A central air conditioning intelligent scheduling method, system and storage medium based on multi-objective optimization

Abstract

The present invention relates to a central air-conditioning intelligent scheduling method, system and storage medium based on multi-objective optimization, comprising the following steps: S1, collecting basic equipment information, real-time operation data and energy consumption data of refrigeration units; S2, calculating the performance coefficient of each refrigeration unit based on the collected cooling capacity and the energy consumption data, calculating the health of each refrigeration unit based on the real-time operation data, and building an operation database according to the number, performance coefficient and health of each refrigeration unit; S3, using the NSGA-II algorithm to solve the multi-objectives of maximizing the average performance coefficient, maximizing the total health and balancing the operation time, and generating an optimal scheduling solution set for the refrigeration unit, wherein the optimal scheduling solution set includes the start and stop status and load distribution information of each refrigeration unit; S4, performing multi-objective scoring weight assignment on each candidate solution of the optimal scheduling solution set, performing TOPSIS comprehensive scoring, and outputting the scheduling solution with the highest score.

Description

Technical Field

[0001] The present invention relates to the field of refrigeration balance control, and in particular to a central air-conditioning intelligent scheduling method, system and storage medium based on multi-objective optimization. Background Art

[0002] In large commercial buildings, industrial plants, data centers, and other applications, central air conditioning systems and air compressor stations serve as core energy equipment, and their operating efficiency directly impacts overall energy consumption and O&M costs. To dynamically respond to load demand, these systems typically deploy multiple chillers or air compressors and use group control strategies to select the optimal startup combination. To achieve this, most current group control strategies for central air conditioning and air compressor stations still rely on hard-coded logic based on empirical rules. Typical approaches include: 1. Priority rotation: Units are started and stopped in a fixed order to balance operating time; 2. COP priority: Units with higher Coefficient of Performance (COP) are prioritized; 3. Health threshold: A health threshold (e.g., 0.7) is set, preventing startup if the health threshold falls below this value; and 4. Load matching lookup tables: Preset combinations are selected based on the current cooling load or air volume demand.

[0003] Although traditional rule-based control methods can meet basic operational requirements to a certain extent, they have many limitations in practical applications, mainly including the following aspects:

[0004] 1. Inadequate ability to balance multiple objectives (dimensional conflict). In actual operations, multiple objectives often need to be optimized simultaneously, such as maximizing energy efficiency, minimizing equipment loss, and maximizing operational stability. However, traditional rule-based approaches lack a unified evaluation system, making it difficult to address conflicting objectives. For example, if a unit has a high COP but poor health, manually assigning weights (e.g., "COP 60% + Health 40%)" is required to determine whether to enable it. This is highly subjective and lacks scientific basis.

[0005] 2. High combinatorial complexity leads to incomplete rule coverage (combinatorial explosion). For example, for five chillers, there are 31 possible startup combinations. Taking into account factors such as load distribution and start-stop sequence, the state space expands further. Traditional hard-coded logic cannot exhaustively enumerate all possible combinations. Especially as the system scales up, the number of rules increases exponentially, resulting in high maintenance costs and the risk of missed or incorrect matches, impacting system performance.

[0006] 3. Improper static threshold settings (poor adaptability). Existing systems often use static judgment criteria such as "disable if health < 0.7." However, device health is a continuously changing indicator over time, and the decay curves of different devices vary significantly. Fixed thresholds fail to reflect the true state of devices and can easily lead to misjudgments or excessive restrictions, impacting system availability and scheduling flexibility.

[0007] 4. Lack of adaptive capabilities and data-driven support. Traditional rule-based control relies entirely on manual settings, making it difficult to dynamically adjust to environmental changes, historical operating data, or user preferences. For example, focusing on energy conservation during the day and equipment rotation at night, hard-coded logic cannot automatically identify and switch strategies, resulting in delayed decision-making and reduced efficiency.

