Ventilation energy-saving control method and system for marine engine room

Abstract

The invention provides a ventilation energy-saving control method and system for a marine engine room, and relates to the technical field of intelligent control, and the method comprises the steps: dividing the marine engine room into a plurality of sub-regions in an equal proportion, obtaining the multi-modal data of the marine engine room in the operation process, and enabling each sub-region to correspond to a ventilation pipeline; according to the multi-modal data, calculating a comprehensive comfort index of each sub-region; according to the comprehensive comfort index, the target tail end air volume of each ventilation pipeline corresponding to each sub-area is calculated; constructing a data driving model; with the minimum fan energy consumption as the target, all the target tail end air volumes are input into the data driving model, and the optimal air door angles of all the ventilation pipelines are output; and the current air door angle of each ventilation pipeline is controlled to the corresponding optimal air door angle, and ventilation energy-saving control over the marine engine room is completed.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent control, and in particular to a ship engine room ventilation energy-saving control method and system. BACKGROUND

[0002] With the continuous enhancement of the large-scale and high-power trend of ship power systems, the heat sources are concentrated in the engine room, the number of equipment is increasing, and the energy consumption is continuously rising, which leads to a significant increase in the difficulty of engine room environment control. As an important infrastructure for ensuring personnel comfort, maintaining air quality and ensuring safe operation of equipment, the operation energy consumption of the engine room ventilation system accounts for a large proportion of the whole ship energy consumption. The superposition of high temperature, high humidity, noise, air pollution and environmental load brought by personnel activities makes the engine room ventilation system must continuously provide sufficient fresh air and cooling airflow under complex working conditions to maintain a safe, comfortable and efficient operating environment. However, in the case of significant differences in environmental requirements in different areas and frequent dynamic changes in equipment heat dissipation, how to effectively reduce the ventilation energy consumption while ensuring the environmental quality has become an important research topic in the field of ship energy-saving control.

[0003] In the prior art, the traditional ship engine room ventilation energy-saving control method mainly relies on fixed air volume, empirical rough adjustment or single variable control strategy. Such methods usually regard the engine room as a whole environment, lacking fine perception of the environmental requirements of different areas.

[0004] In addition, the traditional ship engine room ventilation energy-saving control method generally lacks fusion analysis ability and intelligent optimization mechanism for real-time environmental data, so that the air volume distribution cannot be dynamically adjusted according to the regional requirements, resulting in excessive air supply, energy waste or local ventilation deficiency. SUMMARY

[0005] In view of the above deficiencies of the prior art, the purpose of the embodiments of the present application is to provide a ship engine room ventilation energy-saving control method, which can solve the technical problems that the traditional ship engine room ventilation energy-saving control method mainly relies on fixed air volume, empirical rough adjustment or single variable control strategy, usually regards the engine room as a whole environment, lacks fine perception of the environmental requirements of different areas, and generally lacks fusion analysis ability and intelligent optimization mechanism for real-time environmental data, so that the air volume distribution cannot be dynamically adjusted according to the regional requirements, resulting in excessive air supply, energy waste or local ventilation deficiency.

[0006] The first aspect of the embodiments of the present application provides a ship engine room ventilation energy-saving control method, comprising: S1: proportionally dividing a ship engine room into a plurality of sub-regions, and acquiring multi-modal data of the ship engine room in the running process, each sub-region corresponding to a ventilation pipeline; S2: calculating a comprehensive comfort index of each sub-region according to the multi-modal data; S3: Calculate the target terminal air volume of each ventilation duct corresponding to each sub-area based on the comprehensive comfort index; S4: Building a data-driven model; S5: With the goal of minimizing fan energy consumption, input the air volume of each target terminal into the data-driven model and output the optimal damper angle of each ventilation duct. S6: Control the current damper angle of each ventilation duct to the corresponding optimal damper angle to complete the energy-saving ventilation control of the ship's engine room.

[0007] A second aspect of the present invention provides a ship engine room ventilation energy-saving control system, comprising: a processor and a memory; The memory stores programs or instructions that can run on the processor, which, when executed by the processor, implement the steps of the energy-saving control method for ship engine room ventilation as described in the first aspect.

[0008] A third aspect of the present invention provides a readable storage medium on which a program or instructions are stored, which, when executed by a processor, implement the steps of the energy-saving control method for ship engine room ventilation as described in the first aspect.

[0009] The beneficial effects of the technical solutions provided by the embodiments of the present invention include at least the following: In this embodiment of the invention, the ship's engine room is divided into multiple sub-regions proportionally, and multimodal data of the ship's engine room during operation is acquired. Each sub-region corresponds to a ventilation duct. Based on the multimodal data, the comprehensive comfort index of each sub-region is calculated, and based on the comprehensive comfort index, the target terminal air volume of each ventilation duct corresponding to each sub-region is calculated. This eliminates reliance on fixed air volume, coarse adjustment based on experience, or single-variable control strategies. By dividing the engine room into multiple sub-regions, it has a refined perception of the environmental needs of different regions. By constructing a data-driven model and aiming to minimize fan energy consumption, the target terminal air volume of each region is input into the data-driven model, and the optimal damper angle of each ventilation duct is output. It has the ability to fuse and analyze real-time environmental data and has an intelligent optimization mechanism, so that the air volume distribution can be dynamically adjusted according to the regional needs, without causing excessive air supply, energy waste, or insufficient local ventilation. Attached Figure Description

[0010] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Throughout the drawings, the same reference numerals denote the same parts. Obviously, the drawings described below are merely some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without any creative effort.

[0011] Figure 1 This is a schematic flowchart of a ship engine room ventilation energy-saving control method provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of a ship engine room ventilation energy-saving control system provided in an embodiment of the present invention. Detailed Implementation

[0012] To enable those skilled in the art to better understand the technical solutions in the embodiments of the present invention, the technical solutions 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, not all embodiments. It should be understood that these descriptions are merely exemplary and are not intended to limit the scope of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0013] The energy-saving control method for ship engine room ventilation provided by the present invention will be described in detail below with reference to the accompanying drawings, through specific embodiments and application scenarios.

