Control method of building cooling system

Abstract

The present invention discloses a control method for a building cold source system, comprising: predicting a load prediction value at time τ based on the load measurement value of the current unit combination at time τ-Δτ: forming a plurality of selectable unit combinations based on the load prediction value and the load value of the current unit combination; optimizing and calculating the operating parameters of each unit combination to obtain the operating optimization parameters of each unit combination; calculating the comprehensive operating power of each unit combination based on the load rate of each unit combination, and selecting a unit combination as the optimization result according to the comprehensive operating power. The present invention takes the comprehensive energy efficiency of the chiller as the goal, comprehensively optimizes the cold source system operation strategy composed of the start and stop of the unit, the load rate of each unit, the cold water supply temperature of each unit, the chilled water pump frequency corresponding to each unit, and the control frequency setting value of each cooling tower fan, so as to enable the cold source system to operate in a more efficient range while ensuring the cooling effect, thereby improving the system operation energy efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of building cooling system control, in particular to a control method for a building cooling system. Background Art

[0002] Buildings play a significant role in global energy use and carbon dioxide emissions, accounting for one-third of the world's energy use and one-quarter of its carbon dioxide emissions. Since HVAC (heating, ventilation, and air conditioning) systems account for 38% of a building's energy consumption, the potential for energy savings in buildings is enormous. A building cooling system is a system designed to meet a building's air conditioning needs and consists of chillers, distribution pipelines, heat exchange equipment, and a cooling control system. Cooling equipment in a building cooling system includes air-cooled chillers, ground-source heat pumps, and direct expansion chillers. Distribution pipelines include supply pipes, return pipes, and branch pipes. Heat exchange equipment includes radiant plate heat exchangers and fan coil units. The cooling control system is responsible for controlling the operation of the building cooling system, including cooling load calculation, temperature control, and water flow control.

[0003] With the advancement of building energy management and intelligent technologies, the development of cooling source group control algorithms has evolved through the following stages: Initial Stage: Cooling source group control algorithms first appeared in the 1980s, primarily to address energy waste and comfort issues in traditional cooling systems. This stage primarily relied on experience and rules-based approaches, with control strategies primarily consisting of fixed schemes based on factors such as season, time of day, and building type. Mid-Stage: With advances in computer technology and the application of intelligent control techniques, cooling source group control algorithms have gradually evolved towards intelligent and adaptive approaches. This stage primarily employed cooling source group control algorithms based on technologies such as model predictive control and artificial neural network control. These control strategies dynamically adjusted based on real-time cooling load demand, achieving efficient energy utilization and improved comfort. Modern Stage: With the widespread application of technologies such as big data, cloud computing, and the Internet of Things, cooling source group control algorithms have further evolved towards intelligent and networked approaches. This stage primarily employed cooling source group control algorithms based on cloud computing platforms and the Internet of Things, enabling remote monitoring and control of cooling source systems and improving their operational efficiency and stability.

[0004] At present, the cold source group control algorithm still has some shortcomings, which are manifested in the following aspects:

[0005] 1. Inaccurate load forecasting: The core of the cold source group control algorithm is to dynamically adjust according to the real-time cooling load demand. However, there are errors in the prediction of cooling load, resulting in unsatisfactory control effects of the cold source system. 2. Dependence on control strategy: The control effect of the cold source group control algorithm depends largely on the formulation of the control strategy. If the control strategy is unreasonable or not applicable to the specific building, it will lead to poor control effects. 3. High system complexity: The cold source group control algorithm needs to establish a model of the cold source system, collect a large amount of real-time data, and perform complex calculations and controls, resulting in high system complexity and difficulty in implementation.

[0006] Therefore, further research and improvement are needed to improve the application effect and stability of the technology. Summary of the Invention

[0007] The purpose of the present invention is to provide a control method for a building cooling system in order to solve the problems in the prior art.

[0008] A method for controlling a cooling system of a building, comprising the steps of:

[0009] Based on the load measurement value of the current unit combination at time τ-Δτ, the load forecast value at time τ is predicted:

[0010] Based on the relationship between the load forecast value and the load value of the current unit combination, forming a plurality of selectable unit combinations;

[0011] Optimize and calculate the operating parameters of each unit combination to obtain the operating optimization parameters of each unit combination;

[0012] The comprehensive operating power of each unit combination is calculated based on the load rate of the operation optimization parameters. According to the comprehensive operating power, a unit combination is selected as the optimization result, and an operation strategy composed of the corresponding operation optimization parameters is output.

