CN110107989A - Small-sized based on chilled water return water temperature optimum set point determines frequency water cooler and becomes temperature control method of water - Google Patents

Abstract

The invention relates to a variable water temperature control method for a small fixed-frequency chiller based on the optimal set point of the return temperature of chilled water. First, a model of the air-conditioning system of the small fixed-frequency chiller is established for simulation. Then collect the design parameters of the building and the typical annual meteorological data in the area, and combine the above three to calculate the maximum allowable temperature of the chilled water return under different operating conditions. Then check and correct the maximum allowable return temperature of chilled water to determine the optimal set point of return water temperature and establish different operating condition data sets. Based on the data set, a GRNN prediction model for the optimal set point of the return water temperature of the fixed-frequency chiller was established and verified. Finally, the optimal set point of the return water temperature of the unit is determined with the help of the prediction model for timely adjustment. Finally, the position controller is used to control the start and stop of the chiller according to the difference between the return water temperature and the optimal set water temperature. The invention can improve the efficiency of a small fixed-frequency chiller on the premise of ensuring indoor thermal comfort, and achieve better energy-saving effects.

Description

Variable water temperature of small fixed-frequency chiller based on optimal set point of chilled water return temperature Control Method

technical field

The invention relates to a small fixed-frequency chiller, in particular to an energy-saving method for controlling the return water temperature of the small fixed-frequency chiller. Small fixed-frequency chillers are units with a cooling capacity of no more than 50Kw.

Background technique

At present, small fixed-frequency chillers are widely used in air-conditioning systems of small and medium-sized buildings in my country due to their advantages such as stable cooling performance, simple control system, and low price. The small fixed-frequency chiller does not have the ability to adjust the load. In order to adapt to the dynamic load change of the air conditioning system, the unit is controlled to start and stop according to the return water temperature. The return water temperature control of the existing small fixed-frequency chillers usually adopts the constant water temperature control method, that is, the fixed return water temperature set point (usually the design temperature of the return water is 12°C). The fixed water temperature control method can meet the cooling demand of the building under the most unfavorable working conditions, but in actual operation of the building, the time of the most unfavorable working conditions is usually less than 10%, and most of the time is in the partial load condition. The fixed water temperature control method makes the air conditioning system Chilled water temperature is too low, which is not conducive to the energy-saving operation of small fixed-frequency chillers, resulting in waste of energy.

The variable water temperature control method is an energy-saving adjustment strategy that automatically adjusts the chilled water temperature set point of the chiller according to the requirements of the building conditions, and realizes the variable water temperature operation of the chiller. According to data provided by chiller manufacturers, for every 1°C increase in chilled water temperature, the unit coefficient of performance (COP) can increase by 2.0% to 4.0%. Therefore, the chilled water temperature can be increased as much as possible to improve the performance of the chiller. In the actual operation and research of small fixed-frequency chillers, some explorations have been made on variable water temperature control, mainly in the following two ways: 1. Manually adjust the return water temperature setting value based on experience; 2. Divide the cooling season into several 3. Divide the outdoor temperature into several stages, and give the set temperature of chilled water supply in stages. The existing method needs to rely on experience and can only be determined through repeated tests. On the one hand, it cannot effectively ensure a comfortable indoor thermal environment; on the other hand, it cannot dynamically adjust the water temperature set point according to the actual operating conditions, and it cannot maximize the operating efficiency of the chiller.

According to the above problems, on the premise of ensuring the comfort of the indoor thermal environment of the building, and taking the best operating performance of the unit as the optimization goal, a real-time prediction of the optimal set point of the chilled water return temperature of the unit based on the actual operating conditions was invented to realize the small fixed A variable water temperature operation control method for frequency chillers.