[0014] Reference manual attached Figure 1 The diagram shows a flowchart of a ship engine room ventilation energy-saving control method provided by an embodiment of the present invention.

[0015] This invention provides a method for energy-saving control of ship engine room ventilation, which may include the following steps: S1: Divide the ship's engine room into multiple sub-regions proportionally and acquire multimodal data of the ship's engine room during operation. Each sub-region corresponds to a ventilation duct.

[0016] It should be noted that each sub-region corresponds to one ventilation duct, and the number of sub-regions is the same as the number of ventilation ducts. Each sub-region and its corresponding ventilation duct are represented by parameter j, and the number of sub-regions and the number of ventilation ducts are represented by parameter M.

[0017] Optionally, multimodal data specifically includes environmental data and operating condition data.

[0018] The environmental data specifically includes: the current temperature, current humidity, current noise level, current carbon dioxide concentration, and current number of workers in each sub-area.

[0019] The specific operating data includes: current global fan pressure, current fan speed, number of operating fans, and current terminal air volume, current terminal pressure drop, current damper pressure drop, and current damper angle for each ventilation duct.

[0020] It should be noted that environmental data is used to characterize the real-time environmental conditions of each sub-area within the cabin, primarily reflecting personnel comfort and air quality. This includes current temperature to measure thermal environment levels, current humidity to characterize air humidity, current noise levels to reflect the noise intensity generated by mechanical equipment operation, current carbon dioxide concentration to assess air freshness, and current number of personnel to describe the impact of personnel density on environmental requirements. Operating data describes the workload and ventilation behavior of the cabin ventilation and power systems under current operating conditions. This includes current global fan pressure to reflect the pressure provided by the fans to overcome duct resistance, current fan speed to indicate fan output capacity, number of operating fans to characterize the total number of fans currently supplying air, current terminal airflow of each ventilation duct to represent the actual airflow delivered to each area, current terminal pressure drop to reflect the resistance at the end of the duct, current damper pressure drop to represent the pressure loss caused by damper throttling, and current damper angle to describe the degree of damper opening and directly affect the ventilation volume regulation effect.

[0021] In this embodiment of the invention, by dividing the cabin into multiple sub-regions and collecting multimodal data including environmental and operating conditions, the environmental differences and ventilation status of each region can be precisely characterized, transforming ventilation control from overall control to zoned, on-demand control. Regionalized data acquisition not only accurately reflects personnel comfort needs but also simultaneously monitors equipment operating load and pipeline resistance changes, providing high-quality input for subsequent comprehensive comfort index calculations, target terminal airflow allocation, and data-driven model predictions. This achieves precise air delivery, reduces unnecessary energy consumption, and enhances the system's adaptability to changes in operating conditions, laying a reliable foundation for fan energy consumption optimization.

[0022] In one possible implementation, after S1 and before S2, step S1A is further included: S1A: Perform data standardization processing on environmental data. Among them, Z T T represents the current temperature of the sub-region after data standardization. ref The recommended temperature for the sub-region is represented by ΔT. max Z represents the maximum allowable temperature deviation of the sub-region. RH RH represents the current humidity of the sub-region after data standardization. ref Indicates the recommended humidity for the sub-region, ΔRH max Z represents the maximum allowable humidity deviation for a sub-region. L L represents the current noise level of the sub-region after data standardization. refIndicates the recommended noise level for the sub-region. This represents the current carbon dioxide concentration of the sub-region after data standardization, where CO2 represents the current carbon dioxide concentration of the sub-region. This indicates the recommended carbon dioxide concentration for the sub-region.

[0023] In this embodiment of the invention, by standardizing environmental data, data with different dimensions and magnitudes, such as temperature, humidity, noise, and carbon dioxide concentration, can be uniformly converted into comparable dimensionless deviation values. This avoids the bias caused by differences in the scale of the original data in the calculation of the comfort index and model training. The standardized data can more accurately reflect the degree of deviation of each environmental indicator from the recommended value, making the calculation of the comfort index more sensitive and reliable, while improving the rationality of weight allocation and the stability of ventilation demand analysis. In addition, this processing method enhances the adaptability of the model under different ship operating conditions, improves the training efficiency and generalization ability of the data-driven model, and provides a stable and accurate input foundation for subsequent airflow optimization and energy-saving control.

[0024] S2: Calculate the comprehensive comfort index for each sub-region based on multimodal data.

[0025] It should be noted that the comprehensive comfort index is used to uniformly and quantitatively evaluate the environmental conditions of various sub-areas within the cabin. Its core function is to integrate multiple heterogeneous environmental indicators such as temperature, humidity, noise, and air quality into a single metric that can be directly used for ventilation control. This index first calculates the corresponding comfort indices for standardized temperature, humidity, noise, and carbon dioxide concentration, reflecting the degree of impact of each environmental factor on personnel work experience and equipment operational safety. Then, combined with the personnel density ratio, the weights of different environmental factors are dynamically determined, ensuring that thermal and humidity comfort is prioritized in densely populated areas, while noise and air quality are emphasized in sparsely populated or unoccupied areas. Simultaneously, the temperature and humidity influence coefficient of the equipment is introduced to ensure that equipment safety requirements are not ignored. After obtaining the initial comprehensive comfort index through weighted fusion, a nonlinear mapping is used to enhance the index's sensitivity to environmental degradation, thus forming the final comprehensive comfort index. This provides a quantitative basis for airflow distribution and energy-saving control, enabling the ventilation system to be adjusted differentiated according to the environmental needs of different areas.

[0026] In one possible implementation, the calculation of the comprehensive comfort index specifically includes sub-steps S201 to S205: S201: Based on the standardized data of current temperature, humidity, noise level, and carbon dioxide concentration, calculate the temperature comfort index, humidity comfort index, noise comfort index, and air quality comfort index for each sub-region. Among them, IT This represents the temperature comfort index, where e represents the natural base, | represents the absolute value, and I RH I represents the humidity comfort index. L This represents the noise comfort index, where 'max' indicates the maximum value. This indicates the air quality comfort index.