[0013] The method of predicting the load forecast value at time τ based on the load measurement value at time τ-Δτ of the current unit combination includes:

[0014] Based on the load measurement value E at the time τ-Δτ of the current unit combination d,τ-Δτ , get the deterministic load forecast value at time τ and the random load measurement value X at time τ-Δτ d,τ-Δτ ;

[0015] Execute the weighted recursive algorithm objective function FFRLS and correct the random load autoregressive coefficient Then the random load autoregressive model is executed to obtain the random load forecast value at time τ

[0016] The deterministic load forecast value and random load forecast value The sum is the load forecast value at time τ

[0017] The deterministic load forecast value It is obtained by the deterministic model based on the single moving average algorithm, and the expression is as follows:

[0018]

[0019] Where, is the deterministic prediction value at time τ-Δτ, E d,τ-Δτ is the load measurement at time τ-Δτ, λ is the exponential smoothing coefficient;

[0020] Load measurement value E of unit combination at time τ d,τ It is calculated based on the real-time cooling capacity of the chiller system and the indoor temperature monitoring value, and is expressed as follows:

[0021]

[0022] Where, ρ water C p,water is the specific gravity of water, which is given by the density of frozen water ρ water Specific heat capacity C with water p,water Multiply to obtain; is the chilled water flow at time τ-1; T chws and T chwr are the cold water supply temperature and return water temperature at time τ-1; V B is the building cooling volume, which is the product of the cooling area of each floor of the building and the floor height; and h n,des are the detected value and design value of indoor air enthalpy at time τ, t τ is the building cooling adjustment time; the random load forecast value Described as the random load measurement value X for the previous n moments τ-jΔτ and error value The error value is the random load measurement value X of the first n moments. τ-jΔτ The corresponding random load forecast value of the previous n moments The difference is expressed as follows:

[0023]

[0024] The objective function of the weighted recursive algorithm FFRLS is:

[0025]

[0026] α is the forgetting factor. If the intermediate transfer matrix P is non-positive, the latest load data is used to retrain the load forecasting model composed of the deterministic model and the random load autoregressive model. Otherwise, the load measurement value E at time τ is obtained. d,τ After the training data is constructed, the deterministic model is executed to obtain the random load measurement value X of the first n moments, which is the same as the number of training data. τ-jΔτ , then the least squares method is used to initialize the random load autoregressive coefficient Calculate random load forecast values

[0027] The multiple unit combinations are formed based on the current unit combination, including corresponding unit combinations obtained by adding, reducing or replacing a currently running chiller and maintaining the addition and subtraction strategies of the currently running chillers on the current unit combination. The unit combinations obtained by the addition, reduction and replacement operations are not unique. For example, the number of unit combinations obtained by the addition operation should be consistent with the number of chillers that are not currently turned on. The unit combinations formed can be the first unit combination formed by adding a unit on the current unit combination, the second unit combination formed by subtracting a unit on the current unit combination, the third unit combination formed by replacing a unit in the current unit combination and keeping the unit operation control strategy unchanged, and the fourth unit combination by keeping the operation control strategy of the current unit combination.

[0028] Each of the multiple selectable unit combinations satisfies the following conditions:

[0029] The load prediction value is less than the product of the total load of the unit combination and the maximum load rate. At the same time, if the number of units in the unit combination is greater than 1, the load prediction value is greater than the product of the rated load of the smallest unit in the unit combination and the minimum load rate.

[0030] The operation optimization parameters include the load rate PLR of each unit in the unit combination. ch,i , each unit cold water supply temperature setting value T chws , the water pump frequency setting value f corresponding to each unit set,i , the cooling tower fan frequency setting value f corresponding to each unit twr,i .

[0031] Among them, the cold water supply temperature setting value T of each unit combination chws,i The optimization is calculated by:

[0032] The cold water supply temperature setting value T of each unit combination chws and indoor wet bulb temperature T n,wb Difference T n,wb -T chws, satisfies a linear relationship with the total load rate PLR, namely:

[0033]

[0034]

[0035]

[0036] Among them, T n,wb,des Indicates the indoor wet-bulb temperature set point, T chws,des Indicates the unit's cold water supply temperature setting value, T chws,min Indicates the minimum temperature of the unit’s cold water supply, T chws,max Indicates the maximum temperature of the unit's cold water supply, P ch ,des is the rated power of the chiller, P chwp,des is the rated power of the chilled water pump, S chws is the sensitivity coefficient of the chiller operating power to the cooling water temperature. For a single unit, it can be obtained for any unit i by the following formula:

[0037]

[0038] Where, P ch,i Indicates the operating power of the chiller, T chws,i represents the water supply temperature of chiller i, T cwr,set Indicates the cooling water return temperature setting value, PLR ch,set Indicates the chiller operating power setting value, P rated,i Indicates the rated power of chiller i, coefficient a i ,b i From the performance characteristic equation of chiller i;

[0039] If there are multiple units operating in combination, S chws Perform weighted calculation using the rated power of each chiller as the weight;

[0040] Where, P ch,des,i It represents the rated power of chiller i under design conditions;

[0041] The chiller performance characteristic equation is the unit power P ch Described as the cold water supply temperature T chws , Cooling water return temperature T cwr , the total load rate PLR of the unit, the operating power of a chiller is expressed as follows:

[0042]

[0043] The cold water supply temperature setting value T of each unit is obtained through the linear relationship chws,i .

[0044] Among them, the water pump frequency setting value f set,i The optimization is calculated by the following formula:

[0045]

[0046]

[0047] Where, is the chilled water supply flow rate set value at time τ-1, represents the chilled water pump frequency at time τ-1; Respectively represent the rated load of each chiller under the current optimized unit combination and the rated load of each chiller under the unit combination at the previous moment. Finally, the pump frequency setting value f is obtained by the following method set,i , Chilled water supply flow setting value:

[0048]

[0049]

[0050]

[0051] Where, f set,min 、f set,max Respectively represent the minimum and maximum values of the pump frequency setting value, ΔT set Indicates the set value of the chilled water supply and return temperature difference, f max Indicates the maximum frequency of the pump.

[0052] Among them, the load rate PLR of each unit combination ch,i The optimization is achieved by the following method:

[0053] For the unit combination, the optimization function of unit load and chilled water supply temperature is described as:

[0054]

[0055] Where Q rated,i represents the rated load of chiller i, T chws,min Represents the lower limit of the chiller water supply temperature set value. The equality constraint matrix equation is obtained by the unit load and chilled water supply temperature optimization function. The equality constraint matrix equation is solved to obtain the optimization result X. The optimization result X includes the load rate PLR of each unit. ch,i and cold water supply temperature T chws,i, i is the chiller number: The equality constraint matrix equation is as follows:

[0056]

[0057] Monitor the calculation results of X. If the inequality constraint is not satisfied, the inequality constraint is transformed into an equality constraint by constructing an augmented matrix, and then solve Until the calculation results meet all inequality constraints;

[0058]

[0059] Wherein, the cooling tower fan control frequency f twr,i Based on the load factor PLR of each unit combination ch,i The calculation optimization results are:

[0060]

[0061]

[0062]

[0063]

[0064] Where, f twr,max Indicates the maximum power of the cooling tower fan, a twr,des is the difference between the air wet bulb temperature and the cooling water supply temperature under design conditions; r twr,des is the cooling water supply and return temperature difference under design conditions; P ch,des is the rated power of the chiller; P twr,des is the rated power of the cooling tower fan; S cwr Is the sensitivity coefficient of the chiller operating power to the cooling water temperature. If the operating condition of a single unit i is used, the expression of the sensitivity coefficient of the chiller operating power to the cooling water temperature is as follows:

[0065]

[0066] Where a i ,b i The coefficient comes from the chiller performance characteristic equation, T cwr,i is the cooling water return temperature of chiller i; T chws,i is the chilled water supply temperature of chiller i;

[0067] If there are multiple units operating in combination, the sensitivity coefficient S of the operating power of multiple chillers to the cooling water temperature is cwr,des Perform weighted calculation using the rated power of each chiller as the weight;

[0068]

[0069] The step of calculating the comprehensive operating power of each unit combination based on the load rate of each unit combination and selecting a unit combination as an optimization result according to the comprehensive operating power includes:

[0070] Based on the load factor PLR of each unit combination ch,i , get the comprehensive operating power P of each unit combination;

[0071] Compare the lowest unit comprehensive operating power P min The combined operating power of the unit combined with the running unit like The unit combination with the lowest power is selected as the optimization result; otherwise, the unit combination that is in operation is selected as the optimization result.

