprehensive performance of T and P.

[0059] It is worth noting that if T or P exceeds the extreme value range, additional processing is required (such as truncation or expansion of the extreme value).

[0060] Step S102: Using the pre-processed real-time data and the temperature-pressure coupling coefficient K T-P , calculate the rule weight W i ;

[0061] This step involves the dynamic weight cooperative control algorithm.

[0062] As an optional embodiment, a weight adaptive mechanism is adopted (i.e., the rule weight is dynamically adjusted according to the real-time working conditions). In step S102, the rule weight W i The calculation formula is:

[0063] W i =α·(C SO2 / C HCl )+(1-α)·K T-P

[0064] Among them, C SO2 / C HCl is the SO2 / HCl concentration ratio, and the weight coefficient covers extreme operating conditions;

[0065] α is the adaptive coefficient, and the adaptive mechanism is adopted: initial α=0.6, according to C HCl / C SO2 The concentration ratio is dynamically adjusted. When the typical ratio is 1:2 to 1:5, α is automatically adapted to 0.5 to 0.7 to ensure that the weight is increased under high SO2 conditions.

[0066] Example rule (W for high HCl concentration and low temperature conditions) i calculate):

[0067] Real-time data of working conditions: HCl = 50mg / Nm 3 (Set value 10mg / Nm 3 , deviation e HCl =-40mg / Nm 3 , which belongs to the "negative big NB").

[0068] SO2=100mg / Nm 3 , concentration ratio C SO2 / C HCl =100 / 50=2 (equal to the lower limit, then α=0.5).

[0069] Temperature T=130℃(T min =120°C, normalized value (130-120) / (180-120)=0.167).

[0070] Pressure P = -300Pa (P min =-500Pa, normalized value (-300+500) / 500=0.4).

[0071] Core contradiction: HCl concentration far exceeds the set value, and a large amount of slurry needs to be added. However, the deacidification efficiency is good at this temperature, and the slurry increase can be moderately reduced. i dynamic equilibrium.

[0072] Temperature-pressure coupling coefficient K T-P =0.3×0.167+0.7×0.4=0.33;

[0073] Dynamic weight W i =α·(C SO2 / C HCl )+(1-α)·K T-P =0.5×2+0.5×0.33=1.165;

[0074] Compared with traditional fuzzy control: when there is no weight mechanism, the conflict between the rules of "increasing slurry when pollutants exceed the standard" and "reducing slurry when the temperature is appropriate" may lead to insufficient slurry increase, thereby prolonging the time of exceeding the standard.

[0075] Advantages of the present invention: Through the calculation of the dynamic weight Wi, the "positive" slurry increase can be forced to be triggered. Combined with the correction of the prediction module, the system response speed can be effectively improved, the problem of fixed rule weights in traditional fuzzy control can be solved, and the system control delay or system oscillation caused by the simultaneous triggering of the "slurry increase" and "slurry reduction" rules can be avoided.

[0076] Step S103: Using the rule weight W i After fuzzy rule matching, the output of the preliminary control quantity ΔQ slurry , where ΔQ slurry is the change in (slaked lime) slurry flow rate;

[0077] Step S104: perform prediction-energy efficiency optimization, solve the optimal control quantity through the objective function and constraint conditions, and output the optimization instruction to the execution mechanism;

[0078] This step involves the prediction-energy efficiency integrated optimization algorithm.

[0079] like Figure 4 As shown, as an optional embodiment, step S104 includes:

[0080] Step S1041: Based on real-time data and historical data, a prediction module is used to output predicted values of HCl and SO2 concentrations within a preset time period in the future;

[0081] In this step, a prediction module can be used to output a predicted value of HCl / SO2 concentration within a preset future time period (e.g., the next 5 minutes) based on the sampling of real-time data and historical data. The construction of the prediction module is conventional in the art and will not be described in detail here. The real-time data and historical data can both include HCl concentration, SO2 concentration, temperature, and pressure.

[0082] Step S1042: constructing an objective function;

[0083] Preferably, in step S1042, the objective function is:

[0084] J=Σe 2 +λ·ΣQ slurry

[0085] Here, e represents the deviation of the pollutant (HCl / SO2) concentration, that is, the difference between the actual measured pollutant concentration and the set pollutant concentration.