Contents of the invention

On the premise of ensuring the indoor comfort of the building, the present invention establishes the prediction model of the optimal set point of the return water temperature based on GRNN, predicts the optimal set point of the chilled water return temperature of the unit in real time according to the actual operating conditions, and adjusts the water temperature setting in a timely manner value, so as to realize the variable water temperature optimal operation of the air-conditioning system of the small fixed-frequency chiller. The technical scheme adopted by the present invention to solve its technical problems is: firstly, establish a model of the air-conditioning system of the small-scale fixed-frequency chiller, then carry out simulation with the help of the model and collect the design parameters of the building and the typical annual meteorological data of the area, and calculate the difference The maximum allowable temperature of chilled water return water based on thermal comfort under operating conditions. Check and correct the maximum allowable return temperature of chilled water to determine the optimal set point of return water temperature, establish data sets of different operating conditions, and establish and verify the GRNN prediction of the optimal set point of return water temperature of fixed-frequency chillers based on the data sets Model. Use the prediction model of the best set point of return water temperature to predict the best set point of return water temperature under the current working conditions in real time, adjust the return water temperature of small fixed-frequency chillers in a timely manner, and then use the position controller to adjust the return water temperature according to the return water temperature. The difference from the optimal set water temperature controls the start and stop of the fixed frequency chiller.

(1) Establish a simulation model. Firstly, sub-models such as room model, fixed-frequency chiller model, fan coil model, and control model are established in the simulation software Modelica according to the architectural design parameters, and the dynamic operation process of the air-conditioning system is simulated through the joint operation of the sub-models.

(2) Calculate the maximum allowable temperature of the chilled return water of the chiller. Through the simulation method, the air-conditioning terminal equipment is set to work at a fixed air volume and rated flow rate, and the indoor comfort of the air-conditioning room is guaranteed by adjusting the temperature of the chilled water. Using the typical annual meteorological data in the area where the building is located, different operating conditions can be calculated. Below, the maximum allowable return temperature of chilled water based on thermal comfort.

(3) Determine the optimal set point of return water temperature and establish a data set. First, set the maximum allowable value of the return water temperature as the return water temperature of the chiller. Through simulation, check whether the return water temperature under this working condition meets the indoor comfort requirements. If so, the maximum allowable value of the return water temperature under this working condition The optimum set point for the return water temperature. If the indoor temperature is not satisfied, the set value needs to be corrected. According to the above method, the optimal set point of the return water temperature of the small fixed-frequency chiller under different working conditions is determined. Based on the optimal set point of the return water temperature determined above and the operating condition data, a data set is established.

(4) Establish and verify the optimal return water temperature set point GRNN prediction model. Firstly, the input and output parameters of the prediction model are determined, and the outdoor temperature, outdoor relative humidity, outdoor solar irradiance and indoor set temperature are selected as input parameters, and the optimal set point of return water temperature is selected as output parameters. Then, the data set of input and output parameters is divided into two parts, 50% is used to establish the optimal set point prediction model, and 50% is used to verify the prediction model; finally, GRNN technology is introduced to establish and verify the optimal setting of return water temperature Point GRNN prediction model.

(5) Based on the GRNN prediction model of the optimal set point of return water temperature, an optimization method for variable return water temperature of small fixed-frequency chillers is proposed. Use the best set point prediction model of return water temperature to predict the best set point of return water temperature in real time under current working conditions, adjust the return water temperature in a timely manner, and then use the position controller to optimally set the return water temperature The difference between the point and the feedback temperature controls the start and stop of the small fixed-frequency chiller.

The beneficial effects of the present invention are as follows: (1) ensure indoor thermal comfort, and avoid room overcooling or overheating caused by different regions, different buildings, and different seasons; (2) improve the energy-saving effect of small fixed-frequency chillers, and increase the temperature of chilled water as much as possible , thereby improving the operating efficiency of small fixed-frequency chillers, which will produce significant energy-saving effects.

Description of drawings

The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

Fig. 1 is the logic diagram of training data set acquisition method;

Figure 2 is a logic diagram of the calculation method for the maximum allowable temperature of the return water;

Figure 3 is a logic diagram of the verification method for the optimal set point of the return water temperature;

Fig. 4 is the logic diagram of the optimal control method for variable return water temperature;

Figure 5 is a floor plan of a certain building in Beijing;