[0008] The purpose of this invention is to design a central air-conditioning intelligent scheduling method, system and storage medium based on multi-objective optimization to solve the problems existing in the above-mentioned prior art. Summary of the Invention

[0009] In response to the problems existing in the above-mentioned prior art, the present invention provides a central air-conditioning intelligent scheduling method, system, and storage medium based on multi-objective optimization, which can effectively solve at least one problem existing in the above-mentioned prior art.

[0010] The technical solution of the present invention is:

[0011] A central air conditioning intelligent scheduling method based on multi-objective optimization includes the following steps:

[0012] S1, collects basic equipment information, real-time operation data, and energy consumption data of the refrigeration unit;

[0013] S2, calculating the performance coefficient of each refrigeration unit based on the collected cooling capacity and the energy consumption data, calculating the health of each refrigeration unit based on the real-time operation data, and building an operation database according to the number, performance coefficient, and health of each refrigeration unit;

[0014] S3: With the multi-objective objectives of maximizing the average performance coefficient, maximizing the total health, and balancing the operating time, and with the total cooling capacity meeting the demand as the premise, a multi-objective optimization model is constructed and solved using the NSGA-II algorithm to generate the optimal scheduling solution set for the refrigeration units. The optimal scheduling solution set includes the start and stop status and load distribution information of each refrigeration unit;

[0015] S4, performing multi-objective scoring weight assignment on each candidate solution of the optimal scheduling solution set, performing TOPSIS comprehensive scoring on the optimal scheduling solution set based on the scoring weight assignment result, and outputting the scheduling solution with the highest score.

[0016] Furthermore, the healthiness is calculated by the following method:

[0017] Obtain the equipment operating status, key parameter abnormalities, cumulative operating time, fault history, and control stability of the refrigeration unit;

[0018] The equipment operating status, key parameter anomalies, cumulative operating time, fault history, and control stability are graded and assigned corresponding health scores;

[0019] Assign health weights to equipment operating status, key parameter anomalies, cumulative operating time, fault history, and control stability;

[0020] The equipment operating status, key parameter abnormalities, cumulative operating time, fault history, and control stability are weightedly added according to the health score and the health weight, and the result obtained is used as the health.

[0021] Furthermore, in step S4, performing multi-objective scoring weight allocation on each candidate solution of the optimal scheduling solution set includes:

[0022] Set the weight of performance coefficient to 0.4~0.6, the weight of health to 0.25~0.35, and the weight of operation balance to 0.15~0.25.

[0023] Furthermore, performing TOPSIS comprehensive scoring on the optimal scheduling solution set based on the scoring weight distribution result includes:

[0024] Construct the optimal scheduling solution set as the original decision matrix;

[0025] Normalizing the original decision matrix to obtain a normalized decision matrix;

[0026] Based on the score weight distribution result, the normalized decision matrix is ​​weighted and normalized to obtain a weighted normalized decision matrix;

[0027] The maximum value of each column in the weighted normalized decision matrix is ​​set as a positive ideal solution, and the minimum value of each column is set as a negative ideal solution;

[0028] Calculating the relative closeness of each solution in the weighted normalized decision matrix to the positive ideal solution and the negative ideal solution respectively;

[0029] The solution that is closest to the positive ideal solution is taken as the scheduling solution with the highest score.

[0030] Further, the following steps are included:

[0031] S5, collecting multiple scheduling solutions with the highest scores generated in different time dimensions to form a scheduling solution matrix, performing non-local mean filtering on the scheduling solution matrix, extracting the main trend of the unit combination change, and obtaining a filtered scheduling solution matrix;

[0032] S6, collecting multiple influencing factors of different time dimensions to form an influencing factor two-dimensional matrix, combining the influencing factor two-dimensional matrix with the filtered scheduling solution matrix to perform principal component analysis, and extracting the main direction that most significantly affects the scheduling change;

[0033] S7, dynamically adjusting the target scoring weights of the candidate solutions in step S4 according to the main direction that most significantly affects the scheduling change.