[0072] The present invention realizes the group control of the chillers of the primary pump (variable frequency pump) system, and quickly responds to the changing trend of the building load according to the real-time changes in the heat demand of the building, so as to ensure the heat demand of the terminal heat exchange equipment and the thermal comfort of the indoor environment; it can efficiently identify the performance characteristics of each chiller, and can track the characteristic distortion of the chiller during operation to make it more in line with the actual equipment performance; with the comprehensive energy efficiency of the chiller as the objective function, the operation strategy of the cold source system composed of the start and stop of the unit, the load rate of each unit, the cold water supply temperature of each unit, the water pump frequency setting value corresponding to each unit, and the control frequency setting value of each cooling tower fan is comprehensively optimized, so as to ensure the operation of the cold source system in a more efficient range and improve the system operation energy efficiency under the premise of ensuring the cooling effect.

[0073] The present invention dynamically adjusts the operating mode and parameters of the cold source system based on the real-time changes in the building's cooling demand, thereby achieving efficient energy utilization, reducing energy consumption and reducing the emission of greenhouse gases such as carbon dioxide, thereby achieving the goal of energy conservation and emission reduction; by optimizing the operation of the cold source system, the indoor temperature control accuracy of the building can be improved, temperature fluctuations can be reduced, thermal comfort can be improved, and the quality of the living and working environment can be improved; by optimizing the operation of the cold source system, the system stability and reliability can be improved, the system failure rate can be reduced, and the maintenance and replacement costs can be reduced, thereby reducing the maintenance cost of the building.

[0074] The present invention is suitable for the operation scenario of one machine, one pump and one tower unit connection, requiring real-time monitoring of water temperature, frequency, chiller power, indoor and outdoor temperature and humidity status data. It is applicable to all chiller equipment that can achieve self-regulation of the unit by resetting the cold water supply temperature. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] Figure 1 This is a flow chart of a control method for a building cooling system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0076] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in combination with the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments.

[0077] The control method of the building cooling source system in the embodiment of the present invention uses the real-time supply and return water temperature difference and indoor temperature and humidity status monitoring to adjust the load in real time, realizes load prediction by performing time series analysis on historical load measurement values, and understands the historical load change trend. At the same time, it quickly responds to changes in indoor temperature and humidity status, predicts the operating load of the chiller, and guarantees the heating needs of the terminal equipment.

[0078] By optimizing the chiller, coupling optimization is performed on the chilled water supply temperature, chilled water pump frequency, chiller chilled water supply temperature set value, and chiller load distribution of various unit combinations, giving the optimal chiller supply temperature set value and chilled water pump frequency set value, realizing comparison and selection of multiple unit combinations, determining the optimal operating strategy, and outputting the start and stop control instructions of the chiller equipment, chiller supply temperature set value, and water pump frequency set value. The chiller optimization includes: a) under a specific total load, selecting multiple chiller addition and subtraction strategies including the current operating strategy, and for a specific chiller combination, determining the chilled water supply temperature based on the unit power characteristics and indoor temperature and humidity conditions to ensure that the supply temperature meets the indoor cooling demand; b) coupling optimization is performed on the chilled water pump operating frequency, chiller chilled water supply temperature set value, and load distribution of each chiller under a specific chiller combination to obtain the optimal strategy for the chiller combination. Compare various operating strategies to determine the optimal operating strategy with the goal of ensuring smooth operation of the unit and reducing the energy consumption of the system's cooling machine operation.

[0079] By optimizing the cooling tower frequency, based on the optimal cooling machine optimization strategy and unit load distribution, the optimal real-time operating frequency of the cooling tower fan is determined, and the cooling tower fan frequency set value is output. The energy consumption on the cooling water side is reduced while ensuring the cooling effect of the cooling water, and the cooling tower fan operating frequency is optimized.

[0080] See also Figure 1 As shown, the control method of the building cooling system according to the embodiment of the present invention includes the following steps:

[0081] S1. Based on the load measurement value of the current unit combination at time τ-Δτ, the load forecast value at time τ is predicted:

[0082] S2. Based on the load forecast value and the calculated load value of the current unit combination, a plurality of optional unit combinations are formed;

[0083] S3. Optimize the operating parameters of each unit combination to obtain the optimal operating parameters of each unit combination;

[0084] S4. Calculate the comprehensive operating power of each unit combination based on the load rate in the operation optimization parameters, select a unit combination as the optimization result according to the comprehensive operating power, and output the operation strategy corresponding to the operation optimization parameters.