Detailed ways

Further examples are described below for the present invention

(1) Figure 5 is a building plan of a certain floor of a certain university in Beijing. The building area is 174m ² , the air-conditioning area is 147m ² , the floor height is 2.87m, and there are 11 rooms in total. The building exterior wall materials are 3mm architectural color steel plate, 150mm polystyrene and 3mm architectural color steel plate from outside to inside, and the heat transfer coefficient is 0.773W/(m ² ·K); the inner wall materials are 3mm architectural color steel plate, 100mm polystyrene and 3mm architectural color steel plate, the heat transfer coefficient is 1.23W/(m ² ·K); the ground material is 150mm thick reinforced concrete floor, the heat transfer coefficient is 0.574W/(m ² ·K); the roof material is outward 3mm building color steel plate, 200mm polystyrene and 3mm building color steel plate are respectively inside, the heat transfer coefficient is 0.58W/(m ² ·K); the building exterior window area is 2.25m ² (1.5m×1.5m), glass The materials from outside to inside are 6mm flat glass, 6mm air interlayer and 6mm flat glass respectively, and the heat transfer coefficient is 3W/(m ² ·K). Indoor equipment and human body heat dissipation and humidity are set to fixed values, and the schedule is used to simulate the commute of personnel and the intermittent operation of equipment. The air conditioning system consists of 1 fixed frequency chiller, 2 circulating pumps, 1 water tank, and 13 fan coil units. The fixed-frequency chiller contains two identical scroll compressors, using R22 as the refrigerant, the rated cooling capacity is 25kW, and the rated heat output is 30kW; the input power of a single unit is 5.8kW for cooling and 6.1kW for heating ; There is a fan with an input power of 0.4kW. The head of the circulating pump is 42m, the flow rate is 8t/s, the speed is 2900rpm, and the power is 750W. The three cooling capacities of the fan coil unit are 1900W, 1600W, and 1340W, and the corresponding power is 31W, 26W, and 21W, and the corresponding air volume is 340m ³ / h, 260m ³ /h, 170m ³ /h, the capacity of the water tank is 0.5m ³ , through these actual parameters and then use the simulation software Dymola to establish the simulation model of the room, the simulation model of the fixed frequency chiller, the fan coil model, etc.

(2) Determine the maximum allowable temperature of the return water of the small fixed-frequency chiller. In this method, the fan of the fan coil unit is set to a fixed gear, and the water flow is set to a rated value. According to the deviation between the indoor temperature of the room and the set temperature, the water supply temperature of the fan coil unit is adjusted through PID to control the stability of the indoor temperature. The fan coil water supply temperature output by PID adjustment is the highest water supply temperature that satisfies indoor thermal comfort, and the corresponding fan coil return water temperature is the highest return water temperature that satisfies indoor thermal comfort. According to the above method, using the meteorological files of Beijing area, the simulation of the whole cooling season is carried out, and the simulation step is 20 minutes. In the simulation calculation, the adjustment range of the water supply temperature is set at 7-18.3°C, the lower limit is determined according to the rated water outlet temperature of the heat pump at 7°C, and the upper limit is determined according to the corresponding dew point temperature when the design temperature of the room is 25°C and the humidity is 50%. Make sure to ensure that the fan coil unit has dehumidification capability. Through simulation calculation, the maximum allowable temperature of chilled water return water that meets the thermal comfort of the environment can be obtained under different working conditions.

(3) Based on the maximum allowable return temperature of chilled water, verify the optimal set point of the return water temperature of the small fixed-frequency chiller. First, set the maximum allowable value of the return water temperature as the return water temperature of the small fixed-frequency chiller. Through simulation, it is checked whether the return water temperature under this working condition meets the indoor comfort requirements. If so, the return water temperature under this working condition The highest allowable value is the optimum set point for return water temperature. If the indoor temperature is not satisfied, the set value needs to be corrected. The corrected temperature is the optimal set point of the return water temperature. Afterwards, the meteorological parameters in the Beijing area were selected for simulation. The room temperature was set at 25°C, and the water temperature setting value adjustment frequency was set at 1 time/h. There are 16698 sets of optimal set point of return water temperature and operating condition data obtained through simulation calculation.

(4) According to the optimal set point of return water temperature determined above, select outdoor temperature, outdoor relative humidity, outdoor solar irradiance and indoor set temperature as input parameters, and the optimal set point of return water temperature as output parameters. The optimal set point of return water temperature GRNN consists of an input layer, a model layer, a summation layer and an output layer. The corresponding network input is X=[x ₁ , x ₂ , x ₃ , x ₄ ] ^T , and the output is The number of training samples is 8349.

① Input layer

The number of neurons in the input layer is 4, and each neuron is a simple distribution unit that can directly transmit input variables to the pattern layer.