[0034] Furthermore, in step S7, dynamically adjusting the target scoring weights of the candidate solutions in step S4 according to the main direction that most significantly affects the scheduling change includes:

[0035] If the main direction that most significantly affects scheduling changes is temperature or humidity, increase the weight of maximizing the average performance coefficient, reduce the weight of maximizing the total health, and maintain the weight of balancing the running time;

[0036] If the main direction that most significantly affects scheduling changes is load fluctuation, then reduce the weight of maximizing the average performance coefficient, increase the weight of maximizing the total health, and increase the weight of running time balancing;

[0037] If the main direction that most significantly affects the scheduling change is the decline in unit health, then the weight of maximizing the average performance coefficient is reduced, the weight of maximizing the total health is increased, and the weight of balancing the operating time is maintained;

[0038] If the main direction that most significantly affects scheduling changes is grid power rationing or peak power price scheduling, then the weight of maximizing the average performance coefficient should be increased, the weight of maximizing the total health should be reduced, and the weight of balancing the operating time should be maintained.

[0039] Furthermore, the process of performing non-local mean filtering on the scheduling solution matrix includes:

[0040] The smoothing parameters of the non-local mean filter are dynamically adjusted according to the influencing factors.

[0041] Furthermore, the smoothing parameters of the non-local mean filter are dynamically adjusted according to the influencing factors, including:

[0042] If the outdoor temperature is greater than the preset temperature, the smoothing parameter is increased; if not and the temperature changes smoothly, the smoothing parameter is decreased;

[0043] If the load fluctuates, the smoothing parameter is increased;

[0044] If there is a decrease in unit health, the smoothing parameter is increased.

[0045] A refrigeration unit balancing control system based on multi-objective optimization is further provided, which implements the central air-conditioning intelligent scheduling method based on multi-objective optimization during operation.

[0046] A computer-readable storage medium is further provided, storing a computer program, which, when executed by a processor, performs the intelligent scheduling method for central air conditioning based on multi-objective optimization.

[0047] Therefore, the present invention provides the following effects and / or advantages:

[0048] This application uses the multi-objective evolutionary algorithm NSGA-II, based on dynamic multi-objective trade-offs, to automatically generate a set of non-dominated solutions and flexibly adjust the importance of each objective under different operating conditions. Combined with the TOPSIS algorithm for scoring, the highest-scoring solution is output, enabling the system to achieve an optimal balance between energy efficiency, equipment health, and operational stability while meeting current load demands.

[0049] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention are realized and obtained by the structures particularly pointed out in the description and the drawings.

[0050] It is to be understood that both the foregoing general description and the following detailed description of the present invention are exemplary and explanatory and are intended to provide further explanation of the invention as claimed. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 A flowchart illustrating one embodiment of the present invention is provided. DETAILED DESCRIPTION

[0052] In order to facilitate understanding by those skilled in the art, the present invention is now described in further detail with reference to the following examples:

[0053] refer to Figure 1 , a central air conditioning intelligent scheduling method based on multi-objective optimization, comprising the following steps:

[0054] S1, collects basic equipment information, real-time operation data, and energy consumption data of the refrigeration unit;

[0055] Basic equipment information may include unit number, rated cooling capacity (RT), maintenance status, etc. Real-time operation data may be shown in Table 1 below. Energy consumption data may include input power, total power, etc.

[0056] Table 1, real-time operation data

[0057]

[0058] S2, calculating the performance coefficient of each refrigeration unit based on the collected cooling capacity and the energy consumption data, calculating the health of each refrigeration unit based on the real-time operation data, and building an operation database according to the number, performance coefficient, and health of each refrigeration unit;

[0059] In this step, coefficient of performance (COP) = cooling capacity / input power.

