Intelligent Control Method and System for Ammonia Synthesis Production Load Based on Wind and Solar Prediction

Abstract

This invention provides a method and system for intelligent load control in ammonia synthesis production based on wind and solar forecasting, relating to the fields of energy and chemical production technology. The method includes collecting wind and solar data and ammonia synthesis process data; using a dual-current variational autoencoder combined with a residual connection mechanism to predict power generation curves and process parameter trends; determining process parameter constraints and safety boundaries; establishing a process parameter safety constraint model; obtaining the load adjustment space; generating optimal load commands based on optimization algorithms; and achieving intelligent load control through cascaded PID control. This invention achieves precise dynamic control of the load on ammonia synthesis units, improving wind and solar energy absorption capacity and production efficiency.

Description

Technical Field

[0001] This invention relates to the field of energy and chemical production technology, and in particular to a method and system for intelligent control of synthetic ammonia production load based on wind and solar forecasts. Background Technology

[0002] With the transformation of the energy structure and the development of renewable energy, the installed capacity of renewable energy sources such as wind power and photovoltaics is constantly expanding. However, due to the intermittent, fluctuating, and random characteristics of wind and solar power, their large-scale grid connection poses challenges to the stable operation of the power system. As a major energy consumer, the synthetic ammonia industry, under the traditional model, suffers from high energy consumption, low production efficiency, and stringent requirements for process conditions, resulting in a relatively rigid production process. With the increasing proportion of renewable energy, how to utilize wind and solar power generation forecasts to intelligently regulate the load of synthetic ammonia plants and achieve efficient matching between the energy system and production load has become an urgent problem for the industry to solve.

[0003] Currently, the coordinated operation of ammonia synthesis production and renewable energy mainly relies on manual experience and simple control strategies, which have many shortcomings. The main defects of existing technologies are:

[0004] The existing wind and solar power forecasting and ammonia synthesis production system lack an effective coordination mechanism. The forecasting of wind and solar power output is often handled separately from the adjustment of ammonia synthesis process parameters, which fails to fully consider the correlation between the two. This results in untimely or unreasonable process adjustments when renewable energy fluctuates significantly.

[0005] Traditional methods for controlling the load of synthetic ammonia lack comprehensive constraint analysis of process parameters. They usually only consider the limitations of one or a few key parameters, ignoring the complex coupling relationship between process parameters. This can easily cause excessive fluctuations in process parameters during load adjustment, affecting production safety and product quality.

[0006] Existing load optimization scheduling strategies mostly adopt static programming methods, which cannot adapt to the dynamic changes in wind and solar resources. Furthermore, they lack consideration of the fluctuation patterns of historical process data during ammonia synthesis as constraints, resulting in a mismatch between the generated load control instructions and actual process requirements, thus reducing overall operating efficiency and economy. Summary of the Invention

[0007] This invention provides a method and system for intelligent control of synthetic ammonia production load based on wind and solar forecasting, which can solve the problems in the prior art.

[0008] A first aspect of this invention provides a method for intelligent control of ammonia synthesis production load based on wind and solar forecasts, comprising:

[0009] Data from wind farms and photovoltaic power plants were collected as a meteorological sample set, while data from ammonia synthesis units were collected as a process sample set.

[0010] The meteorological sample set and the process sample set are processed by a dual-flow variational autoencoder, and the temporal features of the meteorological data and process data are extracted by the coding network. The latent features are generated by combining the residual connection mechanism. Based on the latent features, the wind and solar power generation prediction curves and the change trends of process parameters of the ammonia synthesis unit are generated for future periods.

[0011] Based on the wind and solar power generation prediction curve and the process parameter variation trend, calculate the upper and lower limits of process parameters and the operating condition safety boundary of the ammonia synthesis unit at each time point; determine the coupling relationship between process parameters based on historical operating data and establish a safety constraint model for process parameters; combine the upper and lower limits of process parameters and the safety constraint model for process parameters to obtain the load adjustment space of the ammonia synthesis unit under different wind and solar power output conditions.

[0012] Based on the load adjustment space, and combined with the load optimization algorithm, the fluctuation pattern of historical process data is used as the optimization constraint to generate the initial solution set of load allocation for each time period and calculate the fitness value. The release amount and volatility coefficient of pheromones are dynamically adjusted according to the fitness value. When the change value of the optimal solution in continuous iteration is less than the preset convergence threshold, the time-sharing optimal load control command of the ammonia synthesis unit is output.

[0013] The optimal load control command is converted into control parameters, which are then decomposed into target values ​​of process parameters. These parameters are then adjusted step by step using a cascade PID controller to achieve intelligent control of the ammonia synthesis production load.

[0014] The meteorological sample set and the process sample set are processed separately using a dual-flow variational autoencoder. Temporal features of the meteorological and process data are extracted through an encoding network. Latent features are generated using a residual connection mechanism. Based on these latent features, a future wind and solar power generation prediction curve and the trend of process parameters in the ammonia synthesis unit are generated, including:

[0015] The meteorological sample set is used as the first data stream, and the process sample set is used as the second data stream. The first data stream and the second data stream are respectively input into the first variational autoencoder and the second variational autoencoder.

[0016] Residual connection paths are established between the encoder and decoder of the first variational autoencoder and the second variational autoencoder, respectively. The residual connection paths add the original features output from the previous layer to the transformed features output from the current layer to generate the first latent feature and the second latent feature.

[0017] Calculate the temporal cross-correlation coefficient between the first latent feature and the second latent feature, determine the feature fusion weight based on the temporal cross-correlation coefficient, and sum the first latent feature and the second latent feature according to their corresponding weights to obtain the fused feature;

[0018] The fused features and historical data are combined to form a sliding time window. Based on the data sequence in the sliding time window, the power generation curves of wind farms, power generation curves of photovoltaic power plants, and the trend of process parameter changes of ammonia synthesis units for future periods are generated.

[0019] Based on the wind and solar power generation prediction curves and the changing trends of the process parameters, the upper and lower limits of the process parameters and the safety boundary of the operating conditions for the ammonia synthesis unit at each time point are calculated; the coupling relationship between the process parameters is determined based on historical operating data, and a safety constraint model for the process parameters is established; combining the upper and lower limits of the process parameters and the safety constraint model for the process parameters, the load adjustment space of the ammonia synthesis unit under different wind and solar power output conditions is obtained, including:

[0020] Based on the predicted value of the wind and solar power generation prediction curve at the current moment, determine the upper and lower limits of the raw gas pressure, the upper and lower limits of the circulating gas pressure, and the upper and lower limits of the synthesis tower temperature, and obtain the upper and lower limits of the process parameters and the safety boundary of the operating conditions.

[0021] Extract steady-state operating data of feed gas pressure and circulating gas pressure from historical operating data, calculate the linear correlation coefficient between feed gas pressure and circulating gas pressure, extract dynamic operating data, determine the response time difference and response amplitude ratio between synthesis tower temperature and feed gas pressure and circulating gas pressure, and input the linear correlation coefficient, response time difference and response amplitude ratio into the safety constraint evaluation function to generate a process parameter safety constraint model.

[0022] Substituting the upper and lower limits of the process parameters and the safety boundary of the operating conditions into the safety constraint model of the process parameters, the steady-state constraint range of the feed gas pressure and the circulating gas pressure, and the dynamic response constraint range of the synthesis tower temperature are calculated respectively. The intersection of the steady-state constraint range and the dynamic response constraint range is taken to obtain the load adjustment space of the ammonia synthesis unit under different wind and solar power output conditions.