[0027] It should be noted that the temperature comfort index, humidity comfort index, noise comfort index, and air quality comfort index are used to quantify the impact of various environmental factors on the working experience and health of cabin personnel. The core idea is to convert standardized physical quantities such as temperature, humidity, noise, and carbon dioxide concentration into "comfort levels," allowing environmental indicators with different dimensions and properties to be compared and integrated on a unified scale. Specifically, the temperature comfort index and humidity comfort index use exponential decay functions to evaluate the degree of deviation from the recommended temperature and humidity range; the greater the deviation, the lower the index. The noise comfort index and air quality comfort index employ a suppressive mapping for positive deviations, focusing on reflecting the negative impact of increased noise or deteriorating air quality on comfort. Through these four types of indices, complex and diverse environmental quality factors can be transformed into a unified dimensionless comfort measure, providing a fundamental quantitative basis for subsequent weight allocation and comprehensive comfort index calculation.

[0028] S202: Calculate the personnel density ratio of the sub-region based on the current number of workers in the sub-region: Where, r p N represents the ratio of population density in a sub-region. p N represents the current number of workers in the sub-region. max This indicates the maximum number of workers in the sub-region.

[0029] It should be noted that the personnel density ratio describes the relative level of personnel occupancy in a sub-area of ​​the cabin. It is defined as the ratio of the current number of personnel to the maximum capacity of that area, thus characterizing the intensity of environmental demands on personnel activity. A high personnel density ratio indicates that someone is working in that area, and the demand for a warm, humid environment and fresh air is significantly increased. A low personnel density ratio, or even zero, indicates that there is no personnel activity in that area, and environmental control will focus more on noise control, maintaining air quality, and ensuring the safe operation of equipment. The personnel density ratio not only reflects changes in people's subjective environmental needs but also dynamically adjusts the weight of the comfort index in the overall calculation, enabling ventilation control strategies to be differentiated and refined according to changes in personnel distribution.

[0030] S203: Calculate the weighting coefficients for the temperature comfort index, humidity comfort index, noise comfort index, and air quality comfort index based on the personnel density ratio. Among them, w i Let k represent the weighting coefficient of the i-th comfort index. i k represents the initial weighting coefficient of the i-th comfort index under the influence of the population density ratio in the sub-region. T γ represents the initial weighting coefficient of the temperature comfort index under the influence of the population density ratio in a sub-region. T k represents the temperature influence coefficient of the operating equipment in the sub-region. RH γ represents the initial weighting coefficient of the humidity comfort index under the influence of the population density ratio in the sub-region. RH k represents the humidity influence coefficient of the equipment operating in the sub-region. L This represents the initial weighting coefficient of the noise comfort index under the influence of the population density ratio in the sub-region. This represents the initial weighting coefficient of the air quality comfort index under the influence of the population density ratio in the sub-region.

[0031] It should be noted that formula k T 0.5, k RH 0.3, k L 0.4 and The values ​​of 0.3 represent comfort impact factors. Based on cabin operation experience and relevant comfort evaluation standards, it is believed that when personnel density is relatively high, temperature has a greater impact on comfort than humidity. Therefore, the temperature comfort impact factor is set to 0.5, and the humidity comfort impact factor is set to 0.3. When personnel density is relatively low, noise and air quality have a more significant impact on environmental safety and equipment operating status. Therefore, the noise comfort impact factor and air quality comfort impact factor are set to 0.4 and 0.3, respectively.

[0032] Furthermore, the ship's engine room is filled with various precision equipment and power systems, which have specific requirements for the temperature and humidity of their operating environment: excessively high temperatures can lead to overheating, performance degradation, or even damage to the equipment, while excessively high humidity can cause safety hazards such as decreased electrical insulation and metal corrosion. Therefore, the temperature influence coefficient γ introduced into the equipment is crucial. T Humidity influence coefficient γ RHThis is precisely to compensate for this design flaw. These two coefficients ensure that in unmanned areas, the system can still maintain a minimum temperature and humidity control weight based on equipment protection needs, thereby expanding the goal of ventilation control from simply "personnel comfort" to a comprehensive environmental protection that takes into account both "personnel comfort" and "equipment safety," thus improving the engineering practicality and reliability of the method.

[0033] S204: Using weighting coefficients, the temperature comfort index, humidity comfort index, noise comfort index, and air quality comfort index are weighted to obtain the initial comprehensive comfort index for the sub-region: Among them, CCI initial w represents the initial overall comfort index of the sub-region. T The weighting coefficient w represents the temperature comfort index. RH The weighting coefficient w represents the humidity comfort index. L The weighting coefficients representing the noise comfort index. This represents the weighting coefficient of the air quality comfort index.

[0034] S205: Perform a nonlinear mapping on the initial comprehensive comfort index to obtain the comprehensive comfort index for the sub-region: CCI stands for Comprehensive Comfort Index of a sub-region.

[0035] In this embodiment of the invention, S201 maps standardized temperature, humidity, noise, and carbon dioxide concentration to dimensionless comfort indices, enabling comparison and fusion of environmental data with different dimensions on a unified scale. Simultaneously, exponential decay and positive deviation suppression functions enhance sensitivity to environmental deviations, facilitating accurate evaluation of the independent impact of each environmental factor. S202 calculates the personnel density ratio to characterize the actual activity intensity of personnel in each sub-area, dynamically coupling ventilation demand with personnel distribution. This improves ventilation assurance in densely populated areas while avoiding unnecessary air supply in unoccupied areas, thereby enhancing overall energy efficiency. S203 calculates the weighting coefficients of each comfort index based on the personnel density ratio, allowing the importance of temperature, humidity, noise, and air quality to adaptively adjust with changes in personnel activity. Furthermore, by introducing an equipment temperature and humidity influence coefficient, equipment safety requirements are considered, fundamentally solving the problem of fixed weights being difficult to adapt to different operating conditions. S204 weights and fuses various comfort indices to obtain an initial comprehensive comfort index that comprehensively reflects the regional environmental state, achieving a unified quantitative evaluation of multi-source heterogeneous environmental factors, significantly simplifying the decision-making process for airflow allocation and improving evaluation robustness. S205 employs nonlinear mapping to perform interval compression and sensitivity enhancement on the initial comprehensive comfort index, enabling the system to quickly sense and respond to environmental deterioration. This ensures that the final comprehensive comfort index not only conforms to the nonlinear laws of environmental changes but is also more suitable as a control input for subsequent airflow calculation and energy-saving optimization, thereby improving the accuracy, timeliness, and reliability of overall ventilation control.