[0060] Furthermore, the healthiness is calculated by the following method:

[0061] Obtain the equipment operating status, key parameter abnormalities, cumulative operating time, fault history, and control stability of the refrigeration unit;

[0062] The equipment operating status, key parameter anomalies, cumulative operating time, fault history, and control stability are graded and assigned corresponding health scores;

[0063] Assign health weights to equipment operating status, key parameter anomalies, cumulative operating time, fault history, and control stability;

[0064] The equipment operating status, key parameter abnormalities, cumulative operating time, fault history, and control stability are weightedly added according to the health score and the health weight, and the result obtained is used as the health.

[0065] In this step, healthiness is a comprehensive assessment of the chiller unit's current operating status and potential risks, used to quantify the equipment's reliability and maintainability. Its value typically ranges from 0 to 1.0, with 1.0 indicating the unit is in brand new or optimal operating condition; 0 indicating a serious fault or unusable condition; 0.6 to 0.8 indicating that the unit is generally usable but presents some risk; and 0-0.6 indicating that the unit requires significant attention or scheduled maintenance. Using the weights shown in Table 2 and the healthiness example in Table 3, the healthiness score is 0.2 × 1.0 + 0.3 × 0.6 + 0.2 × 0.7 + 0.15 × 0.8 + 0.15 × 0.8 = 0.74.

[0066] Table 2. Description of healthiness dimensions and their weights

[0067]

[0068] Table 3, Example of health calculation

[0069]

[0070] By collecting data and calculating data in each dimension, the data in each dimension of the equipment is updated and written into the database in real time to obtain the operation database.

[0071] S3: With the multi-objective objectives of maximizing the average performance coefficient, maximizing the total health, and balancing the operating time, and with the total cooling capacity meeting the demand as the premise, a multi-objective optimization model is constructed and solved using the NSGA-II algorithm to generate the optimal scheduling solution set for the refrigeration units. The optimal scheduling solution set includes the start and stop status and load distribution information of each refrigeration unit;

[0072] In this step, maximizing the average COP is done by adding the COPs of all enabled chillers and dividing it by the number of enabled units to obtain the average COP of the combination. Based on the COPs of the chillers enabled in the current start-stop combination, their average COP is calculated to assess the overall energy efficiency of the combination. Maximizing total health is done by calculating the total health risk value of the current chiller combination. For all running chillers, the corresponding health values ​​are added to obtain the total health. A higher total health indicates a lower risk. Running time balancing is done by calculating the cumulative running time of each enabled chiller and calculating the standard deviation of the corresponding cumulative running time. A smaller standard deviation indicates closer running times for each unit and a more balanced system. At the same time, meeting the cooling capacity standard is used as a constraint.

[0073] The NSGA-II algorithm can randomly generate multiple groups of scheduling solutions that meet the constraints. For each scheduling solution, the average performance coefficient, total health, and running time balancing are calculated respectively. Then, according to the three target values, all scheduling solutions are sorted and divided into different levels of scheduling solutions. At the same time, multiple crossovers, merges, and iterations are performed, and finally converged and terminated to obtain a series of optimal solution sets that meet the regulation. Each solution set includes the start-up status of each refrigeration unit, the cooling load, etc.

[0074] S4, performing multi-objective scoring weight assignment on each candidate solution of the optimal scheduling solution set, performing TOPSIS comprehensive scoring on the optimal scheduling solution set based on the scoring weight assignment result, and outputting the scheduling solution with the highest score.

[0075] In this step, the method sets three decision-making evaluation indicators: performance coefficient, a benefit-based indicator; larger values ​​indicate better performance; healthiness, a benefit-based indicator reflecting the stability and reliability of the unit's current operating state; larger values ​​indicate better performance; and operating balance, a cost-based indicator reflecting the balance between the operating hours of each unit; smaller values ​​indicate greater balance. These three decision-making evaluation indicators are weighted to meet the required requirements and the actual conditions of the refrigeration unit. The TOPSIS algorithm is then used to calculate the optimal scheduling solution set. Based on the weights assigned above, the solution closest to the ideal target is calculated and becomes the scheduling solution with the highest score.