[0023] Based on the aforementioned load adjustment space, and combined with the load optimization algorithm, the fluctuation patterns of historical process data are used as optimization constraints to generate an initial solution set for load allocation in each time period and calculate the fitness value. The release amount and volatility coefficient of pheromones are dynamically adjusted according to the fitness value. When the change value of the optimal solution in continuous iterations is less than a preset convergence threshold, the time-segmented optimal load control command for the ammonia synthesis unit is output, including:

[0024] Extract the continuous load change sequence from historical process data, and statistically analyze the distribution intervals of the absolute value of load change and the absolute value of load change rate between two adjacent time points. Set the upper limit of the distribution interval as the upper limit of load change amount and the upper limit of load change rate, respectively, to obtain the load optimization constraint boundary.

[0025] The load optimization algorithm is constructed based on the ant colony algorithm. Load solutions that meet the ant colony size requirements are randomly generated within the load adjustment space. It is determined whether the load change and load change rate at adjacent time points in each load solution are within the load optimization constraint boundary. Load solutions that meet the judgment conditions are retained as the initial load solution set. The load target deviation and load fluctuation amplitude of each load solution are calculated. The load target deviation and load fluctuation amplitude are assigned corresponding weights respectively. The result of the weighted sum is used as the fitness value.

[0026] The baseline pheromone release is determined based on the number of iterations completed. The ratio of the optimal fitness value to the average fitness value of the current iteration is multiplied by the pheromone concentration decay coefficient to obtain the pheromone concentration adjustment factor. The baseline release is adjusted based on the pheromone concentration adjustment factor to obtain the actual pheromone release. The actual pheromone release is then updated to the pheromone concentration matrix.

[0027] A new load solution set is generated according to the numerical distribution of the pheromone concentration matrix. The fitness value is repeatedly calculated and the pheromone concentration matrix is ​​updated. When the continuous iteration reaches a predetermined number of times and the change in the optimal fitness value is less than the preset convergence threshold, the corresponding load solution is output as the time-sharing optimal load control command for the ammonia synthesis unit.

[0028] The optimal load control command is converted into control parameters, which are then decomposed into target values ​​for process parameters. Intelligent load control for ammonia synthesis production is achieved through step-by-step adjustment using a cascaded PID controller.

[0029] The optimal load control command is converted into ammonia synthesis process control parameters. The time-series characteristics of the process parameters are obtained based on the response relationship between parameters in the historical operating data of the process parameters. Based on the time-series characteristics, the ammonia synthesis process control parameters are decomposed into process parameter target values.

[0030] The target value of the process parameter is used as the setpoint of the main loop PID controller. The deviation sequence between the actual output value of the main loop PID controller and the target value of the process parameter is calculated. The deviation sequence is weighted by time and integrated to obtain the integral time absolute error index. The adjustment direction and adjustment step size of the main loop PID controller parameters are calculated using the integral time absolute error index. The proportional coefficient, integral coefficient and derivative coefficient of the main loop are updated. The updated output value of the main loop PID controller is set as the setpoint of the slave loop PID controller.

[0031] The actual adjustment amount of the process parameters is obtained by adjusting them step by step through a cascade PID controller and then sent to the corresponding actuators to realize intelligent load control of the ammonia synthesis unit.

[0032] Intelligent load control of the ammonia synthesis unit is achieved by using a cascaded PID controller to adjust process parameters step by step, obtaining the actual adjustment values, and then sending them to the corresponding actuators.

[0033] Extract dynamic response data between process parameters, statistically analyze the fluctuation transmission relationship and response timing characteristics between process parameters, generate coupling coefficient matrix and timing adjustment coefficient of process parameters, superimpose the output value of the loop PID controller with the compensation amount of other loops calculated based on the coupling coefficient matrix to obtain the initial control quantity, calculate the execution timing delay of each loop according to the timing adjustment coefficient, and adjust the initial control quantity in stages according to the execution timing delay;

[0034] The response characteristic curves of the actuator under different input signals are collected, and the gain characteristic and delay characteristic in the response characteristic curve are extracted. The feedforward compensation gain is calculated based on the gain characteristic, and the feedforward derivative time is determined based on the delay characteristic. The feedforward compensation gain and the feedforward derivative time are applied to the initial control quantity after graded adjustment to obtain the compensated execution command.

[0035] The compensated execution command is sent to the actuator and the actual output value of the actuator is collected. The execution deviation sequence between the compensated execution command and the actual output value is calculated. The dynamic correction coefficient is determined according to the changing trend of the execution deviation sequence. The execution command corrected based on the dynamic correction coefficient is resent to the actuator to realize intelligent load control of the ammonia synthesis unit.

[0036] A second aspect of the present invention provides an intelligent control system for synthetic ammonia production load based on wind and solar forecasting, comprising:

[0037] The first unit is used to collect data from wind farms and photovoltaic power stations as a meteorological sample set, and at the same time, to collect data from ammonia synthesis units as a process sample set.

[0038] The second unit is used to process the meteorological sample set and the process sample set respectively using a dual-flow variational autoencoder, extract the temporal features of meteorological data and process data through the coding network, generate latent features by combining the residual connection mechanism, and generate the wind and solar power generation prediction curve and the process parameter change trend of the ammonia synthesis unit for future periods based on the latent features.

[0039] The third unit is used to calculate the upper and lower limits of process parameters and the safety boundary of operating conditions of the ammonia synthesis unit at each time point based on the wind and solar power generation prediction curve and the process parameter change trend; determine the coupling relationship between process parameters based on historical operating data and establish a safety constraint model for process parameters; and combine the upper and lower limits of process parameters and the safety constraint model for process parameters to obtain the load adjustment space of the ammonia synthesis unit under different wind and solar power output conditions.

[0040] The fourth unit is used to generate an initial solution set for load allocation in each time period based on the load adjustment space and the load optimization algorithm, taking the fluctuation pattern of historical process data as the optimization constraint, and calculating the fitness value. The release amount and volatility coefficient of pheromones are dynamically adjusted according to the fitness value. When the change value of the optimal solution in continuous iteration is less than the preset convergence threshold, the unit outputs the time-sharing optimal load control command of the ammonia synthesis unit.

[0041] The fifth unit is used to convert the optimal load control command into control parameters, decompose the control parameters into process parameter target values, and realize intelligent control of the synthetic ammonia production load through step-by-step adjustment by a cascade PID controller.

[0042] A third aspect of the present invention,

[0043] An electronic device is provided, comprising:

[0044] processor;

[0045] Memory used to store processor-executable instructions;

[0046] The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0047] Fourth aspect of the embodiments of the present invention,

[0048] A computer-readable storage medium is provided, having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0049] The beneficial effects of this invention are as follows:

[0050] The intelligent control method for synthetic ammonia production load based on wind and solar forecasting provided by this invention, through a dual-flow variational autoencoder combined with a residual connection mechanism, can simultaneously process meteorological data and process data, thereby achieving accurate prediction of wind and solar power generation and accurate grasp of process parameter change trends, improving the accuracy and reliability of forecasting.

[0051] This invention establishes a safety constraint model for process parameters, fully considers the coupling relationship between process parameters, and determines the load adjustment space of the ammonia synthesis unit under different wind and solar power output conditions. This ensures the safe and stable operation of production, maximizes the utilization of renewable energy, and achieves an effective balance between energy consumption and production safety.

[0052] This invention dynamically adjusts the pheromone release and volatility coefficient through a load optimization algorithm, and combines it with a cascade PID controller to achieve precise control of the optimal load of the ammonia synthesis unit in different time periods. This significantly improves the response speed and control accuracy, reduces energy waste, lowers production costs, and provides technical support for the efficient utilization of renewable energy sources such as wind and solar power and the green production of chemical industries such as ammonia synthesis. Attached Figure Description

[0053] Figure 1 This is a schematic flowchart of the intelligent load control method for synthetic ammonia production based on wind and solar forecasting, according to an embodiment of the present invention.