[0036] S3: Calculate the target terminal air volume of each ventilation duct corresponding to each sub-area based on the comprehensive comfort index.

[0037] It should be noted that the target terminal air volume represents the final air volume that the ventilation system should deliver to each ventilation duct while meeting the actual environmental needs of each sub-area. It is the direct control target in the ventilation regulation process. This air volume is jointly determined by the comprehensive comfort index, the population density ratio, and the ventilation demand coefficient, and is uniformly adjusted according to the terminal air volume threshold of the entire system when necessary. This ensures that the overall air supply volume meets the air quality and heat and humidity requirements of each area without exceeding the fan capacity or causing energy waste. The target terminal air volume reflects the optimization result of the ventilation regulation strategy and serves as the input basis for subsequent damper angle calculation and fan pressure adjustment.

[0038] In one possible implementation, S3 specifically includes sub-steps S301 to S304: S301: Calculate the comfort deviation value for each sub-zone based on the comprehensive comfort index: Wherein, ΔCCI j CCI represents the comfort deviation value of the j-th sub-region.ref The Comprehensive Comfort Index (CCI) represents the reference threshold. j This represents the overall comfort index of the j-th sub-region.

[0039] It should be noted that the comfort deviation value measures the difference between the current environmental state and the target comfort level of a sub-area, and is a quantitative indicator describing the degree of "uncomfortable" environment. When the overall comfort index is below the reference threshold, the deviation value reflects the area's need for additional ventilation. When the overall comfort index meets or exceeds the reference threshold, the deviation value is set to zero to avoid unnecessary air supply to areas that have already met the standard, ensuring the rational allocation of ventilation resources among areas. The comfort deviation value plays a crucial role in calculating the ventilation demand coefficient, enabling the system to adjust airflow differently based on the urgency of environmental improvement.

[0040] S302: Calculate the ventilation demand coefficient for each sub-zone based on the various comfort deviation values: Where, k j Let α represent the ventilation demand coefficient for the j-th sub-region, α represent the weighting coefficient for the comfort deviation value, β represent the weighting coefficient for the population density ratio, and r represent the ventilation demand coefficient for the j-th sub-region. p,j This represents the population density ratio of the j-th sub-region.

[0041] It should be noted that the ventilation demand factor is used to comprehensively reflect the relative ventilation demand level of a sub-area under current operating conditions. It is a key intermediate variable derived from the comfort deviation value and the personnel density ratio. This factor increases when the comfort deviation is large or the personnel density is high, indicating that the area needs more air supply to improve environmental quality. It is lower when the environmental conditions are good or the personnel are sparse, avoiding excessive air supply and energy waste. The ventilation demand factor and the baseline air volume of the ventilation duct jointly determine the initial terminal air volume and are important driving forces in the calculation process of the target terminal air volume, enabling the entire air volume distribution process to achieve adaptive and precise control according to regional differences.

[0042] S303: Calculate the initial terminal air volume of each ventilation duct based on the various ventilation demand coefficients. in, This represents the initial terminal airflow of the j-th ventilation duct, where min indicates taking the minimum value, and q max,j q represents the maximum allowable terminal air volume of the j-th ventilation duct. min,j q represents the minimum allowable terminal airflow of the j-th ventilation duct. base,j This represents the design baseline end air volume of the j-th ventilation duct.

[0043] S304: Determine if the sum of all initial terminal air volumes is greater than the terminal air volume threshold. If so, introduce a terminal air volume correction factor to correct the initial terminal air volumes and obtain the target terminal air volume for the ventilation duct. Otherwise, use the initial terminal air volumes as the target terminal air volume for the ventilation duct.

[0044] Optionally, the initial terminal air volume can be corrected according to the following formula to obtain the target terminal air volume: Where, q opt,j Let Q represent the target terminal airflow of the j-th ventilation duct, η represent the terminal airflow correction factor, and Q represent the target terminal airflow of the j-th ventilation duct. max This indicates the terminal airflow threshold, and M indicates the number of ventilation ducts.

[0045] It should be noted that the terminal air volume threshold refers to the total allowable upper limit of the air volume at the end of all ventilation ducts in the entire ventilation system. It is usually determined by factors such as fan capacity, energy budget or ventilation safety standards. It is used to limit the overall air volume of the system from exceeding the design limit to prevent energy consumption from soaring, equipment overload or environmental disturbance from becoming unbalanced due to excessive total air volume, and to ensure that ventilation control operates within safe and energy-saving constraints.

[0046] In this embodiment of the invention, S301 calculates the comfort deviation value of each sub-region, quantifying the difference between the environmental state and the target comfort level into a calculable deviation index. This allows the system to accurately identify the degree of "uncomfortable" environment and avoid providing excessive ventilation to areas where the environment has already met the standards, thereby achieving a reasonable allocation of air volume resources. S302 calculates the ventilation demand coefficient based on the comfort deviation value and the ratio of personnel density. This ensures that the air volume demand not only reflects the urgency of environmental deterioration but also reflects the impact of personnel activity intensity on the thermal and humid environment and air quality. This achieves a precise characterization of the ventilation demand in each region, avoiding excessive air supply in sparsely populated areas and significantly improving energy efficiency. S303 calculates the initial terminal air volume based on the ventilation demand coefficient and combined with the maximum, minimum, and baseline air volume of each ventilation duct. This ensures that the air volume output meets the ventilation demand while being constrained by engineering boundaries, thereby guaranteeing that the air supply volume is neither too low, leading to environmental deterioration, nor too high, causing energy waste or airflow disturbance, thus improving the safety and stability of system operation. When the total air volume may exceed the fan capacity or the safety threshold of the ventilation system, S304 uses a terminal air volume correction factor to uniformly scale the initial terminal air volume, so that all ventilation ducts can meet the total air volume limit without changing the air volume distribution ratio, avoiding system overload and excessive energy consumption, and ensuring that the overall air volume distribution operates under safe, reasonable and energy-saving constraints.

[0047] S4: Build a data-driven model.