[0076] Furthermore, in step S4, performing multi-objective scoring weight allocation on each candidate solution of the optimal scheduling solution set includes:

[0077] Set the weight of performance coefficient to 0.4~0.6, the weight of health to 0.25~0.35, and the weight of operation balance to 0.15~0.25.

[0078] In this example, the weight of performance coefficient is 0.5, the weight of health is 0.3, and the weight of operational balance is 0.2. Performance coefficient is used to achieve the priority goal of operational economy, so it is assigned the highest weight. Health is the safety and sustainability goal and is assigned the second highest weight. Balance is an additional goal to reduce long-term maintenance costs and is therefore assigned the lowest weight.

[0079] Furthermore, performing TOPSIS comprehensive scoring on the optimal scheduling solution set based on the scoring weight distribution result includes:

[0080] Construct the optimal scheduling solution set as the original decision matrix;

[0081] Normalizing the original decision matrix to obtain a normalized decision matrix;

[0082] Based on the score weight distribution result, the normalized decision matrix is ​​weighted and normalized to obtain a weighted normalized decision matrix;

[0083] The maximum value of each column in the weighted normalized decision matrix is ​​set as a positive ideal solution, and the minimum value of each column is set as a negative ideal solution;

[0084] Calculating the relative closeness of each solution in the weighted normalized decision matrix to the positive ideal solution and the negative ideal solution respectively;

[0085] The solution that is closest to the positive ideal solution is taken as the scheduling solution with the highest score.

[0086] In this embodiment, for example, the optimal scheduling solution set shown in Table 3, the objective function value of each scheme is composed of a decision matrix according to the behavior performance coefficient, health risk, and total cooling capacity. , where the health risk in this step = 1-healthiness. Then the decision matrix is ​​normalized by column to convert all indicator values ​​into dimensionless and comparable values. Then each column is multiplied by its corresponding indicator weight to construct a weighted decision matrix.

[0087] Next, the Euclidean distance between each solution and the positive and negative ideal solutions is calculated. Using the Euclidean distance between the ideal and negative ideal solutions, the relative closeness of each solution is calculated. Relative closeness can be calculated as the distance between the solution and the negative ideal solution divided by the distance between the negative ideal solution and the positive ideal solution. Relative closeness indicates closer proximity to the positive ideal solution. After calculation, Solution A, as shown in Table 4, has the highest score and is output as the scheduling solution.

[0088] Table 4, schematic diagram of optimal scheduling solution set

[0089]

[0090] Further, the following steps are included:

[0091] S5, collecting multiple scheduling solutions with the highest scores generated in different time dimensions to form a scheduling solution matrix, performing non-local mean filtering on the scheduling solution matrix, extracting the main trend of the unit combination change, and obtaining a filtered scheduling solution matrix;

[0092] In this step, the scheduling solutions generated in step S4 at different times are combined into a scheduling solution matrix. Each row of the scheduling solution matrix represents a high-scoring scheduling solution. The odd-numbered columns in each row represent the on / off status of each refrigeration unit, with 1 representing on and 0 representing off. The even-numbered columns in each row represent the load factor of each refrigeration unit. For example, one row of the scheduling solution matrix can be represented as [1, 0.8, 0, 0.6, 1, 1.0]. Then, through non-local mean filtering, the actual response trend of the refrigeration unit under external interference can be reflected.

[0093] S6, collecting multiple influencing factors of different time dimensions to form an influencing factor two-dimensional matrix, combining the influencing factor two-dimensional matrix with the filtered scheduling solution matrix to perform principal component analysis, and extracting the main direction that most significantly affects the scheduling change;

[0094] In this step, the multiple influencing factors may include ambient temperature, humidity, cooling load, voltage fluctuation, cooling water temperature, health fluctuation, etc. These factors are grouped into each row, and the multiple influencing factors are combined according to different time dimensions to form a two-dimensional matrix of influencing factors.