[0086] The significance of the square summation: The deviation e is squared and then the squared deviation values are added up over all time periods. This is done to emphasize the impact of larger deviations on the objective function, as larger deviations have larger squared values and therefore contribute more weight to the sum. The squaring operation also ensures that all deviation terms are non-negative, preventing positive and negative deviations from canceling each other out, thus providing a more accurate measure of the overall deviation.

[0087] λ is the energy efficiency weight, with a default value of 0.3. It regulates the relative importance between emission compliance and cost control and can be dynamically adjusted (0.1 to 0.5) based on emission requirements. The impact of λ: A smaller λ value indicates a greater emphasis on emission compliance, meaning pollutant concentrations are kept as close to the set values as possible, with minimal deviation. A larger λ value emphasizes cost control, meaning the use of slaked lime slurry is minimized.

[0088] Q slurry Represents the flow rate of slaked lime slurry. The size of the slurry flow rate will directly affect the treatment cost and treatment effect. slurry The accumulated slaked lime slurry flow rate at each moment reflects the total usage of slaked lime slurry during the entire control period, which is an important indicator of treatment cost.

[0089] Step S1043: Quadratic programming to solve the optimal ΔQ slurry ;

[0090] In this step, the optimal △Q is solved by quadratic programming slurry (Slurry flow rate change, that is, the slurry flow rate increment that needs to be adjusted).

[0091] Preferably, in step S1043, the quadratic programming formula is:

[0092]

[0093] Among them, e HCL,trepresents the deviation of HCl concentration at time t, e SO2,t represents the deviation of SO2 concentration at time t, λ is the energy efficiency weight, Q slurry,t represents the slurry flow rate at time t, ΔQ slurry is the change in the flow rate, that is, (Q slurry,t =Q slurry,t-1 +ΔQ slurry ), N is the number of time steps in the prediction domain or optimization domain, that is, the number of discrete time points in the future preset period.

[0094] Step S1044: Output optimization instructions under pre-set constraints.

[0095] In this step, an optimization instruction is output under the constraint conditions to adjust the slurry regulating valve.

[0096] Preferably, in step S1044, the preset constraint conditions are:

[0097] Emission limit: HCl ≤ 10 mg / m 3 , SO2≤50mg / m 3 ;

[0098] Equipment limitation: Maximum flow rate Q of slurry pump max =500 l / min; the slurry flow rate change ΔQslurry,t≤2.8%·Qslurry,t-1(t≥2), that is, the change / rate of the slurry flow rate does not exceed 2.8% of the flow rate at the previous moment.

[0099] This allows for dual-objective optimization, clearly demonstrating the balance between emission deviation and slaked lime cost, and the dynamic adjustment mechanism of the λ parameter to highlight the coordinated optimization of economy and environmental protection.

[0100] Step S105: Update the rule base according to the current system state and the pre-set reward function.

[0101] This step involves the reinforcement learning rule base algorithm. Figure 5 As shown, as an optional embodiment, step S105 includes:

[0102] Step S1051: Initialize the fuzzy rule base and Q table;

[0103] In this step, after starting the process, the fuzzy rule base and Q-table are initialized. The fuzzy rule base is a set of rules based on expert experience and initial settings, and the Q-table is used to store the value estimate of each state-action pair.

[0104] Step S1052: quantify the current control state (such as emission deviation, energy consumption) into a state value S, and continuously obtain the current system state S;

[0105] Preferably, in step S1052, the formula for the current system state S is:

[0106]

[0107] The state S consists of multiple normalized variables, among which e HCl and e SO2 are the concentration deviations of HCl and SO2, Q slurry is the slurry flow rate, ΔQ slurry is the change in slurry flow rate, C max and Q max are the maximum values of the corresponding variables.

[0108] Step S1053: Determine whether the termination condition is met. If not, execute step S1054. If the termination condition is met, output the optimized fuzzy rule base and end the process.