[0054] Figure 2 A diagram illustrating the performance comparison across different prediction tasks;

[0055] Figure 3 This is a schematic diagram comparing the convergence performance during the load regulation optimization process. Detailed Implementation

[0056] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0057] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0058] Figure 1 This is a schematic flowchart of the intelligent load control method for synthetic ammonia production based on wind and solar forecasting, as described in an embodiment of the present invention. Figure 1 As shown, the method includes:

[0059] Data from wind farms and photovoltaic power plants were collected as a meteorological sample set, while data from ammonia synthesis units were collected as a process sample set.

[0060] The meteorological sample set and the process sample set are processed by a dual-flow variational autoencoder, and the temporal features of the meteorological data and process data are extracted by the coding network. The latent features are generated by combining the residual connection mechanism. Based on the latent features, the wind and solar power generation prediction curves and the change trends of process parameters of the ammonia synthesis unit are generated for future periods.

[0061] Based on the wind and solar power generation prediction curve and the process parameter variation trend, calculate the upper and lower limits of process parameters and the operating condition safety boundary of the ammonia synthesis unit at each time point; determine the coupling relationship between process parameters based on historical operating data and establish a safety constraint model for process parameters; combine the upper and lower limits of process parameters and the safety constraint model for process parameters to obtain the load adjustment space of the ammonia synthesis unit under different wind and solar power output conditions.

[0062] Based on the load adjustment space, and combined with the load optimization algorithm, the fluctuation pattern of historical process data is used as the optimization constraint to generate the initial solution set of load allocation for each time period and calculate the fitness value. The release amount and volatility coefficient of pheromones are dynamically adjusted according to the fitness value. When the change value of the optimal solution in continuous iteration is less than the preset convergence threshold, the time-sharing optimal load control command of the ammonia synthesis unit is output.

[0063] The optimal load control command is converted into control parameters, which are then decomposed into target values ​​of process parameters. These parameters are then adjusted step by step using a cascade PID controller to achieve intelligent control of the ammonia synthesis production load.

[0064] In one optional implementation, a dual-flow variational autoencoder is used to process the meteorological sample set and the process sample set respectively. Temporal features of the meteorological and process data are extracted through an encoding network, and latent features are generated using a residual connection mechanism. Based on these latent features, a future wind and solar power generation prediction curve and a trend of process parameter changes in the ammonia synthesis unit are generated, including:

[0065] The meteorological sample set is used as the first data stream, and the process sample set is used as the second data stream. The first data stream and the second data stream are respectively input into the first variational autoencoder and the second variational autoencoder.

[0066] Residual connection paths are established between the encoder and decoder of the first variational autoencoder and the second variational autoencoder, respectively. The residual connection paths add the original features output from the previous layer to the transformed features output from the current layer to generate the first latent feature and the second latent feature.

[0067] Calculate the temporal cross-correlation coefficient between the first latent feature and the second latent feature, determine the feature fusion weight based on the temporal cross-correlation coefficient, and sum the first latent feature and the second latent feature according to their corresponding weights to obtain the fused feature;

[0068] The fused features and historical data are combined to form a sliding time window. Based on the data sequence in the sliding time window, the power generation curves of wind farms, power generation curves of photovoltaic power plants, and the trend of process parameter changes of ammonia synthesis units for future periods are generated.

[0069] Specifically, the meteorological sample set includes historical data on meteorological parameters such as wind speed, wind direction, temperature, humidity, air pressure, and light intensity. This data was collected from meteorological stations and preprocessed. The process sample set includes historical data on process parameters such as wind farm power generation, photovoltaic power plant power generation, feed rate, temperature, pressure, and catalyst activity in ammonia synthesis units. This data was collected from industrial control systems and preprocessed. The preprocessing process includes steps such as data normalization, missing value imputation, and outlier handling to ensure data quality and consistency.

[0070] Meteorological sample sets are used as the first data stream, and process sample sets are used as the second data stream, which are input into the first variational autoencoder (VAC) and the second variational autoencoder, respectively. Each VAC consists of an encoder and a decoder. In the first VAC, the encoder is composed of three convolutional neural network layers, each with a kernel size of 3×3, a stride of 1, and 32, 64, and 128 channels, respectively. Each convolutional layer is followed by a batch normalization layer and a ReLU activation function. The decoder is composed of three transposed convolutional network layers, with a structure symmetrical to the encoder. Similarly, the second VAC has the same network structure, but its parameters are trained independently.

[0071] In the first variational autoencoder, the output features of the first layer are added to the output features of the second layer, and the output features of the second layer are added to the output features of the third layer to generate the first latent feature. Similarly, in the second variational autoencoder, the same residual connection method is used to generate the second latent feature. Taking the first variational autoencoder as an example, if the output feature of the first convolutional layer is F1 with dimensions [batch_size, 32, height, width], and the output feature of the second convolutional layer is F2 with dimensions [batch_size, 64, height / 2, width / 2], then the number of channels of F1 needs to be adjusted from 32 to 64 through a 1×1 convolution, and the spatial dimension of F1 needs to be halved through average pooling with a stride of 2 to match the dimension of F2. Then, the two are added together to obtain the residual feature.

[0072] For a real-world example, suppose the meteorological sample set contains wind speed data for a wind farm over the past 30 days, with a sampling frequency of once per hour and a data dimension of [720, 1] (wind speed values ​​over 720 hours). After processing by the first variational autoencoder, the first latent feature is obtained, with a dimension of [batch_size, 128, 9, 9]. Similarly, the process sample set contains wind farm power generation data for the corresponding time period. After processing by the second variational autoencoder, the second latent feature is obtained, also with a dimension of [batch_size, 128, 9, 9].

[0073] When the time lag is 0, calculate the correlation coefficient between v1 and v2; when the time lag is 1, calculate the correlation coefficient between v1 and v2[1:].

[0074] The feature fusion weight calculation method is as follows: The time-series cross-correlation coefficient is normalized. If the cross-correlation coefficient is positive and greater than the threshold of 0.5, the weight of the first latent feature is the cross-correlation value, and the weight of the second latent feature is 1 minus the cross-correlation value. If the cross-correlation coefficient is less than 0.5, both features have a weight of 0.5. In the example above, since the cross-correlation coefficient is 0.72, which is greater than the threshold of 0.5, the weight of the first latent feature is 0.72, and the weight of the second latent feature is 0.28.

[0075] Multiply each element of the first latent feature by its weight of 0.72, multiply each element of the second latent feature by its weight of 0.28, and then add them together at the corresponding positions to obtain the fused feature, which still has the dimension [bat ch_size,128,9,9].

[0076] The fused features and historical data are combined to form a sliding time window with a window size of 24 hours and a sliding step of 1 hour. For each time point, the fused features and historical data from the previous 24 hours are taken and input into the prediction model to generate prediction results for the future period. The prediction model uses a Long Short-Term Memory (LSTM) network, containing two LSTM layers with a hidden layer size of 256, a fully connected output layer, and the Sigma-Ald function as the activation function. The final outputs are the power generation curves of wind farms, photovoltaic power plants, and the trend of process parameters of the ammonia synthesis unit for the next 24 hours.

[0077] The above implementation method effectively utilizes the correlation between meteorological and process data to improve prediction accuracy. Experimental results show that, compared with traditional prediction methods, this method reduces the average absolute percentage error in wind power prediction by 15.3%, the root mean square error in photovoltaic power prediction by 12.7%, and improves the accuracy of predicting key process parameters for ammonia synthesis plants by 18.6%.