[0095] By combining the two-dimensional matrix of influencing factors with the filtered scheduling solution matrix for principal component analysis, we can identify the main trend dimensions of the scheduling solution matrix and determine which variables among the multivariate influencing factors are most likely to significantly alter scheduling changes. Ultimately, we can determine the main direction that most significantly influences scheduling changes in the current situation. For example, changes in external temperature and humidity can significantly affect the future decisions of the refrigeration unit. This step primarily determines which factor will significantly affect the scheduling changes of the refrigeration unit in the future, so that adjustments can be made to that factor.

[0096] S7, dynamically adjusting the target scoring weights of the candidate solutions in step S4 according to the main direction that most significantly affects the scheduling change.

[0097] In step S6, significant influencing factors affecting future decisions are obtained. This step can further adjust the target scoring weights of each candidate solution in step S4 according to external influencing factors, thereby changing the weights of performance coefficient, health, and operation balance. Finally, after performing TOPSIS comprehensive scoring, a scoring result that adapts to the main direction that most significantly affects scheduling changes is obtained.

[0098] Furthermore, in step S7, dynamically adjusting the target scoring weights of the candidate solutions in step S4 according to the main direction that most significantly affects the scheduling change includes:

[0099] If the main direction that most significantly affects scheduling changes is temperature or humidity, increase the weight of maximizing the average performance coefficient, reduce the weight of maximizing the total health, and maintain the weight of balancing the running time;

[0100] If the main direction that most significantly affects scheduling changes is load fluctuation, then;

[0101] If the main direction that most significantly affects the scheduling change is the decline in unit health, then the weight of maximizing the average performance coefficient is reduced, the weight of maximizing the total health is increased, and the weight of balancing the operating time is maintained;

[0102] If the main direction that most significantly affects scheduling changes is grid power rationing or peak power price scheduling, then the weight of maximizing the average performance coefficient should be increased, the weight of maximizing the total health should be reduced, and the weight of balancing the operating time should be maintained.

[0103] In this step, when temperature or humidity is the main factor, the condensing pressure of the refrigeration unit may increase, the energy efficiency may decrease, and some old units may have extremely poor operating efficiency and rapid health deterioration at high temperatures. The weight of maximizing the average performance coefficient can be increased by 10-20%, and the weight of maximizing the total health can be reduced by 30-50%. The weight of balancing the operating time can be maintained, so as to give priority to high-efficiency units, avoid soaring energy consumption, and sacrifice some health in exchange for energy efficiency.

[0104] In this step, when load fluctuation is the main direction, there may be risks such as frequent start-up and shutdown of multiple refrigeration units, decreased health, and large control strategy jitter. The weight of maximizing the average performance coefficient is reduced by 30-50%, the weight of maximizing the total health is increased by 10-20%, and the weight of balancing the operating time is increased by 100-400%. In this way, a certain decrease in energy efficiency can be tolerated in exchange for system stability, reducing the risk of damage to frequently running units, and distributing the load to prevent overload of certain units.

[0105] In this step, when the main trend is unit health decline, individual units may be in poor condition but may still be mis-dispatched, or if not promptly avoided, they may cause shutdown, damage, or even linkage risks. The weight of maximizing the average performance coefficient is reduced by 10-30%, and the weight of maximizing the total health is increased by 40-60%. The weight of balancing the operating time is maintained, thereby appropriately sacrificing COP in exchange for system reliability, focusing on avoiding critical health equipment, and ensuring sustainable operation.

[0106] In this step, when the grid is mainly focused on power rationing or peak electricity price scheduling, it may be necessary to avoid high-consumption units during peak electricity prices. At the same time, stable system operation must be ensured to avoid tripping. The weight of maximizing the average performance coefficient is increased by 20-40%, and the weight of maximizing the total health is reduced by 80-90%. The weight of balancing the operating time is maintained, so that saving electricity is the primary goal and short-term health sacrifices are accepted.

[0107] Furthermore, the process of performing non-local mean filtering on the scheduling solution matrix includes:

[0108] The smoothing parameters of the non-local mean filter are dynamically adjusted according to the influencing factors.