[0048] Optionally, the data-driven model specifically includes: a first data-driven model for describing the relationship between the terminal air volume and the terminal pressure drop of the ventilation duct, and a second data-driven model for describing the relationship between the damper pressure drop and the damper angle of the ventilation duct.

[0049] It should be noted that the first and second data-driven models are used to characterize the intrinsic relationships between different physical quantities in the ventilation system. The first model, by learning the nonlinear mapping between terminal airflow and terminal pressure drop, achieves accurate prediction of changes in airflow resistance in the ventilation ducts, providing crucial input for subsequent global fan pressure optimization. The second model describes the functional relationship between damper pressure drop and damper angle. By learning the characteristics of damper throttling at different angles, it achieves a quantifiable expression of damper control behavior, laying the foundation for solving the optimal damper angle for each ventilation duct. Together, these two models constitute a data-driven physical proxy for the ventilation system, enabling complex airflow characteristics to participate in fan energy consumption optimization in a computable form.

[0050] In one possible implementation, S4 specifically includes sub-steps S401 and S402: S401: Combining a denoising autoencoder neural network and a multilayer perceptron neural network to construct the first data-driven model.

[0051] It should be noted that the denoising autoencoder neural network is a type of deep learning model capable of recovering original features from input data contaminated by noise. It extracts the latent representation of the data through reconstruction learning of noisy samples. In this method, it is used to enhance the robustness of end-point airflow features and improve the model's tolerance to measurement errors. The multilayer perceptron neural network is a typical feedforward fully connected neural network structure with powerful nonlinear fitting capabilities. Through a combination of multiple hidden layers and activation functions, it can accurately approximate the complex mapping relationship between end-point airflow and end-point pressure drop. In model training, the denoising autoencoder is used to initialize the hidden layer weights of the multilayer perceptron, thereby improving training efficiency and model generalization performance. Together, they constitute a high-precision first data-driven model.

[0052] Specifically, firstly, a denoising autoencoder is used to perform layer-by-layer reconstruction learning on the terminal airflow samples after adding Gaussian noise, extracting stable and robust latent feature representations. The weight parameters obtained from training each denoising autoencoder are then used as the initial weights for the corresponding hidden layers of the multilayer perceptron. Subsequently, the multilayer perceptron is trained under supervision using real terminal airflow and terminal pressure drop samples. Based on this, the network parameters are iteratively updated using the error backpropagation algorithm, enabling the model to accurately approximate the nonlinear mapping relationship between terminal airflow and terminal pressure drop, thus forming a first data-driven model with high-precision prediction capabilities and strong noise resistance.

[0053] Optionally, the first data-driven model includes multiple sequentially connected denoising autoencoder neural networks. The number of denoising autoencoder neural networks is the same as the number of hidden layers in a multilayer perceptron neural network, with each denoising autoencoder neural network corresponding to one hidden layer.

[0054] In one possible implementation, the training method of the first data-driven model specifically includes sub-steps S4011 to S4014: S4011: Obtain the terminal air volume sample dataset and the corresponding terminal pressure drop sample dataset.

[0055] S4012: Add Gaussian noise to all end air volume samples in the end air volume sample dataset to obtain a noisy end air volume sample dataset.

[0056] S4013: Use the noisy terminal air volume sample dataset as the first training dataset, iteratively train each denoising autoencoder neural network in the order of sequential connection, and determine the weight coefficients of each denoising autoencoder neural network.

[0057] Specifically, the first training dataset (i.e., the noisy end-point airflow sample dataset) is input into the first denoising autoencoder neural network, and the training objective is to make the output result approximate the original noiseless end-point airflow sample. Then, the output of the first denoising autoencoder neural network is directly used as the input of the second denoising autoencoder neural network, and the second denoising autoencoder neural network is trained with the same objective of making the output result approximate the original noiseless end-point airflow sample. This process continues, using the output of the previous denoising autoencoder neural network as the input of the next, and so on, until the training of the last denoising autoencoder neural network is completed and the weight coefficients of each denoising autoencoder neural network are determined.

[0058] S4014: Use the weight coefficients of each denoising autoencoder neural network as the initial weight coefficients of each corresponding hidden layer, and use the terminal air volume sample dataset and the terminal pressure drop sample dataset as the second training dataset to train the multilayer perceptron neural network and determine the optimal set of network parameters for the multilayer perceptron neural network.

[0059] Specifically, the terminal airflow sample dataset is used as input, and the terminal pressure drop sample dataset is used as output target. Supervised learning is performed through a multilayer perceptron neural network, and the network weights and bias parameters are iteratively optimized using the backpropagation algorithm to minimize the mean square error between the predicted output and the actual pressure drop. In each iteration, based on the residual error between the network output and the target output, the weight matrices and bias vectors of each hidden layer and output layer are updated sequentially until the preset convergence condition of the loss function is met or the maximum number of training epochs is reached, ultimately obtaining the optimal set of network parameters for the multilayer perceptron neural network.

[0060] In this embodiment of the invention, S4011 acquires the terminal airflow sample dataset and its corresponding terminal pressure drop sample dataset, enabling the first data-driven model to learn the physical characteristics of the ventilation duct based on real-world operating data, thereby ensuring the engineering reliability of the model's prediction results. S4012 adds Gaussian noise to the terminal airflow samples to construct noisy training data, enhancing the model's robustness to sensor errors, turbulence disturbances, and environmental noise, preventing the model from being effective only under ideal conditions and experiencing performance degradation in actual operating conditions. S4013 sequentially inputs the noisy data into multiple denoising autoencoder neural networks for layer-by-layer denoising pre-training, enabling each layer of the neural network to extract airflow features at different levels and effectively suppress noise interference. This also solves the gradient vanishing and random initialization instability problems in deep neural network training, thereby improving the model's feature representation ability, training convergence, and generalization performance. S4014 uses weights obtained from pre-training with a denoised autoencoder as the initial weights for the hidden layer of the multilayer perceptron, and performs supervised learning fine-tuning with samples of terminal airflow and terminal pressure drop. This enables the model to accurately approximate the complex nonlinear relationship between terminal airflow and terminal pressure drop, achieving high-precision and high-stability pressure drop prediction. This provides reliable data support for subsequent minimization of wind turbine energy consumption and the solution of optimal damper angle.