[0109] In this step, it is determined whether the preset termination conditions are met, such as reaching the maximum number of iterations, system state convergence, etc. If the termination conditions are met, the optimized fuzzy rule base is output and the process ends; if not, the process continues.

[0110] Step S1054: Generate control action A according to the current rule base;

[0111] Step S1055: Execute the selected action A, the system state changes to the new state S', and the reward R is calculated based on the state change and system performance;

[0112] Preferably, in step S1055, the reward function formula is:

[0113]

[0114] Among them, C meas is the actual measured concentration, C lim is the emission limit; Cost actual Cost is the actual cost. ref is the reference cost; oscillation_penalty is the oscillation of the system.

[0115] The reward R takes into account the emission compliance situation (C meas is the actual measured concentration, C lim Emission limit), cost consumption (Cost actual Cost is the actual cost. ref is the reference cost) and the oscillation of the system (oscillation_penalty). This formula optimizes the system to operate safely (away from C lim), cost control (lower than Cost ref ) and stability (reducing oscillations).

[0116] Among them, performance safety

[0117] Purpose: To evaluate the measurement value C meas Relative to the limit value C lim safety margin.

[0118] Effect: When C meas Close to or above C lim When , the score is greatly reduced (even negative), encouraging people to stay away from the limit value to ensure safety.

[0119] Weight: 10 (highest), highlighting the priority of security or stability.

[0120] Cost-effectiveness items

[0121] Purpose: To measure the difference between actual cost and reference cost.

[0122] Effect: When the actual cost is lower than the reference value, the score will increase, otherwise the score will be deducted, which encourages cost savings.

[0123] Weight: 5, second only to performance and safety.

[0124] Stability penalty: -2·oscillation_penalty

[0125] Purpose: To suppress system oscillation or instability.

[0126] Effect: The more severe the oscillation, the more the total score will be deducted (such as the number of fluctuations, amplitude, etc.).

[0127] Weight: 2, emphasizing stability but with low priority.

[0128] Step S1056: Update Q table;

[0129] Preferably, the step S1056 includes:

[0130] Update the Q table according to the reward R and the new state S', and use the Q-learning algorithm to update the Q value. The specific formula is as follows:

[0131] Q(S,A)←Q(S,A)+η·[R+γ·maxQ(S',A′)-Q(S,A)]

[0132] Among them, maxQ(S',A') is the maximum expected reward of the next state, η is the learning rate, and γ is the discount factor.

[0133] In the specific implementation, the learning rate η = 0.1 and the discount factor γ = 0.9 are used to balance the immediate rewards and future rewards.

[0134] Step S1057: Update the fuzzy rule base according to the new Q value;

[0135] In this step, the fuzzy rule base is updated according to the updated Q value to optimize the weights or actions of the rules.

[0136] Step S1058: Update the current state to the new state S', go to step S1053, and continue the loop.

[0137] The interaction relationship between the above steps S101 to S105 is as follows:

[0138] (1) Dynamic weight → fuzzy control: providing rule weight matrix;

[0139] (2) Fuzzy control → predictive optimization: transfer preliminary control quantity;

[0140] (3) Prediction optimization → actuator: output final control instructions;

[0141] (4) Execution agency → reinforcement learning: feedback emission data and cost information;

[0142] (5) Reinforcement learning → fuzzy control: periodically update the rule base.

[0143] Figure 2 The closed-loop control process from data acquisition to control instruction output and then to rule base optimization is fully demonstrated, which is the core idea of "multi-variable dynamic collaborative optimization" described in this application.

[0144] In one embodiment of the present invention, to implement the above invention, the following changes can be made at the hardware level:

[0145] 1. Add a plug-and-play external control box:

[0146] Protocol communication module: supports industrial protocols such as OPC UA, Modbus, Profinet, Profibus-DP, and is compatible with common PLC and DCS control systems.

[0147] Virtual simulation unit: Built-in digital twin model, which can simulate the control effect offline without affecting the original system operation during debugging.

[0148] During specific implementation, according to the actual communication protocol, the plug-and-play external control box is connected to the original DCS or PLC system via industrial Ethernet or other communication methods.