[0078] Figure 2This diagram illustrates the performance comparison between the method of this invention and three traditional prediction methods across different prediction tasks. The diagram clearly shows that the dual-flow variational autoencoder method of this invention significantly outperforms the traditional methods in all three prediction metrics. Regarding the mean absolute percentage error (MAPE) in wind power prediction, the error rate is 28.4% for the ARIMA model, 24.7% for support vector regression, and 22.1% for the BP neural network, while the method of this invention is only 10.3%, a reduction of 53.5% compared to the best traditional method, the BP neural network. In terms of the root mean square error (RMSE) in photovoltaic power prediction, the error rates of traditional methods are 21.3%, 18.9%, and 16.4%, respectively, while the method of this invention reduces the error rate to 6.2%, demonstrating superior prediction accuracy. Most significantly, in terms of the accuracy of process parameter prediction for ammonia synthesis plants, the accuracy of traditional methods is below 75%, with the ARIMA model at 67.2%, support vector regression at 72.8%, and the BP neural network at 75.3%, while the method of this invention achieves a high accuracy of 91.0%, an improvement of over 15 percentage points. This result fully verifies the technical advantages of this invention by integrating meteorological and process data and using a residual connection mechanism to extract potential features, providing a more reliable predictive basis for the intelligent control of ammonia synthesis units.

[0079] In one optional implementation, based on the wind and solar power generation prediction curve and the process parameter variation trend, the upper and lower limits of the process parameters and the operating condition safety boundary of the ammonia synthesis unit at each time are calculated; the coupling relationship between process parameters is determined based on historical operating data, and a safety constraint model for the process parameters is established; the upper and lower limits of the process parameters and the safety constraint model for the process parameters are combined to obtain the load adjustment space of the ammonia synthesis unit under different wind and solar power output conditions, including:

[0080] Based on the predicted value of the wind and solar power generation prediction curve at the current moment, determine the upper and lower limits of the raw gas pressure, the upper and lower limits of the circulating gas pressure, and the upper and lower limits of the synthesis tower temperature, and obtain the upper and lower limits of the process parameters and the safety boundary of the operating conditions.

[0081] Extract steady-state operating data of feed gas pressure and circulating gas pressure from historical operating data, calculate the linear correlation coefficient between feed gas pressure and circulating gas pressure, extract dynamic operating data, determine the response time difference and response amplitude ratio between synthesis tower temperature and feed gas pressure and circulating gas pressure, and input the linear correlation coefficient, response time difference and response amplitude ratio into the safety constraint evaluation function to generate a process parameter safety constraint model.

[0082] Substituting the upper and lower limits of the process parameters and the safety boundary of the operating conditions into the safety constraint model of the process parameters, the steady-state constraint range of the feed gas pressure and the circulating gas pressure, and the dynamic response constraint range of the synthesis tower temperature are calculated respectively. The intersection of the steady-state constraint range and the dynamic response constraint range is taken to obtain the load adjustment space of the ammonia synthesis unit under different wind and solar power output conditions.

[0083] After obtaining the predicted wind and solar power generation curves, a mapping relationship between the process parameters of the ammonia synthesis unit and the wind and solar power generation was established based on historical operating data. At a wind and solar power generation capacity of 50MW, the upper limit of the feed gas pressure was 28MPa, and the lower limit was 22MPa; the upper limit of the recirculating gas pressure was 25MPa, and the lower limit was 19MPa; the upper limit of the synthesis tower temperature was 520℃, and the lower limit was 450℃. At a wind and solar power generation capacity of 100MW, the upper limit of the feed gas pressure was 32MPa, and the lower limit was 25MPa; the upper limit of the recirculating gas pressure was 28MPa, and the lower limit was 22MPa; the upper limit of the synthesis tower temperature was 530℃, and the lower limit was 460℃. The parameter constraint ranges at other power points were obtained through interpolation calculations, thus establishing complete upper and lower limit constraints on process parameters and safety boundaries for operating conditions.

[0084] From the historical database of process parameters of the ammonia synthesis unit, process parameters were selected over a continuous 24-hour period. 900 data points of feed gas pressure and circulating gas pressure were acquired at 15-second sampling intervals. Data segments of feed gas pressure and circulating gas pressure during stable operation were selected as steady-state operating data. The steady-state operating data were standardized, and the standardized feed gas pressure and circulating gas pressure data were substituted into the Pearson correlation coefficient calculation formula. The linear correlation coefficient between feed gas pressure and circulating gas pressure was found to be 0.87. Based on this linear correlation coefficient, the linear regression equation for feed gas pressure and circulating gas pressure was determined as: Circulating gas pressure = 0.85 × Feed gas pressure + 2.3 (MPa). Dynamic operating data were extracted from the historical process parameter database, showing the feed gas pressure increasing from 24 MPa to 28 MPa, the circulating gas pressure increasing from 20 MPa to 26 MPa, and the synthesis tower temperature increasing from 470°C to 500°C during load switching. Based on this dynamic operating data, the response time difference of the synthesis tower temperature to changes in feed gas pressure was calculated to be 15 minutes, and the response time difference to changes in circulating gas pressure was calculated to be 12 minutes. According to the correspondence of parameter changes in the dynamic operating data, the response amplitude ratio of the synthesis tower temperature change to the feed gas pressure change was calculated to be 7.5°C / MPa, and the response amplitude ratio of the synthesis tower temperature change to the circulating gas pressure change was calculated to be 5°C / MPa.

[0085] The safety constraint evaluation function includes: a static coupling term for feed gas pressure and circulating gas pressure calculated based on the linear correlation coefficient; a dynamic response term for synthesis tower temperature to feed gas pressure calculated based on the response time difference and response amplitude ratio; and a dynamic response term for synthesis tower temperature to circulating gas pressure calculated based on the response time difference and response amplitude ratio. The process parameter variation boundaries are calculated based on the safety constraint evaluation function, including a feed gas pressure variation range of 24 MPa to 28 MPa with a variation rate not exceeding 0.5 MPa / min, a circulating gas pressure variation range of 20 MPa to 26 MPa with a variation rate not exceeding 0.4 MPa / min, and a synthesis tower temperature variation range of 470℃ to 500℃ with a variation rate not exceeding 3℃ / min. These process parameter variation boundaries form the process parameter safety constraint model.

[0086] When the predicted wind and solar power generation is 75MW, substituting the upper and lower limits of process parameters and the operating condition safety boundary into the process parameter safety constraint model, the constraint range for feed gas pressure is obtained as [23.5MPa, 30MPa], and the constraint range for recirculating gas pressure is [20.5MPa, 26.5MPa]. Considering the linear correlation between the two, the steady-state constraint ranges for feed gas pressure and recirculating gas pressure are adjusted to [24MPa, 29MPa] and [21MPa, 26MPa], respectively.

[0087] Based on the dynamic response characteristics of the synthesis tower temperature, the dynamic response constraint range of the synthesis tower temperature under changes in feed gas pressure and circulating gas pressure is calculated to be [460℃, 510℃]. Taking the intersection of the steady-state constraint range and the dynamic response constraint range, the load adjustment space of the ammonia synthesis unit at a wind and solar power generation capacity of 75MW is finally obtained as follows: feed gas pressure [24MPa, 29MPa], circulating gas pressure [21MPa, 26MPa], synthesis tower temperature [460℃, 510℃].

[0088] Based on historical data analysis, when the feed gas pressure is 24 MPa, the circulating gas pressure is 21 MPa, and the synthesis tower temperature is 460℃, the ammonia synthesis unit's output is approximately 70% of its rated output; when the feed gas pressure is 29 MPa, the circulating gas pressure is 26 MPa, and the synthesis tower temperature is 510℃, the ammonia synthesis unit's output is approximately 95% of its rated output. Therefore, under a wind and solar power generation capacity of 75 MW, the load adjustment range of the ammonia synthesis unit is 70% to 95% of its rated output.

[0089] By repeating the above calculation process, the load adjustment space of the ammonia synthesis unit under different wind and solar power output conditions can be obtained, establishing a complete mapping relationship between wind and solar power output and ammonia synthesis load adjustment space. This method, by comprehensively considering the upper and lower limits of process parameters, the coupling relationship between parameters, and dynamic response characteristics, achieves accurate calculation of the load adjustment space of the ammonia synthesis unit, providing technical support for the coordinated operation of wind and solar power generation and ammonia synthesis units.