[0085] In step S1, based on the load measurement value of the current unit combination at the time τ-Δτ, the load forecast value at the time τ can be predicted by a load forecast model. The load forecast model consists of a deterministic model and an autoregressive model. The deterministic model gives the deterministic load forecast value at the time τ. The autoregressive model gives the random load forecast value at time τ Load forecast value at time τ Deterministic load forecast value and random load forecast values The expression is as follows:

[0086]

[0087] Among them, the deterministic load forecast value in the load forecast value It is obtained through a moving average algorithm as follows:

[0088]

[0089] in is the deterministic load forecast value at time τ-Δτ, E d,τ-Δτ is the load measurement value at time τ-Δτ, λ is the exponential smoothing coefficient, λ = 0.1.

[0090] Among them, the load measurement value E at time τ d,τ It is calculated based on the real-time cooling capacity of the chiller system and the indoor temperature monitoring value, namely:

[0091]

[0092] The first part on the right side of the equation is the cooling capacity of the chiller system, ρ water C p,water is the specific gravity of water (kj / (m 3 K)); is the chilled water flow rate at time τ-1 (m 3 / h); T chws and T chwr are the chilled water supply and return water temperatures at time τ-1 (℃). The second part is the evaluation of indoor cooling capacity, which is achieved by comparing the indoor air enthalpy value with the design working condition, where V BIt is the cooling volume of the building, which is the product of the cooling area of each floor of the building and the floor height (m 3 ); and h n,des are the detected value and design value of indoor air enthalpy at time τ, which can be calculated by the corresponding indoor temperature / humidity values T and RH; t τ It is the building cooling adjustment time, generally 3600s;

[0093] Among them, the random load forecast value at time τ in the load forecast value is Obtained through the autoregressive model, the random load forecast value at time τ in the autoregressive model Described as the random load measurement value X at time τ-jΔτ τ-jΔτ and its error value The linear combination of is the random load forecast value at time τ-jΔτ, and the autoregressive coefficient Calculate using weighted recursive least squares FFRLS:

[0094]

[0095] The objective function of the weighted recursive algorithm FFRLS is set as:

[0096]

[0097] Using the weighted recursive least squares method FFRLS to recursively correct the random load autoregressive coefficient φ i , the forgetting factor α can be set to 0.995. In addition, due to the FFRLS algorithm, if the intermediate transfer matrix P is not positive definite, it will lead to coefficient It is getting smaller and smaller, so it is necessary to monitor the positive definiteness of the matrix P. If the transfer matrix P is not positive definite, the load forecasting model is retrained using the latest load data.

[0098] During the load forecasting model pre-training process, when sufficient random load forecast values E at time τ are obtained as training data, d,τ After that (generally 3 to 7 days of operation data of the building cooling system, the period is 15 to 30 minutes), by executing the deterministic model, the random load measurement value X at the time τ-jΔτ, which is the same as the number of training data, is obtained. τ-jΔτ As a random load measurement sample. Then, the least squares method initializes the random load autoregressive coefficient

[0099] After the pre-training is completed, the load forecasting model is based on the updated load measurement value E of the chiller at the time τ-Δτ d,τ-Δτ , first execute the deterministic model to obtain the deterministic load forecast value at time τ and the random load measurement value X at time τ-Δτ d,τ-Δτ , then perform weighted recursive least squares FFRLS to correct the random load autoregressive coefficient Finally, the random load autoregressive model is executed to obtain the random load forecast value at time τ Finally, the load forecast value at time τ is obtained

[0100] In the embodiment of the present invention, in step S2, the plurality of selectable chiller combinations are formed based on the load prediction value and the calculated load of the current chiller combination, specifically including:

[0101] Based on the current operation combination of the chiller, a variety of operation combinations of the chiller are proposed according to the following four processes + two requirements.

[0102] a) Reduction of units: If the historical load is gradually decreasing and the number of units in operation is greater than 1, reduction of units is allowed;

[0103] b) Maintain: Maintain the control strategy of the original unit (i.e. the currently operating unit);

[0104] c) Replacement: Under the original unit operation control strategy, replace one of the currently operating units in the unit combination;

[0105] d) Adding a unit: If the historical load is gradually increasing and there are units that have not yet been operated, adding a unit is allowed;

[0106] In addition, the proposed chiller operation combination must meet the following requirements:

[0107] a) Load forecast value <Total load of the unit combination*maximum load rate of the chiller PLR max ;

[0108] b) If the number of units is > 1, the load forecast value is required > Rated load of the smallest unit in the unit combination × minimum load rate (PLR) of the chiller min .