② pattern layer

The number of neurons in the pattern layer is N=8349, and each neuron corresponds to a different training sample. When the transfer function of the mode layer neuron is selected as a Gaussian function, the output of the nth neuron in the mode layer is:

Among them, ||XX _n || is the Euclidean distance between the input vector X of the network and the weight vector X _n as the input of the neuron. When the input of the neuron is 0, the output of the neuron is the maximum value 1,. The sensitivity of a neuron to an input is modulated by a smoothing factor σ.

③Summation layer

The summation layer includes two types of neurons. The first type of neurons calculates and sums the neuron outputs of all pattern layers. The transfer function is:

The second type of neuron is to weight and sum the outputs of all pattern layers, the weight is the jth element of the nth training sample output Y _n , and the transfer function is:

④ Output layer

Each neuron divides the output of the summation layer to obtain the formula The prediction result of , the output of the jth neuron corresponds to the prediction result of the jth element, namely:

Once the training samples are determined, the network structure of GRNN and the connection weights between neurons are also determined. The smoothing factor σ is the only estimated value required by GRNN, and this parameter plays an important role in the performance of GRNN.

The optimal smoothing factor σ of the GRNN prediction model is determined by the cross-validation method, and the specific steps are as follows:

①Set the σ value, starting from 0.01, and increase by 0.01 each time within the range of [0.01,0.9];

②Take one of the training samples for inspection, and the rest are used to construct a generalized regression neural network model to predict the sample;

③ Repeat this process for each sample to get the predicted value of all samples;

④ The expected deviation percentage of the predicted value of the training sample is used as the evaluation standard of the network performance, and its calculation formula is:

in, y _i are the predicted value and observed value of the load respectively, and y _max is the maximum observed value of the load. The minimum expected deviation percentage corresponding to the σ value is the optimal smoothing factor.

Afterwards, the optimal smoothing factor σ and the cross-validation error EEP of the prediction model are obtained, and the accuracy of the model is verified by comparing the training samples with their predicted values, and comparing the predicted samples with their predicted values.

(5) Based on the GRNN prediction model of the optimal set point of return water temperature, an optimization method for variable return water temperature of fixed-frequency chillers is proposed. Use the prediction model of the optimal return water temperature set point to predict the optimal return water temperature set point under the current working condition in real time, and adjust the return water temperature of the small fixed-frequency chiller in a timely manner. Using the position controller, according to the return water temperature The difference between the optimal set point and the feedback temperature controls the start and stop of a small fixed frequency chiller.

Claims (4)