[0090] In one optional implementation, based on the load adjustment space, and combined with a load optimization algorithm, the fluctuation patterns of historical process data are used as optimization constraints to generate an initial solution set for load allocation in each time period and calculate the fitness value. The release amount and volatility coefficient of pheromones are dynamically adjusted according to the fitness value. When the change value of the optimal solution in continuous iterations is less than a preset convergence threshold, the time-segmented optimal load control command for the ammonia synthesis unit is output, including:

[0091] Extract the continuous load change sequence from historical process data, and statistically analyze the distribution intervals of the absolute value of load change and the absolute value of load change rate between two adjacent time points. Set the upper limit of the distribution interval as the upper limit of load change amount and the upper limit of load change rate, respectively, to obtain the load optimization constraint boundary.

[0092] The load optimization algorithm is constructed based on the ant colony algorithm. Load solutions that meet the ant colony size requirements are randomly generated within the load adjustment space. It is determined whether the load change and load change rate at adjacent time points in each load solution are within the load optimization constraint boundary. Load solutions that meet the judgment conditions are retained as the initial load solution set. The load target deviation and load fluctuation amplitude of each load solution are calculated. The load target deviation and load fluctuation amplitude are assigned corresponding weights respectively. The result of the weighted sum is used as the fitness value.

[0093] The baseline pheromone release is determined based on the number of iterations completed. The ratio of the optimal fitness value to the average fitness value of the current iteration is multiplied by the pheromone concentration decay coefficient to obtain the pheromone concentration adjustment factor. The baseline release is adjusted based on the pheromone concentration adjustment factor to obtain the actual pheromone release. The actual pheromone release is then updated to the pheromone concentration matrix.

[0094] A new load solution set is generated according to the numerical distribution of the pheromone concentration matrix. The fitness value is repeatedly calculated and the pheromone concentration matrix is ​​updated. When the continuous iteration reaches a predetermined number of times and the change in the optimal fitness value is less than the preset convergence threshold, the corresponding load solution is output as the time-sharing optimal load control command for the ammonia synthesis unit.

[0095] Process data from a synthetic ammonia unit over the past three months, including hourly load records, totaling 2160 data points, were collected. Analysis of this data revealed a maximum absolute load change of 240 tons / hour and a maximum absolute load change rate of 8% between adjacent time points. These two maximum values ​​were set as upper limits for load change and load change rate, respectively, forming load optimization constraint boundaries. These boundary values ​​ensure that the generated load control scheme conforms to the actual process's regulation capacity.

[0096] Assuming the load adjustment range of the ammonia synthesis unit is 3000 to 4500 tons / day, the day is divided into 24 time periods, each corresponding to a load value. The initial ant colony size is set to 100. Within the load adjustment range, 100 load solutions are randomly generated, each containing 24 load values, corresponding to the load setpoints for the 24 time periods of the day. For each load solution, it is necessary to determine whether the load change at adjacent time points is less than or equal to 240 tons / hour and whether the load change rate is less than or equal to 8%. Only load solutions that satisfy these two constraints are retained as the initial load solution set.

[0097] For each valid load solution, calculate its target load deviation and load fluctuation amplitude. The target load deviation is the square root of the sum of squares of the differences between the load values ​​for each time period and the target load value, and the load fluctuation amplitude is the square root of the sum of squares of the load changes for each adjacent time period. Assume that during peak electricity price periods (e.g., 10:00-15:00 and 18:00-21:00), the desired load reduction is to 3500 tons / day; during off-peak electricity price periods (e.g., 23:00-7:00), the desired load increase is to 4300 tons / day; and during other times, a moderate load of 4000 tons / day is maintained. Set the weight of the target load deviation to 0.7 and the weight of the load fluctuation amplitude to 0.3, and sum the two weighted values ​​to obtain the fitness value. The smaller the fitness value, the better the load solution.

[0098] Assuming a maximum number of iterations of 200 and a current iteration number of t, the baseline pheromone release can be set to 100 × (1 - t / 200). This ensures that the baseline pheromone release gradually decreases as the number of iterations increases, which is beneficial for algorithm convergence. The ratio of the optimal fitness value to the average fitness value in the current iteration is multiplied by a pheromone concentration decay coefficient (e.g., set to 0.9) to obtain the pheromone concentration adjustment factor. If the optimal fitness value is 10 and the average fitness value is 25, then the pheromone concentration adjustment factor is 10 / 25 × 0.9 = 0.36. Based on this adjustment factor, the baseline release is adjusted to obtain the actual pheromone release of 100 × (1 - t / 200) × 0.36. This value is then updated to the corresponding position in the pheromone concentration matrix.

[0099] The predetermined number of consecutive iterations is set to 20, and the preset convergence threshold is 0.5%. When the change in the optimal fitness value is less than 0.5% for 20 consecutive iterations, the algorithm is considered to have converged, and the corresponding load solution is output as the time-sharing optimal load control command for the ammonia synthesis unit.

[0100] In practical applications, the optimal load control scheme obtained through this method is assumed to be as follows: From 00:00 to 07:00, the load value stabilizes at 4300 tons / day; from 07:00 to 10:00, the load value gradually decreases from 4300 to 4000 tons / day; from 10:00 to 15:00, the load stabilizes at 3500 tons / day; from 15:00 to 18:00, the load gradually increases from 3500 to 4000 tons / day; from 18:00 to 21:00, the load stabilizes at 3500 tons / day; from 21:00 to 23:00, the load gradually increases from 3500 to 4200 tons / day; and from 23:00 to 00:00, the load stabilizes at 4300 tons / day. This scheme not only satisfies electricity price considerations but also ensures smooth load changes, meets process requirements, and achieves the goal of load optimization.

[0101] Using the above methods, the ammonia synthesis unit can flexibly adjust the load according to external factors such as electricity prices and energy supply conditions, while ensuring that load changes are within the process safety range, thereby improving energy utilization efficiency and reducing operating costs.

[0102] Figure 3 This diagram illustrates the convergence performance comparison between the ant colony optimization algorithm of this invention and three traditional optimization algorithms in the load regulation optimization process. The diagram clearly shows significant differences in the fitness value trends of the four algorithms over 200 iterations. Simulated annealing is the most conservative, with an initial fitness value of 95.3 that only decreases to 63.2 after 200 iterations, exhibiting slow convergence and limited final optimization effect. Particle swarm optimization has an initial fitness value of 87.1, eventually converging to 35.8. Although the optimization range is improved compared to simulated annealing, the convergence speed remains slow. Genetic algorithm demonstrates better optimization ability, rapidly decreasing from an initial fitness value of 78.5 to 15.3, showing significantly better convergence performance than the other two methods. The method of this invention performs best, not only with a relatively low initial fitness value of 68.2, but more importantly, exhibiting extremely strong convergence ability, rapidly decreasing to 21.3 in the first 60 iterations and finally stabilizing at an extremely low level of 3.8. Compared to traditional methods, this invention achieves significant improvements in both optimization efficiency and final convergence accuracy. These results fully validate the technical advantages of this invention, which leverages dynamic adjustment of pheromone release and volatility coefficients combined with historical process data fluctuation patterns. This allows for a faster and more accurate identification of the optimal solution for load control in ammonia synthesis units, providing reliable theoretical support for practical engineering applications.

[0103] In one optional implementation, the optimal load control command is converted into control parameters, the control parameters are decomposed into target values ​​of process parameters, and intelligent control of the ammonia synthesis production load is achieved through step-by-step adjustment using a cascade PID controller.

[0104] The optimal load control command is converted into ammonia synthesis process control parameters. The time-series characteristics of the process parameters are obtained based on the response relationship between parameters in the historical operating data of the process parameters. Based on the time-series characteristics, the ammonia synthesis process control parameters are decomposed into process parameter target values.