[0061] S402: Construct a second data-driven model based on the ridge regression algorithm.

[0062] It should be noted that ridge regression is a parameter estimation method that introduces an L2 regularization term on top of traditional linear regression. By applying a squared penalty to the regression coefficients, it effectively suppresses the impact of multicollinearity on model stability, avoids parameter overfitting, and improves the model's generalization ability. In this method, ridge regression is used to fit the relationship between damper angle and damper pressure drop, enabling the model to maintain robust predictive performance even when input features are correlated or subject to noise interference. This ensures that the second data-driven model can accurately characterize the throttling characteristics under different damper angles.

[0063] Specifically, a linear regression model is established with the damper angle as input and the damper pressure drop as output. An L2 regularization term is introduced into the loss function to penalize excessive fluctuations in the regression coefficients. The parametric relationship between the damper angle and the damper pressure drop is solved by minimizing the ridge regression loss function, which includes the squared error and the regularization term. Based on this, a damper pressure drop prediction model is obtained that can suppress the influence of feature collinearity, improve model stability and generalization performance, and accurately represent the throttling characteristics under different damper angles in mathematical form, thus constructing a second data-driven model.

[0064] In this embodiment of the invention, S401 constructs a first data-driven model by combining a denoising autoencoder neural network and a multilayer perceptron neural network. This model first uses the denoising autoencoder to extract stable and robust latent features from noisy airflow data, and then uses a multilayer perceptron to accurately fit the highly nonlinear mapping relationship between terminal airflow and terminal pressure drop. This significantly improves the accuracy and anti-interference capability of terminal pressure drop prediction, and enhances the convergence performance and generalization ability of model training, providing a reliable data foundation for subsequent global fan pressure optimization. S402 constructs a second data-driven model based on the ridge regression algorithm. By introducing an L2 regularization term to suppress multicollinearity and noise interference, it ensures that the functional relationship between damper angle and damper pressure drop can be accurately expressed in a stable, interpretable, and computationally efficient form. This allows the damper throttling characteristics to be quickly solved in real-time optimization, thus providing a lightweight and highly reliable mathematical model for calculating the optimal damper angle and optimizing overall energy consumption.

[0065] S5: With the goal of minimizing fan energy consumption, the air volume of each target terminal is input into the data-driven model, and the optimal damper angle of each ventilation duct is output.

[0066] It should be noted that fan energy consumption quantifies the energy consumed by a fan under specific operating conditions to maintain the required airflow of the system. Its magnitude is determined by both the overall fan pressure and the total airflow, and is a core indicator for measuring the operating efficiency of a ventilation system. The optimal damper angle, on the other hand, represents the damper opening angle of each ventilation duct that minimizes overall fan pressure and fan energy consumption while meeting the corresponding target terminal airflow. It is the most energy-efficient control command obtained through data-driven models and mathematical equations. The optimal damper angle reflects the collaborative optimization results of the ventilation system under multi-zone demand constraints and is the final execution quantity for achieving energy-saving control.

[0067] Optionally, the formula for calculating wind turbine energy consumption is as follows: Where Power represents the energy consumption of the wind turbine, p fan This indicates the current global fan pressure, M represents the number of ventilation ducts, and q opt,jφ represents the target terminal air volume of the j-th ventilation duct, and φ represents the fan efficiency.

[0068] In one possible implementation, S5 specifically includes: With the goal of minimizing fan energy consumption, i.e. minimizing the current global fan pressure, the air volume of each target terminal is input into the data-driven model, and the optimal damper angle of each ventilation duct is output.

[0069] It should be noted that under constant ventilation conditions, the energy consumption of the fan is directly proportional to "fan output pressure × total air volume". The total air volume is determined by the target terminal air volume in each operating condition and is a fixed value. Therefore, the only adjustable factor for fan energy consumption is the global fan pressure p. fan When p fan When p decreases, the amount of work required for the fan to overcome pipeline resistance decreases accordingly, and the overall energy consumption also decreases. Conversely, if p increases... fan If the airflow is too high, the fan must expend more energy to maintain the airflow for the same air volume. Therefore, the essence of achieving energy saving is to rationally adjust the damper angles of each ventilation duct so that the system can meet the target terminal airflow while reducing p. fan To reduce to a minimum, thereby minimizing wind turbine energy consumption.

[0070] In one possible implementation, S5 specifically includes sub-steps S501 to S508: S501: Input the air volume of each target terminal into the first data-driven model, and output the terminal pressure drop of each ventilation duct: , where Δp j This represents the pressure drop at the end of the j-th ventilation duct, where j = 1, 2, ..., M, and M represents the number of ventilation ducts.

[0071] S502: Establish the first mathematical equation describing the relationship between terminal pressure drop and global fan pressure: Where, p j This represents the pressure drop of the damper in the j-th ventilation duct.

[0072] It should be noted that the first mathematical equation describes the fundamental mechanical relationship between the pressure drop at the end of each ventilation duct and the overall fan pressure. By adding the damper pressure drop to the end pressure drop, the minimum pressure required by the fan is obtained, thus realizing the correlation calculation between the local resistance of the ventilation duct and the overall system pressure. This equation reflects the total resistance that the fan must overcome, which is an important constraint in the subsequent energy consumption minimization solution.

[0073] S503: Using a second data-driven model, establish a second mathematical equation describing the relationship between damper pressure drop and damper angle in ventilation ducts: Where, δ j θ represents the damper angle of the j-th ventilation duct. j The ridge regression intercept term, θ j ε represents the current damper angle of the j-th ventilation duct. j θ represents the damper angle of the j-th ventilation duct. j The ridge regression coefficient term.

[0074] It should be noted that the second mathematical equation describes the functional relationship between damper pressure drop and damper angle. A ridge regression model is used to establish an expression for the impact of angle changes on airflow throttling capability, allowing for accurate quantification of pressure drop at different damper angles. This equation reveals a calculable mapping between damper adjustment behavior and airflow resistance, and is an important predictive model for solving the optimal damper angle.