1. A variable water temperature control method for small fixed-frequency chillers based on the optimal set point of chilled water return temperature, characterized in that: First, establish a model of the air-conditioning system of a small fixed-frequency chiller, then use the model to simulate and collect the design parameters of the building and the typical annual meteorological data in the area, and calculate the chilled water return water based on thermal comfort under different operating conditions The maximum allowable temperature; check and correct the maximum allowable return temperature of chilled water to determine the optimal set point of return water temperature, establish data sets of different operating conditions, and establish and verify the optimal return temperature setting of fixed-frequency chillers based on the data sets Fixed-point GRNN prediction model; use the best return water temperature prediction model to predict the best return water temperature set point in the current working condition in real time, adjust the return water temperature of small fixed-frequency chillers in a timely manner, and then use the bit formula The controller controls the start and stop of the fixed-frequency chiller according to the difference between the return water temperature and the optimal set water temperature. 2. The variable water temperature control method for small fixed-frequency chillers based on the optimal set point of chilled water return temperature according to claim 1, characterized in that: (1) Establish a simulation model; first establish a room model, a fixed-frequency chiller model, a fan coil model, and a control model sub-model in the simulation software Modelica according to the architectural design parameters, and simulate the air-conditioning system through the joint operation of the sub-models dynamic running process; (2) Calculate the maximum allowable temperature of the frozen return water of the chiller; through the simulation method, the air-conditioning terminal equipment is set to work at a fixed air volume and rated flow rate, and the indoor comfort of the air-conditioned room is ensured by adjusting the temperature of the chilled water. Based on typical annual meteorological data, the maximum allowable temperature of chilled water return water based on thermal comfort is calculated under different operating conditions; (3) Determine the optimal set point of the return water temperature and establish a data set; first, set the maximum allowable value of the return water temperature as the return water temperature of the chiller, and verify whether the return water temperature under this working condition meets the indoor temperature requirements through simulation. If the comfort requirement is satisfied, the maximum allowable value of the return water temperature under this working condition is the best set point of the return water temperature; if the indoor temperature is not satisfied, the set value needs to be corrected; The optimal set point of the return water temperature of the small fixed-frequency chiller; according to the optimal set point of the return water temperature determined above, as well as the operating condition data, a data set is established; (4) Establish and verify the GRNN prediction model of the optimal return water temperature set point; firstly, determine the input and output parameters of the prediction model, select outdoor temperature, outdoor relative humidity, outdoor solar irradiance and indoor set temperature as input parameters, and return to The optimal set point of water temperature is the output parameter; then, the data set of input and output parameters is divided into two parts, one part is used to establish the optimal set point prediction model, and the other part is used to verify the prediction model; finally, the GRNN technology is introduced, Establish and verify the GRNN prediction model for the optimal set point of return water temperature; (5) Based on the GRNN prediction model of the optimal set point of return water temperature, an optimization method for variable return water temperature of small fixed-frequency chillers is proposed; the return water temperature of the current working condition is predicted in real time by using the prediction model of the optimal set point of return water temperature The optimum set point is to adjust the return water temperature in a timely manner, and then use the position controller to control the start and stop of the small fixed-frequency chiller according to the difference between the optimum return water temperature set point and the feedback temperature. 3. The variable water temperature control method for small fixed-frequency chillers based on the optimal set point of chilled water return temperature according to claim 1, characterized in that: (2) Determine the maximum allowable temperature of the return water of the small fixed-frequency chiller; this method sets the fan of the fan coil unit to a fixed gear, and the water flow rate to the rated value. According to the deviation between the room temperature and the set temperature, through PID adjusts the water supply temperature of the fan coil unit to control the stability of the indoor temperature; the water supply temperature of the fan coil unit output by PID adjustment is the highest water supply temperature that meets the indoor thermal comfort, and the corresponding fan coil return water temperature is the highest return water temperature that meets the indoor thermal comfort. water temperature. 4. The variable water temperature control method for small fixed-frequency chillers based on the optimal set point of chilled water return temperature according to claim 1, characterized in that: (4) According to the above-mentioned optimal set point of return water temperature, select outdoor temperature, outdoor relative humidity, outdoor solar irradiance and indoor set temperature as input parameters, and the optimal set point of return water temperature as output parameters; The optimal set point of return water temperature GRNN is composed of input layer, model layer, summation layer and output layer; the corresponding network input is X=[x ₁ , x ₂ , x ₃ , x ₄ ] ^T , and the output is The number of training samples is N; ① Input layer The number of neurons in the input layer is 4, and each neuron is a simple distribution unit that directly transmits input variables to the pattern layer; ② pattern layer The number of neurons in the pattern layer is N, and each neuron corresponds to a different training sample; when the transfer function of the pattern layer neuron is selected as a Gaussian function, the output of the nth neuron in the pattern layer is: Among them, ||XX _n || is the Euclidean distance between the input vector X of the neuron and the weight vector X _n ; when the input of the neuron is 0, the output of the neuron is the maximum value 1; The sensitivity of the element to the input is adjusted by the smoothing factor σ; ③Summation layer The summation layer includes two types of neurons. The first type of neurons calculates and sums the neuron outputs of all pattern layers. The transfer function is: The second type of neuron is to weight and sum the outputs of all pattern layers, the weight is the jth element of the nth training sample output Y _n , and the transfer function is: ④ Output layer Each neuron divides the output of the summation layer to obtain the formula The prediction result of , the output of the jth neuron corresponds to the prediction result of the jth element, namely: Once the training samples are determined, the network structure of GRNN and the connection weights between neurons are also determined. The smoothing factor σ is the only estimated value required by GRNN, and this parameter plays an important role in the performance of GRNN; The optimal smoothing factor σ of the GRNN prediction model is determined by the cross-validation method, and the specific steps are as follows: ①Set the σ value, starting from 0.01, and increase by 0.01 each time within the range of [0.01,0.9]; ②Take one of the training samples for inspection, and the rest are used to construct a generalized regression neural network model to predict the sample; ③ Repeat step ② for each sample to obtain the predicted values of all samples; ④ The expected deviation percentage of the predicted value of the training sample is used as the evaluation standard of the network performance, and its calculation formula is: in, y _i are the predicted and observed values of the load, respectively, and y _max is the maximum observed value of the load; the minimum expected deviation percentage corresponding to the σ value is the optimal smoothing factor.