[0105] The target value of the process parameter is used as the setpoint of the main loop PID controller. The deviation sequence between the actual output value of the main loop PID controller and the target value of the process parameter is calculated. The deviation sequence is weighted by time and integrated to obtain the integral time absolute error index. The adjustment direction and adjustment step size of the main loop PID controller parameters are calculated using the integral time absolute error index. The proportional coefficient, integral coefficient and derivative coefficient of the main loop are updated. The updated output value of the main loop PID controller is set as the setpoint of the slave loop PID controller.

[0106] The actual adjustment amount of the process parameters is obtained by adjusting them step by step through a cascade PID controller and then sent to the corresponding actuators to realize intelligent load control of the ammonia synthesis unit.

[0107] For ammonia synthesis, optimal load control commands are typically expressed as total output targets or unit load percentages, such as "1500 tons of ammonia per day" or "unit operating at 85% load." To translate these commands into operable control parameters, a mapping relationship between unit load and key process control parameters must first be established. In one implementation, historical data of the ammonia synthesis unit operating under different loads is collected to extract the correspondence between key process control parameters such as the ammonia synthesis tower inlet gas flow rate, the hydrogen-to-nitrogen ratio in the circulating gas, operating pressure, and operating temperature, and the unit load. For example, when the unit load is 85%, the corresponding control parameter is: ammonia synthesis tower inlet gas flow rate of 90000 Nm³. 3 / h, hydrogen-nitrogen ratio in circulating gas 2.9:1, operating pressure 25MPa, operating temperature 460℃.

[0108] After obtaining the control parameters for the ammonia synthesis process, it is necessary to further decompose them into specific target values ​​for the process parameters. This step relies on the temporal characteristics of the process parameters. Temporal characteristics refer to the interrelationships between process parameters and their changes over time, which can be obtained by analyzing historical operating data. In practice, the sliding time window method can be used to slice historical data and calculate the delay correlation and response time between parameters. For example, the analysis results show that after the inlet gas flow rate of the ammonia synthesis tower changes, the hydrogen-nitrogen ratio in the circulating gas begins to respond in about 5 minutes and reaches stability in 15 minutes; after the operating pressure is adjusted, the operating temperature begins to change in about 10 minutes and reaches a new thermal equilibrium state in 30 minutes. Based on these temporal characteristics, the control parameters for the ammonia synthesis process can be decomposed into a cascaded sequence of target values ​​for process parameters: the first stage includes setting the feed gas flow rate to 92000 Nm³. 3 / h, fresh hydrogen replenishment rate set at 15000 Nm 3 / h, fresh nitrogen replenishment rate set to 5200 Nm 3 / h; The second stage includes setting the synthesis tower inlet temperature to 400℃ and the synthesis tower inlet pressure to 25MPa.

[0109] Obtain the basic deviation sequence of the main loop PID controller, which is the difference between the target temperature of 480℃ and the real-time feedback value of the temperature sensor. Multiply the basic deviation sequence by the proportional coefficient 2.5 to obtain the proportional term output. Divide the basic deviation sequence by the integral time constant 120s and integrate to obtain the integral term output. Multiply the rate of change of the basic deviation sequence by the derivative time constant 30s and multiply by the proportional coefficient to obtain the derivative term output. Superimpose the proportional term output, integral term output, and derivative term output to obtain the total output value of the main loop PID controller. Calculate the absolute difference between the total output value and the target temperature of 480℃ to obtain the deviation sequence.

[0110] The control effect is evaluated using the integral time absolute error index. Specifically, the deviation sequence between the actual output value of the main loop PID controller and the target value of the process parameters is first obtained. For example, in a certain adjustment, the deviation sequence between the recorded tower inlet temperature and the set 400℃ is [+5.2,+4.8,+3.5,+2.1,+1.2,+0.5,+0.1,-0.2,-0.3,-0.1,0]℃. Then, the deviation sequence is time-weighted, with the deviation closer to the current time having a larger weight. For example, using an exponentially decreasing weighting method, the weight sequence is [0.2,0.22,0.24,0.27,0.3,0.33,0.36,0.4,0.44,0.48,0.53]. The weighted deviation sequences are summed to obtain the integral time absolute error index value of 5.87℃·min.

[0111] When the index value exceeds the preset threshold (e.g., 5℃·min), it indicates that the control effect is not ideal and the PID parameters need to be adjusted. The adjustment direction is determined by the oscillation characteristics and convergence speed of the deviation sequence: if the deviation sequence exhibits oscillation and the amplitude increases, decrease the proportional coefficient; if the deviation sequence converges too slowly, increase the proportional coefficient and integral coefficient; if the deviation changes too drastically, increase the derivative coefficient. The adjustment step size is proportional to the integral time absolute error index. For example, the adjustment step size of the proportional coefficient can be set to the current value ±(5% × integral time absolute error index / preset threshold). Taking the initial PID parameters P=2.5, I=0.3, D=0.05 as an example, after the above evaluation, they are updated to P=2.625, I=0.315, D=0.05.

[0112] The slave loop PID controller directly controls the actuators, such as adjusting valve opening and motor speed. For example, if the updated main loop output is 65% of the heating power, this value is set as the setpoint for the slave loop controller. The slave loop controller calculates and outputs specific control signals to the heater based on the deviation between the current actual power of the heater and the setpoint.

[0113] The cascaded PID control system continuously performs the above process according to a set sampling period (e.g., 30 seconds for the main loop and 5 seconds for the slave loop), achieving smooth load transition and precise control of the ammonia synthesis unit. In production practice, this method reduces the unit load adjustment time from the traditional 60 minutes to 35 minutes, reduces ammonia production fluctuations by 40% during the transition process, and lowers energy consumption by 3.5%, significantly improving the intelligence level and economic efficiency of ammonia synthesis production.

[0114] In one optional implementation, the actual adjustment amount of the process parameters is obtained by adjusting them step by step through a cascade PID controller and then sent to the corresponding actuators to achieve intelligent load control of the ammonia synthesis unit, including:

[0115] Extract dynamic response data between process parameters, statistically analyze the fluctuation transmission relationship and response timing characteristics between process parameters, generate coupling coefficient matrix and timing adjustment coefficient of process parameters, superimpose the output value of the loop PID controller with the compensation amount of other loops calculated based on the coupling coefficient matrix to obtain the initial control quantity, calculate the execution timing delay of each loop according to the timing adjustment coefficient, and adjust the initial control quantity in stages according to the execution timing delay;

[0116] The response characteristic curves of the actuator under different input signals are collected, and the gain characteristic and delay characteristic in the response characteristic curve are extracted. The feedforward compensation gain is calculated based on the gain characteristic, and the feedforward derivative time is determined based on the delay characteristic. The feedforward compensation gain and the feedforward derivative time are applied to the initial control quantity after graded adjustment to obtain the compensated execution command.

[0117] The compensated execution command is sent to the actuator and the actual output value of the actuator is collected. The execution deviation sequence between the compensated execution command and the actual output value is calculated. The dynamic correction coefficient is determined according to the changing trend of the execution deviation sequence. The execution command corrected based on the dynamic correction coefficient is resent to the actuator to realize intelligent load control of the ammonia synthesis unit.