[0149] The advantage of the plug-and-play external control box is that it can be seamlessly compatible with old control systems, reducing transformation costs and risks, and making the control system plug-and-play with a transformation cycle of less than 30 days.

[0150] 2. Add testing instruments:

[0151] High-precision in-situ gas analysis instrument: HCl / SO2 laser / UV spectrometer (measuring range HCl: 0-2000mg / Nm 3 , SO2: 0-1200mg / Nm 3 ).

[0152] During specific implementation, a laser / ultraviolet spectrometer is installed at the inlet of the deacidification tower.

[0153] 3. Add intelligent actuator:

[0154] Hydrated lime slurry variable frequency pump (flow rate regulation accuracy ±1%).

[0155] Hydrated lime slurry regulating valve and process water regulating valve (response time <0.5s).

[0156] During the specific implementation, the framework computing unit configuration is as follows: flash the SD card pre-installed with the lightweight LSTM model to Jetson Nano, and set the sampling period to 100ms.

[0157] The following debugging can be done at the software level:

[0158] Virtual simulation verification: Import historical operating data, simulate HCl / SO2 concentration mutation scenarios, and verify that the control response time is less than 2 seconds.

[0159] Dynamic weight adjustment: According to the actual HCl / SO2 ratio range of the waste incineration plant (1:2 to 1:5), the initial weight coefficient α is set to 0.6.

[0160] After running the system, the experimental data obtained are shown in Table 1:

[0161] Table 1 Experimental data

[0162] index Traditional PID Existing fuzzy control Solution of the present invention <![CDATA[Average deviation of HCl / SO2]]> ±8.5 ±5.2 ±4.3 Consumption of slaked lime (kg / ton of garbage) 12.8 11.7 10.8 Overshoot times (times / day) 3.2 1.8 1.5 Response time (seconds) 45 32 23

[0163] Note: The data comes from actual measurements at a waste incineration plant in Sichuan (processing capacity 1,000 tons / day) from March to May 2023. The slaked lime solution of the present invention reduces operating costs by 7.7% (the fly ash treatment cost will also be reduced, but this was not calculated in this case).

[0164] As can be seen from Table 1, the average HCl / SO2 deviation of the present invention is smaller than that of the traditional PID and the existing fuzzy control, and the adjustment accuracy is high; the slaked lime unit consumption of the present invention is smaller than that of the traditional PID and the existing fuzzy control, and the use cost is low; the number of overshoots of the present invention is smaller than that of the traditional PID and the existing fuzzy control, and the use stability is strong; the response time of the present invention is shorter than that of the traditional PID and the existing fuzzy control, and the sensitivity is high.

[0165] As can be seen from the above, the present invention specifically solves the following problems:

[0166] Achieve dynamic coordinated control of multiple variables (HCl, SO2, temperature, pressure) to improve stability under complex working conditions;

[0167] Optimize the fuzzy rule base through self-learning mechanism to reduce manual intervention;

[0168] Solve the control accuracy problem under strong coupling of multiple variables and high fluctuation of working conditions;

[0169] Integrate energy efficiency optimization goals to reduce slaked lime usage and energy consumption;

[0170] Seamless compatibility with old control systems reduces transformation costs and risks, enables plug-and-play control systems, and a transformation cycle of less than 30 days.

[0171] In summary, the present invention has the following beneficial effects:

[0172] Improved control accuracy: Comprehensive application of multi-variable dynamic collaborative optimization control achieves high-precision control of the flue gas desulfurization and deacidification process, and the system response time is improved by 28.1%;

[0173] Energy saving and consumption reduction: Through multi-variable coordinated control and precise parameter adjustment, the use of slaked lime slurry, process water and energy is optimized, reducing operating costs and energy consumption. The operating cost of slaked lime is reduced by 7.7%;

[0174] Enhanced system stability: Advanced control algorithms and real-time monitoring mechanisms improve the system's adaptability to operating condition changes and disturbances, enhancing system stability and reliability. The control system's overshoot rate has increased by 16.7%.

[0175] Easy to maintain and expand: The independent external control box design facilitates system installation, maintenance and function expansion. It can be seamlessly connected to the original control system. There is no need to stop production during debugging, and it does not affect the use of the original control system, which is conducive to the promotion and application of technology.