[0075] S504: Integrating the first and second mathematical equations, a comprehensive mathematical equation is established to describe the relationship between terminal pressure drop, global fan pressure, damper pressure drop, and damper angle. .

[0076] It should be noted that the integrated mathematical equation combines the first and second mathematical equations to form a unified expression for the global fan pressure, damper angle, damper pressure drop, and terminal pressure drop, enabling the multi-source physical relationships to be solved collaboratively within the same equation system. The integrated mathematical equation establishes a complete mapping relationship from local airflow resistance to global fan pressure, and is the core mathematical basis for minimizing fan energy consumption and determining the optimal damper angle.

[0077] S505: Determine the optimal fully open damper in each ventilation duct based on the pressure drop at each end.

[0078] Alternatively, the optimal fully open damper in each ventilation duct can be determined according to the following formula: Where, j o This indicates the optimal fully open damper.

[0079] S506: Based on the optimal fully open damper, solve the comprehensive mathematical equation to obtain the minimum global fan pressure and minimize the fan energy consumption.

[0080] Optionally, the minimum global fan pressure is specifically: in, This represents the minimum global fan pressure. This represents the damper angle of the ventilation duct corresponding to the optimal fully open damper. The ridge regression intercept term, This represents the damper angle of the ventilation duct corresponding to the optimal fully open damper. This represents the target terminal air volume of the ventilation duct corresponding to the optimal fully open damper. This represents the pressure drop at the end of the ventilation duct corresponding to the optimal fully open damper.

[0081] S507: Calculate the optimal damper angle for each ventilation duct based on the minimum global fan pressure and the pressure drop at each terminal.

[0082] Optionally, the optimal damper angle for each ventilation duct can be calculated using the following formula: in, This represents the optimal damper angle for the j-th ventilation duct.

[0083] S508: Outputs the optimal damper angles.

[0084] In this embodiment of the invention, S501 obtains the terminal pressure drop of each ventilation duct by inputting the target terminal airflow into the first data-driven model, achieving accurate prediction of actual resistance under complex airflow conditions and providing reliable input for subsequent optimization. S502 establishes a mathematical relationship between terminal pressure drop and global fan pressure, enabling precise quantification of the total resistance the fan needs to overcome, forming a necessary constraint for energy-saving optimization. S503 constructs a mapping between damper angle and damper pressure drop based on the second data-driven model, making the damper throttling characteristics calculable, thereby achieving an accurate description of resistance changes at different angles. S504 integrates the damper pressure drop equation and the terminal pressure drop equation into a unified comprehensive mathematical model, forming a complete and jointly solvable correlation structure between damper angle, duct pressure drop, and global fan pressure, laying a mathematical foundation for solving the minimum fan pressure. S505 determines the optimal fully open damper by comparing the pressure drop combinations of each ventilation duct, narrowing the optimization search range and ensuring the physical feasibility of the pressure solution. S506 calculates the minimum global fan pressure based on the optimal fully open damper, thus clarifying the energy-saving optimization objective and obtaining the globally optimal pressure value. S507 solves for the optimal damper angle for each duct under the constraint of the minimum global fan pressure, ensuring that each ventilation duct achieves the lowest overall energy consumption while meeting the target terminal airflow. S508 outputs the optimal damper angle for each ventilation duct, forming an energy-saving control command that can be directly issued to the actuator, thereby realizing global collaborative optimization and fan energy consumption minimization control based on damper adjustment.

[0085] S6: Control the current damper angle of each ventilation duct to the corresponding optimal damper angle to complete the energy-saving ventilation control of the ship's engine room.

[0086] Specifically, by using electric damper actuators, the current damper angle of each ventilation duct is controlled to adjust to the corresponding optimal damper angle, thereby completing the energy-saving ventilation control of the ship's engine room.

[0087] In this embodiment of the invention, the target terminal airflow is determined based on a comprehensive comfort index, and a terminal pressure drop and damper pressure drop prediction mechanism is constructed using a data-driven model. Furthermore, the optimal damper angle for each ventilation duct is determined with the goal of minimizing the overall fan pressure. Finally, the optimal opening is precisely adjusted using an electric damper actuator, thereby achieving on-demand air supply, coordinated control, and energy minimization of the naval ventilation system. This invention not only enables refined airflow allocation under varying environmental demands in multiple areas, reducing unnecessary air supply losses, but also dynamically adapts ventilation strategies according to real-time operating conditions, significantly improving ventilation efficiency, equipment safety, and system adaptability, achieving efficient, reliable, and engineering-feasible energy-saving control of the naval cabin.

[0088] Reference manual attached Figure 2 The diagram shows a structural schematic of a ship engine room ventilation energy-saving control system provided in an embodiment of the present invention.

[0089] This invention provides a ship engine room ventilation energy-saving control system 20, including: a processor 201 and a memory 202; The memory 202 stores programs or instructions that can run on the processor 201. When the program or instructions are executed by the processor 201, they implement the steps of the above-mentioned ship engine room ventilation energy-saving control method and achieve the same technical effect. To avoid repetition, the present invention will not repeat the above-mentioned steps.

[0090] It should be understood that the processor 201 in this embodiment of the invention may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0091] It should also be understood that the memory 202 in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDR SDRAM), enhanced synchronous DRAM (ESDRAM), synchronous linked DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0092] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0093] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0094] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0095] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0096] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0097] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0098] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0099] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0100] This invention provides a readable storage medium comprising: storing a program or instructions on the readable storage medium, wherein when the program or instructions are executed by a processor, the program or instructions implement the steps of the above-described energy-saving control method for ventilation of ship engine room, and can achieve the same technical effect. To avoid repetition, this invention will not elaborate further.

[0101] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the embodiments of the present invention, and are not intended 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 of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the protection scope of the present invention.

Claims

1. A method for energy-saving control of ventilation in a ship's engine room, characterized in that, include: S1: Divide the ship's engine room into multiple sub-regions proportionally, and acquire multimodal data of the ship's engine room during operation. Each sub-region corresponds to a ventilation duct. S2: Calculate the comprehensive comfort index of each of the sub-regions based on the multimodal data; S3: Calculate the target terminal air volume of each ventilation duct corresponding to each of the sub-regions based on the comprehensive comfort index; S4: Building a data-driven model; S5: With the goal of minimizing fan energy consumption, input the air volume of each target terminal into the data-driven model and output the optimal damper angle of each ventilation duct. S6: Control the current damper angle of each ventilation duct to the corresponding optimal damper angle to complete the energy-saving ventilation control of the ship's engine room.