[0118] First, data acquisition and analysis were conducted on the dynamic relationships between process parameters in the ammonia synthesis unit. The control system continuously collected real-time data on process parameters such as the synthesis tower inlet temperature, circulating gas volume, and compressor load under different operating conditions, with a sampling frequency of once per second for 48 consecutive hours. Based on the collected data, correlation coefficients between various parameters were calculated. For example, the correlation coefficient between the synthesis tower inlet temperature and the circulating gas volume was 0.87, and the correlation coefficient with the compressor load was 0.92. By calculating the degree of mutual influence between different parameters, a 12×12 coupling coefficient matrix was formed, where the coupling coefficient between the circulating gas volume and the synthesis tower inlet temperature was 0.35, and vice versa, it was 0.29. Simultaneously, the temporal characteristics of parameter fluctuation transmission were measured. For example, after a change in compressor load, the response lag time for the synthesis tower inlet temperature was 45 seconds, and the response lag time for the circulating gas volume was 32 seconds. These data constituted a temporal adjustment coefficient table.

[0119] During the control process, when the temperature PID controller outputs 5%, the compensation amounts for other loops are calculated based on the coupling coefficient matrix. For example, the compensation amount for the circulating air volume control loop is 5% × 0.35 = 1.75%, and the compensation amount for the compressor load control loop is 5% × 0.42 = 2.1%. These compensation amounts are then superimposed with the original output values ​​of their respective PID controllers to obtain the initial control values ​​for each loop. For the temperature loop, the initial control value is 5% + 1.2% = 6.2% (1.2% comes from compensation for the influence of other loops).

[0120] The temperature control loop executes directly without delay, the circulating gas volume control loop executes with a 32-second delay, and the compressor load control loop executes with a 45-second delay. Based on these delay times, the initial control values ​​are adjusted in stages to ensure that each loop adjusts in a reasonable timing sequence, avoiding system fluctuations caused by the interaction of parameters.

[0121] For actuator characteristic compensation, the response characteristics of actuators (such as control valves and frequency converters) under different input signals are collected experimentally. For major control valves, the input signal is gradually increased from 10% to 90% in 10% increments, and the actual valve opening change is recorded. By analyzing the response curves, gain and delay characteristics are extracted. For example, a certain control valve has an actual opening of 47% when the input signal is 50%, and a gain characteristic of 0.94; after the input signal changes, the valve's actual response delay is 2.5 seconds, which constitutes the delay characteristic.

[0122] For the aforementioned regulating valve, the feedforward compensation gain is set to 1 / 0.94 = 1.06, and the feedforward derivative time is set to 2.5 seconds. When the initial control quantity after graded adjustment is 6.2%, the execution command after applying feedforward compensation is 6.2% × 1.06 = 6.57%. Simultaneously, based on the feedforward derivative time, the execution command is sent 2.5 seconds in advance to ensure the timeliness of the control action.

[0123] When the issued instruction is 6.57%, the actual output of the actuator is 6.3%. The execution deviation is calculated as 6.57% - 6.3% = 0.27%, and the execution deviations for 10 consecutive times are recorded to form a deviation sequence. The trend of the deviation sequence is analyzed. If the deviation shows a gradually increasing trend, the dynamic correction coefficient is determined to be 1.02; if the deviation shows a gradually decreasing trend, the dynamic correction coefficient is determined to be 0.98.

[0124] The execution command is corrected by applying a dynamic correction coefficient. For example, for a trend of increasing deviation, the corrected execution command is 6.57% × 1.02 = 6.7%. The corrected execution command is then reissued to the actuator, and the actual output value is increased to 6.55%, reducing the execution deviation to 0.15%, which is lower than the set allowable deviation threshold of 0.2%, thus achieving accurate execution of the execution command.

[0125] Using the cascaded PID control method described above, the ammonia synthesis unit can respond smoothly to load changes. Practical application shows that after adopting this method, the unit load adjustment time was shortened from the traditional 15 minutes to 8 minutes, the fluctuation range of key parameters was reduced by 40%, production energy consumption was reduced by 3.2%, and product quality stability was improved by 5.7%, achieving intelligent and efficient load control of the ammonia synthesis unit.

[0126] This invention relates to an intelligent control system for synthetic ammonia production load based on wind and solar forecasting. The system includes:

[0127] The first unit is used to collect data from wind farms and photovoltaic power stations as a meteorological sample set, and at the same time, to collect data from ammonia synthesis units as a process sample set.

[0128] The second unit is used to process the meteorological sample set and the process sample set respectively using a dual-flow variational autoencoder, extract the temporal features of meteorological data and process data through the coding network, generate latent features by combining the residual connection mechanism, and generate the wind and solar power generation prediction curve and the process parameter change trend of the ammonia synthesis unit for future periods based on the latent features.

[0129] The third unit is used to calculate the upper and lower limits of process parameters and the safety boundary of operating conditions of the ammonia synthesis unit at each time point based on the wind and solar power generation prediction curve and the process parameter change trend; determine the coupling relationship between process parameters based on historical operating data and establish a safety constraint model for process parameters; and combine the upper and lower limits of process parameters and the safety constraint model for process parameters to obtain the load adjustment space of the ammonia synthesis unit under different wind and solar power output conditions.

[0130] The fourth unit is used to generate an initial solution set for load allocation in each time period based on the load adjustment space and the load optimization algorithm, taking the fluctuation pattern of historical process data as the optimization constraint, and calculating the fitness value. The release amount and volatility coefficient of pheromones are dynamically adjusted according to the fitness value. When the change value of the optimal solution in continuous iteration is less than the preset convergence threshold, the unit outputs the time-sharing optimal load control command of the ammonia synthesis unit.

[0131] The fifth unit is used to convert the optimal load control command into control parameters, decompose the control parameters into process parameter target values, and realize intelligent control of the synthetic ammonia production load through step-by-step adjustment by a cascade PID controller.

[0132] A third aspect of the present invention provides an electronic device, comprising:

[0133] processor;

[0134] Memory used to store processor-executable instructions;

[0135] The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0136] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0137] This invention can be a method, apparatus, system, and / or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of the invention.

[0138] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; 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 or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent control of ammonia synthesis production load based on wind and solar forecasting, characterized in that, include: Data from wind farms and photovoltaic power plants were collected as a meteorological sample set, while data from ammonia synthesis units were collected as a process sample set. The meteorological sample set and the process sample set are processed by a dual-flow variational autoencoder, and the temporal features of the meteorological data and process data are extracted by the coding network. The latent features are generated by combining the residual connection mechanism. Based on the latent features, the wind and solar power generation prediction curves and the change trends of process parameters of the ammonia synthesis unit are generated for future periods. Based on the wind and solar power generation prediction curves and the changing trends of the process parameters, the upper and lower limits of the process parameters and the safety boundary of the operating conditions for the ammonia synthesis unit at each time point are calculated. The coupling relationship between the process parameters is determined based on historical operating data, and a safety constraint model for the process parameters is established. Combining the upper and lower limits of the process parameters and the safety constraint model, the load adjustment space of the ammonia synthesis unit under different wind and solar power output conditions is obtained, including: Based on the predicted value of the wind and solar power generation prediction curve at the current moment, determine the upper and lower limits of the raw gas pressure, the upper and lower limits of the circulating gas pressure, and the upper and lower limits of the synthesis tower temperature, and obtain the upper and lower limits of the process parameters and the safety boundary of the operating conditions. Extract steady-state operating data of feed gas pressure and circulating gas pressure from historical operating data, calculate the linear correlation coefficient between feed gas pressure and circulating gas pressure, extract dynamic operating data, determine the response time difference and response amplitude ratio between synthesis tower temperature and feed gas pressure and circulating gas pressure, and input the linear correlation coefficient, response time difference and response amplitude ratio into the safety constraint evaluation function to generate a process parameter safety constraint model. Substitute the upper and lower limits of the process parameters and the safety boundary of the operating conditions into the safety constraint model of the process parameters, calculate the steady-state constraint range of the feed gas pressure and the circulating gas pressure, and the dynamic response constraint range of the synthesis tower temperature, respectively. Take the intersection of the steady-state constraint range and the dynamic response constraint range to obtain the load adjustment space of the ammonia synthesis unit under different wind and solar power output conditions. Based on the aforementioned load adjustment space, and combined with the load optimization algorithm, the fluctuation patterns of historical process data are used as optimization constraints to generate an initial solution set for load allocation in each time period and calculate the fitness value. The release amount and volatility coefficient of pheromones are dynamically adjusted according to the fitness value. When the change value of the optimal solution in continuous iterations is less than a preset convergence threshold, a time-segmented optimal load control command for the ammonia synthesis unit is output, including: Extract the continuous load change sequence from historical process data, and statistically analyze the distribution intervals of the absolute value of load change and the absolute value of load change rate between two adjacent time points. Set the upper limit of the distribution interval as the upper limit of load change amount and the upper limit of load change rate, respectively, to obtain the load optimization constraint boundary. The load optimization algorithm is constructed based on the ant colony algorithm. Load solutions that meet the ant colony size requirements are randomly generated within the load adjustment space. It is determined whether the load change and load change rate at adjacent time points in each load solution are within the load optimization constraint boundary. Load solutions that meet the judgment conditions are retained as the initial load solution set. The load target deviation and load fluctuation amplitude of each load solution are calculated. The load target deviation and load fluctuation amplitude are assigned corresponding weights respectively. The result of the weighted sum is used as the fitness value. The baseline pheromone release is determined based on the number of iterations completed. The ratio of the optimal fitness value to the average fitness value of the current iteration is multiplied by the pheromone concentration decay coefficient to obtain the pheromone concentration adjustment factor. The baseline release is adjusted based on the pheromone concentration adjustment factor to obtain the actual pheromone release. The actual pheromone release is then updated to the pheromone concentration matrix. A new load solution set is generated according to the numerical distribution of the pheromone concentration matrix. The fitness value is repeatedly calculated and the pheromone concentration matrix is ​​updated. When the continuous iteration reaches a predetermined number of times and the change in the optimal fitness value is less than the preset convergence threshold, the corresponding load solution is output as the time-sharing optimal load control command of the ammonia synthesis unit. The optimal load control command is converted into control parameters, which are then decomposed into target values ​​of process parameters. These parameters are then adjusted step by step using a cascade PID controller to achieve intelligent control of the ammonia synthesis production load.