[0176] The core innovation of this invention lies in: dynamic weight collaborative control algorithm (multivariable dynamic collaborative control algorithm) + prediction-energy efficiency integrated optimization algorithm (energy efficiency target integration method) + reinforcement learning rule base algorithm (self-learning rule optimization), which is particularly suitable for the high volatility and strong coupling of waste incineration flue gas deacidification environment. The specific description is as follows:

[0177] Multivariable dynamic weighting mechanism: For the first time, the pressure coupling coefficient and pollutant ratio are combined as weight factors into the control method, solving the problem of traditional fuzzy control ignoring the influence of working conditions (for example, when the temperature fluctuates by ±30°C, the weight adjustment range of the rules of the present invention reaches 35%, while the existing technology does not have such a mechanism).

[0178] Energy efficiency and emissions dual-objective optimization: By introducing the λ parameter (dynamically adjusted from 0.1 to 0.5), the cost of slaked lime was reduced by 11.3% (verified by data from a waste incineration plant in Sichuan) while ensuring that emissions meet standards. This integrated energy efficiency optimization goal reduces slaked lime usage and energy consumption.

[0179] Reinforcement Learning Rule Base: The reinforcement learning mechanism enables the rule base to be updated every 15 minutes, and after long-term operation, the control accuracy is improved by 17.3% (compared to a fixed rule base without a learning mechanism).

[0180] The core innovation and technical effects of this invention are as follows:

[0181] (1) Field innovation of dynamic weighted collaborative control algorithm

[0182] 1. Targeted design of multivariable coupling

[0183] In the prior art, the application of dynamic weight mechanism in other fields (such as chemical process control) does not take into account the specific coupling relationship of waste incineration flue gas deacidification (strong nonlinear interaction of temperature, pressure and pollutant concentration). This invention is the first to use the temperature-pressure coupling coefficient K T-P Ratio of pollutant concentration to C SO2 / C HCl Combined as a weight factor (i.e. formula W i =α·(C SO2 / C HCl )+(1-α)·K T-P ), solving the following problems that traditional methods cannot handle:

[0184] 1) Adaptability to extreme working conditions: When the HCl concentration rises suddenly (e.g., the mixing of chlorine-containing plastics causes an instantaneous change of ±100%), the "slurry increase" rule is forcibly triggered through dynamic weighting, avoiding the adjustment delay caused by rule conflicts in traditional fuzzy control (see the example in the manual, Wi calculation increases the slurry increase by 35%).

[0185] 2) Equipment characteristic matching: To address the problem of slurry pump efficiency decreasing with usage time, the weight factor reflects changes in operating conditions in real time and compensates for control deviations caused by equipment aging (traditional methods rely on fixed rules and cannot adapt).

[0186] 2. Deep binding with waste incineration process parameters

[0187] Temperature range (T min =120℃, T max =180℃), pressure range (P min =-500Pa, Pmax = 0Pa) and weighting (temperature weighting 0.3, pressure weighting 0.7) are calibrated based on actual deacidification tower operating data, rather than general parameters. For example, the impact of pressure on atomization efficiency is quantified as a 70% weight. This directly addresses the core pain point of the SDA process, "pressure fluctuations leading to uneven slurry atomization." This is a domain-specific optimization, not a simple transplant of the existing weighting mechanism.

[0188] (2) The synergistic efficiency mechanism of the three algorithms (different from simple combination)

[0189] 1. Hierarchical progression from dynamic weights to fuzzy control and prediction optimization

[0190] Dynamic weights empower fuzzy control: The traditional fuzzy control rule weights are fixed, which leads to rule conflicts under complex working conditions (such as "high temperature reduction" and "high pollution increase" triggered at the same time). The present invention uses the dynamic weight matrix W i Sort the priority of the rules so that the initial control quantity ΔQ output by the fuzzy controller is slurry More in line with real-time working conditions (see the manual for details Figure 3 , after weight adjustment, the rule matching accuracy increased by 22%).