[0109] In the embodiments of the present invention, the units and chillers are synonymous.

[0110] In the embodiment of the present invention, the process of optimizing the chiller operation strategy is as follows:

[0111] 3.1. Initialize / update the chiller performance characteristic equation

[0112] In the present invention, the unit power P ch,i Described as the cold water supply temperature T chws,i, Cooling water return temperature T cwr,i , unit load factor PLR ch,i The function is:

[0113]

[0114] The correction coefficient a is recursively calculated using the weighted recursive least squares method FFRLS. i ,b i , the forgetting factor α can be set to 0.995. Where the subscript i is used to refer to the parameters of the i-th chiller.

[0115] 3.2. Optimization of chiller group control strategy

[0116] Based on multiple different unit operation combinations, the operation strategy of each unit operation combination is optimized.

[0117] 3.2.1. Determine the cold water supply temperature:

[0118] The chiller's cold water supply temperature T chws,i and indoor wet bulb temperature T n,wb Difference T n,wb -T chws , satisfies a linear relationship with the total load rate PLR of the unit, namely:

[0119]

[0120] Among them, P ch,des is the chiller rated power, P twr,des is the rated power of the cooling tower fan, S chws is the sensitivity coefficient of the chiller operating power to the cooling water temperature;

[0121] The sensitivity coefficient of the operating power of a single unit to the cooling water temperature can be obtained by the following formula: i ,b i Derived from the chiller performance characteristic equation.

[0122]

[0123] If there are multiple units operating in combination, the sensitivity coefficient S of the operating power of multiple units to the cooling water temperature is chws Weighted calculation is performed using the rated power of the unit as the weight:

[0124]

[0125] 3.2.2、Optimize the water pump distribution parameters (including the water pump frequency setting value f set,i and cold water flow setting value );

[0126] For a unit combination, the regulation ratio of its pump frequency is defined as:

[0127] r f,i is the pump frequency setting value f set,i and the pump frequency at time τ-1 The ratio of

[0128] Pump frequency setting value f set,i and the pump frequency at time τ-1 The ratio r f,i , is positively correlated with the front-to-back load ratio of the unit combination and the total rated load of the unit combination. The ratio r f,i The expression is as follows;

[0129]

[0130] By the ratio r f,i Calculate the pump frequency setting value f set,i , cold water flow setting value If the pump frequency does not exist at time τ-1 in the adding machine condition Cold water flow set value Then the assumed value is given according to the preset method;

[0131] The pump frequency f is calculated by the following expression set,i , cold water flow setting value at time τ Optimization:

[0132]

[0133]

[0134]

[0135] Through the above processing, the water pump frequency setting value f can be obtained set,i and cold water flow setting value

[0136] 3.2.3、Chiller load distribution optimization (optimization of unit load rate PLR ch,i and the cold water supply temperature set value T chws,i ):

[0137] For a certain unit combination, the load and chilled water supply temperature optimization function of the unit combination is described as:

[0138]

[0139] The equality constraint matrix equation is obtained from the above function. Solving the equality constraint equation gives the optimization result X. The 3i and 3i+1 items in X (i is the chiller unit number) are the load rates PLR of each unit. ch,i and cold water supply temperature T chws,i :

[0140]

[0141] Monitor the calculation results of X. If the inequality constraint is not satisfied, then construct an augmented matrix to transform the inequality constraint into an equality constraint and then solve Until the calculation results meet all inequality constraints. The augmented matrix is as follows:

[0142]

[0143]

[0144] The inequality constraints are transformed into equality constraints as follows:

[0145]

[0146] The above method can be used to optimize the load rate PLR of the unit combination ch,i and the cold water supply temperature set value T chws,i .

[0147] 3.2.4 Optimize cooling tower fan control parameters (optimize cooling tower fan control frequency setting value f twr,i );

[0148] Load factor (PLR) based on unit commitment ch,i , the cooling tower fan control frequency f is obtained by the following calculation twr,i :

[0149]

[0150]

[0151]

[0152]

[0153] Where a twr,des is the difference between the air wet bulb temperature and the cooling water supply temperature under design conditions; r twr,des is the cooling water supply and return temperature difference under design conditions; P ch,des is the rated power of the chiller; P twr,des is the rated power of the cooling tower fan; S cwr is the sensitivity coefficient of the chiller operating power to the cooling water temperature;

[0154] If facing the single unit working condition, for any chiller i, the coefficient a can be obtained by the following formula. i ,b i Derived from the chiller performance characteristic equation.