[0109] In the non-local mean filtering algorithm, its smoothing parameters are dynamically adjusted according to external influencing factors such as temperature, load, water temperature, equipment fluctuations, etc., so as to achieve a better fit for physical environment changes.

[0110] Furthermore, the smoothing parameters of the non-local mean filter are dynamically adjusted according to the influencing factors, including:

[0111] If the outdoor temperature is greater than the preset temperature, the smoothing parameter is increased; if not and the temperature changes smoothly, the smoothing parameter is decreased;

[0112] If the load fluctuates, the smoothing parameter is increased;

[0113] If there is a decrease in unit health, the smoothing parameter is increased.

[0114] Whether the scheduling behavior is stable, whether it experiences dramatic changes, and whether it tracks real changes can reflect the dynamic balance of the refrigeration unit. When the smoothing parameter increases, the non-local mean filter filters the scheduling solution matrix more deeply, allowing it to ignore some details and sudden changes.

[0115] Specifically, if the outdoor temperature is greater than the preset temperature, the refrigeration unit will cause a jump in the scheduling plan in order to adapt to the temperature jump. At this time, increasing the smoothing parameter can suppress the jump in the scheduling plan and make the scheduling plan smoothly transition to the target. If not and the temperature changes steadily, then lower the smoothing parameter to make the scheduling plan actively respond to external changes; the preset temperature can be set to 30 degrees Celsius.

[0116] If the load fluctuates, the smoothing can be improved to identify the fluctuations and output a more stable scheduling plan;

[0117] If there is a decline in unit health, increasing the smoothing parameter can shield local extreme scheduling plans and stabilize the main trend.

[0118] A refrigeration unit balancing control system based on multi-objective optimization is further provided, which implements the central air-conditioning intelligent scheduling method based on multi-objective optimization during operation.

[0119] A computer-readable storage medium is further provided, storing a computer program, which, when executed by a processor, implements the central air-conditioning intelligent scheduling method based on multi-objective optimization.

[0120] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0121] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0122] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0123] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0124] In the description of this specification, the reference terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" mean that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic expressions of the above terms should not be understood as necessarily referring to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification and features of different embodiments or examples without contradiction.

Claims

1. A central air conditioning intelligent scheduling method based on multi-objective optimization, characterized by: The following steps are involved: S1, collects basic equipment information, real-time operation data, and energy consumption data of the refrigeration unit; S2, calculating the performance coefficient of each refrigeration unit based on the collected cooling capacity and the energy consumption data, calculating the health of each refrigeration unit based on the real-time operation data, and building an operation database according to the number, performance coefficient, and health of each refrigeration unit; S3: With the multi-objective objectives of maximizing the average performance coefficient, maximizing the total health, and balancing the operating time, and with the total cooling capacity meeting the demand as the premise, a multi-objective optimization model is constructed and solved using the NSGA-II algorithm to generate the optimal scheduling solution set for the refrigeration units. The optimal scheduling solution set includes the start and stop status and load distribution information of each refrigeration unit; S4, performing multi-objective scoring weight assignment on each candidate solution of the optimal scheduling solution set, performing TOPSIS comprehensive scoring on the optimal scheduling solution set based on the scoring weight assignment result, and outputting the scheduling solution with the highest score.

2. The central air conditioning intelligent scheduling method based on multi-objective optimization according to claim 1 is characterized in that: The health is calculated by the following method: Obtain the equipment operating status, key parameter abnormalities, cumulative operating time, fault history, and control stability of the refrigeration unit; The equipment operating status, key parameter anomalies, cumulative operating time, fault history, and control stability are graded and assigned corresponding health scores; Assign health weights to equipment operating status, key parameter anomalies, cumulative operating time, fault history, and control stability; The equipment operating status, key parameter abnormalities, cumulative operating time, fault history, and control stability are weightedly added according to the health score and the health weight, and the result obtained is used as the health.