[0191] Predictive optimization corrects fuzzy control defects: The preliminary control quantity output by fuzzy control is based only on the current state. The prediction module (such as the LSTM model) uses historical data to predict the concentration change in the next 5 minutes, and combines it with the energy efficiency objective function (J) for secondary planning, which solves the "short-sightedness" problem of fuzzy control (for example, avoiding excessive slurry increase to meet short-term standards, resulting in long-term cost increases).

[0192] Reinforcement learning closed-loop optimization rule base: Traditional fuzzy control rule bases rely on manual experience and cannot be updated. This invention uses the Q-learning algorithm to update the rule base every 15 minutes to adapt to long-term changes such as equipment aging and sensor drift (see experimental data for details, the control accuracy is improved by 17.3% after long-term operation), forming a complete closed loop of "real-time control-prediction optimization-rule evolution" (such as Figure 2 shown).

[0193] 2. Industry Specificity of Dual-Objective Optimization and Constraints

[0194] Objective function J = Σe 2 +λ·ΣQ slurry The energy efficiency weight λ (dynamically adjusted from 0.1 to 0.5) and constraints (emission limits, slurry pump flow rate change rate ≤ 2.8%) are designed to address the pain points of the waste incineration industry:

[0195] Balance between emission compliance and cost control: The λ parameter enables the system to flexibly switch between strict environmental protection requirements (e.g., focusing on emissions when λ=0.1) and economic operation (e.g., focusing on costs when λ=0.5). Compared with traditional methods that only pursue compliance without considering costs (e.g., the traditional PID slaked lime consumption is 12.8 kg / ton of garbage, while the present invention reduces it to 10.8 kg / ton, a reduction of 15.6%), it has significant economic benefits.

[0196] Necessity of equipment safety constraints: The constraint of slurry flow rate change rate ≤ 2.8% prevents mechanical damage caused by frequent and large adjustments of pumps and valves (traditional methods do not have this constraint, and the measured overshoot frequency reaches 3.2 times / day, while the present invention reduces it to 1.5 times / day, a reduction of 53%), thereby improving equipment life.

[0197] (3) Indivisibility of the technical solution

[0198] 1. Dynamic weight calculation depends on the input of the prediction module: K in step S102 T-P The temperature / pressure data needs to be filtered by the prediction module (such as filtering when the sensor signal-to-noise ratio (SNR) is less than 10dB in a high temperature environment) to ensure the accuracy of the weight calculation.

[0199] 2. The reward function of reinforcement learning contains the goal of predictive optimization: the cost term Cost in the reward function actual The cumulative value of slurry flow rate ΣQ in direct correlation prediction optimization slurry , forming a closed-loop feedback of "control-optimization-learning", none of which can be missing.

[0200] 3. Industry-specific adaptability: All algorithm parameters (such as temperature and pressure range, λ value, and constraints) are customized based on the characteristics of the waste incineration process, rather than parameter adjustments of general control algorithms, reflecting targeted solutions to technical problems in this field.

[0201] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A multivariable dynamic collaborative optimization control method for waste incineration flue gas deacidification, characterized in that: include: Step S101: Acquire real-time data for preprocessing and calculate the temperature-pressure coupling coefficient K T-P ; Step S102: Using the pre-processed real-time data and the temperature-pressure coupling coefficient K T-P , calculate the rule weight W i ; Step S103: Using the rule weight W i After fuzzy rule matching, the output of the preliminary control quantity ΔQ slurry , where ΔQ slurry is the change in slurry flow rate; Step S104: perform prediction-energy efficiency optimization, solve the optimal control quantity through the objective function and constraint conditions, and output the optimization instruction to the execution mechanism; Step S105: Update the rule base according to the current system state and the pre-set reward function.

2. The multivariable dynamic collaborative optimization control method for waste incineration flue gas deacidification according to claim 1 is characterized in that: In step S101, the real-time data includes HCl concentration, SO2 concentration, temperature and pressure; And / or, the temperature-pressure coupling coefficient K T-P The calculation formula is: Among them, T is the real-time temperature value, T min is the lower limit of temperature, T max is the upper temperature limit, P is the real-time pressure value, P min is the lower limit of pressure, P max The upper limit of pressure.