[0155]

[0156] If there are multiple units operating in combination, the sensitivity coefficient S of the operating power of multiple chillers to the cooling water temperature is cwr,des The weighted calculation should be performed using the rated power of the unit as the weight.

[0157]

[0158] Through the above processing, the cooling tower fan control frequency setting value f can be optimized twr,i .

[0159] In the embodiment of the present invention, the load rate PLR of each unit combination is obtained based on the above ch,i , the comprehensive operating power of the chiller of each unit combination is obtained by the following formula.

[0160]

[0161] Compare the lowest unit comprehensive operating power P in the unit combination min The combined operating power of the running units like The output operation strategy is based on the lowest power unit combination; otherwise, the output operation strategy is based on the running unit combination. The output operation strategy includes the chiller load rate PLR ch,i , cold water supply temperature setting value T chws,i , Pump frequency setting value f set,i , Cooling tower fan control frequency setting value f twr,i The above is only a preferred embodiment of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be considered as the scope of protection of the present invention.

Claims

1. A method for controlling a building cooling system, characterized in that: Including steps: Based on the current unit combination Load measurement value at the moment, prediction Load forecast value at the moment: Based on the relationship between the load forecast value and the load value of the current unit combination, forming a plurality of selectable unit combinations; Optimize and calculate the operating parameters of each unit combination to obtain the operating optimization parameters of each unit combination; Calculating the comprehensive operating power of each unit combination based on the load factor of the operation optimization parameters, selecting a unit combination as the optimization result based on the comprehensive operating power, and outputting an operation strategy composed of the corresponding operation optimization parameters; Based on the current unit commitment Load measurement value at the moment, prediction The load forecast value at the moment includes: Based on the current unit combination Load measurement value at the moment ,get Deterministic load forecast value at the moment and Random load measurement value at time ; Execute the weighted recursive algorithm objective function FFRLS and correct the random load autoregressive coefficient , and then perform the random load autoregressive model to obtain Random load forecast value at each moment ; The deterministic load forecast value and random load forecast value The sum is Load forecast value at the moment ; The deterministic load forecast value It is obtained by the deterministic model based on the single moving average algorithm, and the expression is as follows: ； Where, yes The deterministic prediction value at time yes The load measurement value at the time, is the exponential smoothing coefficient; Among them, the unit combination Load measurement value at each moment It is calculated based on the real-time cooling capacity of the unit combination and the indoor temperature monitoring value, and is expressed as follows: ； Where, is the specific gravity of water, given by the density of frozen water Specific heat capacity with water Multiply to obtain; yes Chilled water flow at all times; and They are Cold water supply temperature and return water temperature at all times; is the building cooling volume, which is the product of the cooling area of each floor of the building and the floor height; and The indoor air enthalpy The detection value and design value at the moment, It is the building cooling adjustment time; The random load forecast value Described as Random load measurement value at time and its error value The error value is a linear combination of Random load measurement value at time and the corresponding Random load forecast value at each moment The difference is expressed as follows: ； The objective function of the weighted recursive algorithm FFRLS is: ； is the forgetting factor. If the intermediate transfer matrix P is non-positive, the latest load data is used to retrain the load forecasting model composed of the deterministic model and the random load autoregressive model. Otherwise, Load measurement value at each moment After the training data is constructed, the deterministic model is executed to obtain the random load measurement value at the corresponding time with the same number of training data. , then the least squares method is used to initialize the random load autoregressive coefficient Calculate random load forecast values .

2. The control method of a building cooling system according to claim 1, characterized in that: The multiple unit combinations are formed based on the current unit combination by adding, reducing or replacing a currently running chiller and maintaining the addition and subtraction strategy of the currently running chiller. They are respectively a first unit combination formed by adding a unit to the current unit combination, a second unit combination formed by subtracting a unit from the current unit combination, a third unit combination formed by replacing a unit in the current unit combination and keeping the unit operation control strategy unchanged, and a fourth unit combination as the operation control strategy of the current unit combination is maintained.

3. The control method of a building cooling system according to claim 1, characterized in that: Each of the multiple selectable unit combinations meets the following conditions: The load prediction value is less than the product of the total load of the unit combination and the maximum load rate. At the same time, if the number of units in the unit combination is greater than 1, the load prediction value is greater than the product of the rated load of the smallest unit in the unit combination and the minimum load rate.