2. The energy-saving control method for ship engine room ventilation according to claim 1, characterized in that, The multimodal data specifically includes: environmental data and operating condition data; The environmental data specifically includes: the current temperature, current humidity, current noise level, current carbon dioxide concentration, and current number of workers in each of the sub-regions; The specific operating data includes: current global fan pressure, current fan speed, number of operating fans, and current terminal air volume, current terminal pressure drop, current damper pressure drop, and current damper angle for each ventilation duct.

3. The energy-saving control method for ship engine room ventilation according to claim 2, characterized in that, After S1 and before S2, it also includes: S1A: Perform data standardization processing on the environmental data.

4. The energy-saving control method for ship engine room ventilation according to claim 3, characterized in that, The calculation method for the comprehensive comfort index specifically includes: S201: Based on the current temperature, current humidity, current noise level, and current carbon dioxide concentration after data standardization, calculate the temperature comfort index, humidity comfort index, noise comfort index, and air quality comfort index of the sub-region, respectively. S202: Calculate the personnel density ratio of the sub-region based on the current number of workers in the sub-region; S203: Based on the personnel density ratio, calculate the weighting coefficients of the temperature comfort index, the humidity comfort index, the noise comfort index, and the air quality comfort index respectively; S204: Using the weighting coefficients, the temperature comfort index, the humidity comfort index, the noise comfort index, and the air quality comfort index are weighted to obtain the initial comprehensive comfort index of the sub-region. S205: Perform a nonlinear mapping on the initial comprehensive comfort index to obtain the comprehensive comfort index of the sub-region.

5. The energy-saving control method for ship engine room ventilation according to claim 1, characterized in that, S3 specifically includes: S301: Calculate the comfort deviation value of each of the sub-regions based on the comprehensive comfort index; S302: Calculate the ventilation demand coefficient for each of the sub-regions based on the respective comfort deviation values; S303: Calculate the initial terminal air volume of each ventilation duct based on the ventilation demand coefficients. S304: Determine whether the sum of all the initial terminal air volumes is greater than the terminal air volume threshold; if so, introduce a terminal air volume correction factor to correct the initial terminal air volume and obtain the target terminal air volume of the ventilation duct; otherwise, use the initial terminal air volume as the target terminal air volume of the ventilation duct.

6. The energy-saving control method for ship engine room ventilation according to claim 1, characterized in that, The data-driven model specifically includes: a first data-driven model for describing the relationship between the terminal air volume and the terminal pressure drop of the ventilation duct, and a second data-driven model for describing the relationship between the damper pressure drop and the damper angle of the ventilation duct; S4 specifically includes: S401: Combine the denoising autoencoder neural network and the multilayer perceptron neural network to construct the first data-driven model; S402: Construct the second data-driven model based on the ridge regression algorithm.

7. The energy-saving control method for ship engine room ventilation according to claim 6, characterized in that, The first data-driven model includes multiple sequentially connected denoising autoencoder neural networks. The number of denoising autoencoder neural networks is the same as the number of hidden layers in the multilayer perceptron neural network, and each denoising autoencoder neural network corresponds to one hidden layer. The training methods for the first data-driven model specifically include: S4011: Obtain the terminal air volume sample dataset and the terminal pressure drop sample dataset corresponding to the terminal air volume sample dataset; S4012: Add Gaussian noise to all end air volume samples in the end air volume sample dataset to obtain a noisy end air volume sample dataset; S4013: Using the noisy terminal air volume sample dataset as the first training dataset, iteratively train each of the denoising autoencoder neural networks in the order of sequential connection, and determine the weight coefficients of each of the denoising autoencoder neural networks. S4014: The weight coefficients of each of the denoising autoencoder neural networks are used as the initial weight coefficients of each of the corresponding hidden layers, and the terminal air volume sample dataset and the terminal pressure drop sample dataset are used as the second training dataset to train the multilayer perceptron neural network and determine the optimal set of network parameters of the multilayer perceptron neural network.

8. The energy-saving control method for ship engine room ventilation according to claim 6, characterized in that, The specific formula for calculating the energy consumption of the wind turbine is as follows: ； Where Power represents the energy consumption of the wind turbine, p fan This indicates the current global fan pressure, M represents the number of ventilation ducts, and q opt,j φ represents the target terminal air volume of the j-th ventilation duct, and φ represents the fan efficiency. Specifically, S5 is: With the goal of minimizing the fan energy consumption, i.e. minimizing the current global fan pressure, the target terminal air volume is input into the data-driven model, and the optimal damper angle of each ventilation duct is output.

9. The energy-saving control method for ship engine room ventilation according to claim 8, characterized in that, S5 specifically includes: S501: Input the air volume of each target terminal into the first data-driven model and output the terminal pressure drop of each ventilation duct; S502: Establish a first mathematical equation describing the relationship between the terminal pressure drop and the global fan pressure; S503: Using the second data-driven model, establish a second mathematical equation describing the relationship between the damper pressure drop and the damper angle in the ventilation duct; S504: Integrate the first mathematical equation and the second mathematical equation to establish a comprehensive mathematical equation describing the relationship between the terminal pressure drop, the global fan pressure, the damper pressure drop, and the damper angle; S505: Determine the optimal fully open damper in each of the aforementioned end pressure drops; S506: Based on the optimal fully open damper, solve the comprehensive mathematical equation to obtain the minimum global fan pressure and minimize the fan energy consumption; S507: Calculate the optimal damper angle for each of the ventilation ducts based on the minimum global fan pressure and the pressure drop at each of the terminals; S508: Output the optimal damper angles described above.

10. A ship engine room ventilation energy-saving control system, characterized in that, include: Processor and memory; The memory stores programs or instructions that can run on the processor, which, when executed by the processor, implement the steps of the ship engine room ventilation energy-saving control method as described in any one of claims 1 to 9.

Images