3. The multivariable dynamic collaborative optimization control method for waste incineration flue gas deacidification according to claim 1 is characterized in that: In step S102, the rule weight W i The calculation formula is: W i =α·(C SO2 / C HCl )+(1-α)·K T-P Among them, α is the adaptive coefficient, according to C HCl / C SO2 Dynamic adjustment of concentration ratio, C SO2 / C HCl is the concentration ratio of SO2 and HCl.

4. The multivariable dynamic collaborative optimization control method for waste incineration flue gas deacidification according to claim 1 is characterized in that: The step S104 includes: Step S1041: Based on real-time data and historical data, a prediction module is used to output predicted values of HCl and SO2 concentrations within a preset time period in the future; Step S1042: constructing an objective function; Step S1043: Quadratic programming to solve the optimal ΔQ slurry ; Step S1044: Output optimization instructions under pre-set constraints.

5. The multivariable dynamic collaborative optimization control method for waste incineration flue gas deacidification according to claim 4 is characterized in that: In step S1042, the objective function is: J=In 2 +λ·ΣQ slurry Where e represents the deviation of HCl and SO2 concentrations, λ is the energy efficiency weight, and Q slurry represents the slurry flow rate; And / or, in step S1043, the quadratic programming formula is: Among them, e HCL,t represents the deviation of HCl concentration at time t, e SO2,t represents the deviation of SO2 concentration at time t, λ is the energy efficiency weight, Q slurry,t represents the slurry flow rate at time t, and N is the number of time steps in the prediction domain or optimization domain.

6. The multivariable dynamic collaborative optimization control method for waste incineration flue gas deacidification according to claim 5 is characterized in that: In step S1044, the preset constraint conditions are: Emission limit: HCl ≤ 10 mg / m 3 , SO2≤50mg / m 3 ; Equipment limitation: Maximum flow rate Q of slurry pump max =500 l / min; the change in slurry flow rate ΔQslurry,t≤2.8%·Qslurry,t-1(t≥2), that is, the change in slurry flow rate does not exceed 2.8% of the flow rate at the previous moment.

7. The multivariable dynamic collaborative optimization control method for waste incineration flue gas deacidification according to claim 1 is characterized in that: The step S105 includes: Step S1051: Initialize the fuzzy rule base and Q table; Step S1052: quantify the current control state into a state value S, and continuously obtain the state S of the current system; Step S1053: Determine whether the termination condition is met. If not, execute step S1054. If the termination condition is met, output the optimized fuzzy rule base and end the process. Step S1054: Generate control action A according to the current rule base; Step S1055: Execute the selected action A, the system state changes to the new state S', and the reward R is calculated based on the state change and system performance; Step S1056: Update Q table; Step S1057: Update the fuzzy rule base according to the new Q value; Step S1058: Update the current state to the new state S', and go to step S1053.

8. The multivariable dynamic collaborative optimization control method for waste incineration flue gas deacidification according to claim 7 is characterized in that: In step S1052, the current system state S is expressed as: Among them, e HCl and e SO2 are the concentration deviations of HCl and SO2, Q slurry is the slurry flow rate, ΔQ slurry is the change in slurry flow rate, C max and Q max are the maximum values of the corresponding variables.

9. The multivariable dynamic collaborative optimization control method for waste incineration flue gas deacidification according to claim 7, characterized in that: In step S1055, the reward function formula is: Among them, C meas is the actual measured concentration, C lim is the emission limit; Cost actual Cost is the actual cost. ref is the reference cost; oscillation_penalty is the oscillation of the system.

10. The multivariable dynamic collaborative optimization control method for waste incineration flue gas deacidification according to claim 7, characterized in that: The step S1056 includes: Update the Q table according to the reward R and the new state S', and use the Q-learning algorithm to update the Q value. The specific formula is as follows: Q(S,A)←Q(S,A)+η·[R+γ·maxQ(S',A′)-Q(S,A)] Among them, maxQ(S',A') is the maximum expected reward of the next state, η is the learning rate, and γ is the discount factor.

Images