4. The control method of a building cooling system according to claim 1, characterized in that: The operation optimization parameters include the load rate of the unit combination , chilled water supply temperature setpoint of the unit combination , water pump frequency setting value of the unit combination , Cooling tower fan frequency setting value of unit combination .

5. The control method of a building cooling system according to claim 4, characterized in that: The cold water supply temperature of the unit combination Determined by: Chilled water supply temperature for each chiller combination and indoor wet-bulb temperature The difference between the load factor and the unit combination Satisfies the linear relationship, that is: ； in, Indicates the indoor wet-bulb temperature setting value under design conditions. Indicates the set value of cold water supply temperature under design conditions. Indicates the minimum cold water supply temperature. Indicates the maximum cold water supply temperature. is the chiller rated power, is the rated power of the chilled water pump; is the sensitivity coefficient of the chiller operating power to the cooling water temperature; The sensitivity coefficient of a single chiller is obtained by the following formula: ； Where, Indicates the operating power of the chiller. Indicates the cooling water return temperature setting value. Indicates the chiller operating power setting value, Indicates the rated power of the chiller, coefficient From the chiller performance characteristic equation; If there are multiple units operating in combination, the sensitivity coefficient of the chiller operating power to the cooling water temperature is Perform weighted calculation using the rated power of the unit as the weight; , Indicates the rated power of the chiller under design conditions; The chiller performance characteristic equation converts the chiller operating power Described as cold water supply temperature , cooling water return temperature , total unit load rate As a function of , the operating power of a chiller is expressed as follows: ； The water supply temperature setting value is obtained through the linear relationship .

6. The control method of a building cooling system according to claim 5, characterized in that: The water pump frequency setting value The optimization is calculated by the following formula: ； ； Where, for The chilled water supply flow setting value at the moment, express The chilled water pump frequency at the moment; 、 Respectively represent the rated load of each chiller in the unit combination and the rated load of each chiller under the unit combination at the previous moment; when adding a machine, there is no 、 :Get the pump frequency setting value through the following formula 、 Chilled water supply flow setting value at the moment , 、 Respectively represent the minimum and maximum values of the pump frequency setting value. Indicates the maximum frequency of the pump, Indicates the set value of the chilled water supply and return temperature difference: 。 7. The control method of a building cooling system according to claim 6, characterized in that: Load factor of the unit combination , cold water supply temperature setting value Optimized by: Load factor of unit combination , cold water supply temperature setting value The optimization function is: ； Where, represents the rated load of chiller i, Represents the lower limit of the chiller water supply temperature set value; the equality constraint matrix equation is obtained from the optimization function, and the optimization result is obtained by solving the equality constraint matrix equation , optimization results The load factor of each unit combination is included and cold water supply temperature setpoint , i is the chiller number: The equality constraint matrix equation is as follows: ； monitor If the calculation result does not satisfy the inequality constraint, the inequality constraint is transformed into an equality constraint by constructing an augmented matrix, and then the solution is , until the calculation results meet all inequality constraints; 。 8. The control method of a building cooling system according to claim 7, characterized in that: The cooling tower fan control frequency Based on the load factor of each unit combination The calculation optimization results are: ; Where, Indicates the maximum power of the cooling tower fan. It is the difference between the air wet bulb temperature and the cooling water supply temperature under design conditions; It is the temperature difference between the supply and return water of cooling water under the design working conditions; is the chiller rated power; is the cooling tower fan rated power; is the sensitivity coefficient of the chiller operating power to the cooling water temperature. The sensitivity coefficient of a single chiller is obtained by the following formula: ； Where, The coefficients come from the chiller performance characteristic equation, is the cooling water return temperature of chiller i, is the chilled water supply temperature of chiller i; If there are multiple units operating in combination, the sensitivity coefficient of the operating power of multiple chillers to the cooling water temperature is Perform weighted calculation using the rated power of the unit as the weight; 。 9. The control method for a building cooling system according to claim 8, characterized in that: The comprehensive operating power of each unit combination is calculated based on the load factor of each unit combination. According to the comprehensive operating power, a unit combination is selected as the optimization result, including: Based on the load factor of each unit combination , get the comprehensive operating power of each unit combination ; ; Compare the lowest unit comprehensive operating power in the unit combination The combined operating power of the unit combined with the running unit ,like , the unit combination with the lowest power is selected as the optimization result; otherwise, the unit combination in operation is selected as the optimization result.

Images