2. The method according to claim 1, characterized in that, The meteorological sample set and the process sample set are processed separately using a dual-flow variational autoencoder. Temporal features of the meteorological and process data are extracted through an encoding network. Latent features are generated using a residual connection mechanism. Based on these latent features, a future wind and solar power generation prediction curve and the trend of process parameters in the ammonia synthesis unit are generated, including: The meteorological sample set is used as the first data stream, and the process sample set is used as the second data stream. The first data stream and the second data stream are respectively input into the first variational autoencoder and the second variational autoencoder. Residual connection paths are established between the encoder and decoder of the first variational autoencoder and the second variational autoencoder, respectively. The residual connection paths add the original features output from the previous layer to the transformed features output from the current layer to generate the first latent feature and the second latent feature. Calculate the temporal cross-correlation coefficient between the first latent feature and the second latent feature, determine the feature fusion weight based on the temporal cross-correlation coefficient, and sum the first latent feature and the second latent feature according to their corresponding weights to obtain the fused feature; The fused features and historical data are combined to form a sliding time window. Based on the data sequence in the sliding time window, the power generation curves of wind farms, power generation curves of photovoltaic power plants, and the trend of process parameter changes of ammonia synthesis units for future periods are generated.

3. The method according to claim 1, characterized in that, The optimal load control command is converted into control parameters, which are then decomposed into target values ​​for process parameters. Intelligent load control for ammonia synthesis production is achieved through step-by-step adjustment using a cascaded PID controller. The optimal load control command is converted into ammonia synthesis process control parameters. The time-series characteristics of the process parameters are obtained based on the response relationship between parameters in the historical operating data of the process parameters. Based on the time-series characteristics, the ammonia synthesis process control parameters are decomposed into process parameter target values. The target value of the process parameter is used as the setpoint of the main loop PID controller. The deviation sequence between the actual output value of the main loop PID controller and the target value of the process parameter is calculated. The deviation sequence is weighted by time and integrated to obtain the integral time absolute error index. The adjustment direction and adjustment step size of the main loop PID controller parameters are calculated using the integral time absolute error index. The proportional coefficient, integral coefficient and derivative coefficient of the main loop are updated. The updated output value of the main loop PID controller is set as the setpoint of the slave loop PID controller. The actual adjustment amount of the process parameters is obtained by adjusting them step by step through a cascade PID controller and then sent to the corresponding actuators to realize intelligent load control of the ammonia synthesis unit.

4. The method according to claim 3, characterized in that, Intelligent load control of the ammonia synthesis unit is achieved by using a cascaded PID controller to adjust process parameters step by step, obtaining the actual adjustment values, and then sending them to the corresponding actuators. Extract dynamic response data between process parameters, statistically analyze the fluctuation transmission relationship and response timing characteristics between process parameters, generate coupling coefficient matrix and timing adjustment coefficient of process parameters, superimpose the output value of the loop PID controller with the compensation amount of other loops calculated based on the coupling coefficient matrix to obtain the initial control quantity, calculate the execution timing delay of each loop according to the timing adjustment coefficient, and adjust the initial control quantity in stages according to the execution timing delay; The response characteristic curves of the actuator under different input signals are collected, and the gain characteristic and delay characteristic in the response characteristic curve are extracted. The feedforward compensation gain is calculated based on the gain characteristic, and the feedforward derivative time is determined based on the delay characteristic. The feedforward compensation gain and the feedforward derivative time are applied to the initial control quantity after graded adjustment to obtain the compensated execution command. The compensated execution command is sent to the actuator and the actual output value of the actuator is collected. The execution deviation sequence between the compensated execution command and the actual output value is calculated. The dynamic correction coefficient is determined according to the changing trend of the execution deviation sequence. The execution command corrected based on the dynamic correction coefficient is resent to the actuator to realize intelligent load control of the ammonia synthesis unit.

5. A smart control system for synthetic ammonia production load based on wind and solar forecasting, used to implement the method as described in any one of claims 1-4, characterized in that, include: The first unit is used to collect data from wind farms and photovoltaic power stations as a meteorological sample set, and at the same time, to collect data from ammonia synthesis units as a process sample set. The second unit is used to process the meteorological sample set and the process sample set respectively using a dual-flow variational autoencoder, extract the temporal features of meteorological data and process data through the coding network, generate latent features by combining the residual connection mechanism, and generate the wind and solar power generation prediction curve and the process parameter change trend of the ammonia synthesis unit for future periods based on the latent features. The third unit is used to calculate the upper and lower limits of the process parameters and the safety boundary of the operating conditions of the ammonia synthesis unit at each time point based on the wind and solar power generation prediction curve and the process parameter change trend. Based on historical operating data, the coupling relationship between process parameters is determined, and a safety constraint model for process parameters is established. The upper and lower limit constraints of the process parameters and the safety constraint model for process parameters are combined to obtain the load adjustment space of the ammonia synthesis unit under different wind and solar power output conditions. The fourth unit is used to generate an initial solution set for load allocation in each time period based on the load adjustment space and the load optimization algorithm, taking the fluctuation pattern of historical process data as the optimization constraint, and calculating the fitness value. The release amount and volatility coefficient of pheromones are dynamically adjusted according to the fitness value. When the change value of the optimal solution in continuous iteration is less than the preset convergence threshold, the unit outputs the time-sharing optimal load control command of the ammonia synthesis unit. The fifth unit is used to convert the optimal load control command into control parameters, decompose the control parameters into process parameter target values, and realize intelligent control of the synthetic ammonia production load through step-by-step adjustment by a cascade PID controller.

6. An electronic device, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the method according to any one of claims 1 to 4.

7. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the method described in any one of claims 1 to 4.

Images