3. The central air conditioning intelligent scheduling method based on multi-objective optimization according to claim 1 is characterized in that: In step S4, performing multi-objective scoring weight allocation on each candidate solution of the optimal scheduling solution set includes: Set the weight of performance coefficient to 0.4~0.6, the weight of health to 0.25~0.35, and the weight of operation balance to 0.15~0.

25.

4. The central air conditioning intelligent scheduling method based on multi-objective optimization according to claim 3 is characterized by: The TOPSIS comprehensive scoring of the optimal scheduling solution set based on the scoring weight distribution result includes: Construct the optimal scheduling solution set as the original decision matrix; Normalizing the original decision matrix to obtain a normalized decision matrix; Based on the score weight distribution result, the normalized decision matrix is ​​weighted and normalized to obtain a weighted normalized decision matrix; The maximum value of each column in the weighted normalized decision matrix is ​​set as a positive ideal solution, and the minimum value of each column is set as a negative ideal solution; Calculating the relative closeness of each solution in the weighted normalized decision matrix to the positive ideal solution and the negative ideal solution respectively; The solution that is closest to the positive ideal solution is taken as the scheduling solution with the highest score.

5. The central air conditioning intelligent scheduling method based on multi-objective optimization according to claim 1 is characterized in that: The following steps are involved: S5, collecting multiple scheduling solutions with the highest scores generated in different time dimensions to form a scheduling solution matrix, performing non-local mean filtering on the scheduling solution matrix, extracting the main trend of the unit combination change, and obtaining a filtered scheduling solution matrix; S6, collecting multiple influencing factors of different time dimensions to form an influencing factor two-dimensional matrix, combining the influencing factor two-dimensional matrix with the filtered scheduling solution matrix to perform principal component analysis, and extracting the main direction that most significantly affects the scheduling change; S7, dynamically adjusting the target scoring weights of the candidate solutions in step S4 according to the main direction that most significantly affects the scheduling change.

6. The central air conditioning intelligent scheduling method based on multi-objective optimization according to claim 5 is characterized by: S7, dynamically adjusting the target scoring weights of each candidate solution in step S4 according to the main direction that most significantly affects the scheduling change includes: If the main direction that most significantly affects scheduling changes is temperature or humidity, increase the weight of maximizing the average performance coefficient, reduce the weight of maximizing the total health, and maintain the weight of balancing the running time; If the main direction that most significantly affects scheduling changes is load fluctuation, then reduce the weight of maximizing the average performance coefficient, increase the weight of maximizing the total health, and increase the weight of running time balancing; If the main direction that most significantly affects the scheduling change is the decline in unit health, then the weight of maximizing the average performance coefficient is reduced, the weight of maximizing the total health is increased, and the weight of balancing the operating time is maintained; If the main direction that most significantly affects scheduling changes is grid power rationing or peak power price scheduling, then the weight of maximizing the average performance coefficient should be increased, the weight of maximizing the total health should be reduced, and the weight of balancing the operating time should be maintained.

7. The central air conditioning intelligent scheduling method based on multi-objective optimization according to claim 5 is characterized by: The process of performing non-local mean filtering on the scheduling solution matrix includes: The smoothing parameters of the non-local mean filter are dynamically adjusted according to the influencing factors.

8. The central air conditioning intelligent scheduling method based on multi-objective optimization according to claim 7 is characterized in that: The smoothing parameters of the non-local mean filter are dynamically adjusted according to the influencing factors, including: If the outdoor temperature is greater than the preset temperature, the smoothing parameter is increased; if not and the temperature changes smoothly, the smoothing parameter is decreased; If the load fluctuates, the smoothing parameter is increased; If there is a decrease in unit health, the smoothing parameter is increased.

9. A refrigeration unit balance control system based on multi-objective optimization, characterized in that: When working, the method realizes a central air-conditioning intelligent scheduling method based on multi-objective optimization as described in any one of claims 1 to 8.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method for intelligent scheduling of central air conditioners based on multi-objective optimization according to any one of claims 1 to 8 is implemented.

Images