2。 Chiller model 

制冷机模型

2。1。 Description of the chiller 

2。1制冷机的描述

In this study, a chiller plant installed in an institutional complex was investigated, which comprised of five identical screw chillers connected in parallel。 Each chiller had two refrigeration circuits, namely circuit 1 and circuit 2, and each refrigeration circuit was equipped with two compressors, as shown in Fig。1。 The nominal cooling capacity of the studied chiller was 1116 kW, and the rated power of the studied chiller was 398 kW。 The air-cooled condensers contained 16 identical condenser fans to deliver a total airflow rate of 85。5 m3/s by eight groups, and each refrigeration circuit had four fan groups。 The fan speed was 15。8 r/s, and each fan consumed a power of 2。4 kW。 

在这项研究中，安装在一个机构复杂的冷水机组进行了调查，其中包括五个相同的螺杆式冷水机组并联连接。每个制冷机有两个制冷电路，即电路1和电路2，每个制冷电路配备了两个压缩机，如图1所示。额定制冷量的研究冷水机组为1116千瓦，额定功率的研究冷水机组为398千瓦。空冷冷凝器包含16个相同的冷凝器风扇提供的总气流速率为85。5立方米/秒的由八组，每个制冷电路有四风扇组。风扇转速为15。8转/秒，每个风扇消耗功率为2。4千瓦。

To investigate the performance of the air-cooled chiller, the operating data of the chiller plant were monitored yearround by a building management system, which were used to develop and verify the chiller model。 Due to the accuracy of measured variables, the uncertainty associated with COP was determined by the single sample analysis [8], using the following equation: 

探讨了风冷冷水机组的性能，对冷水机组的运行数据进行监测，全年的楼宇管理系统，它被用于开发和验证机模型。由于测量变量的准确性，与COP相关联的不确定性是由单样本分析[ 8 ]，使用以下方程：

n 2

GCOP(rms) ¦>Gxi ˜ w( COP x/w i )@ (1) 

i 1

In Eq。 (1), xi is the ith independent variable, δxi is the uncertainty of the variable xi, and n is the number of measurements related to COP。 The root sum square error of chiller COP (δCOP(rms)) due to all the uncertainties of the inpidual variables was evaluated to be 0。09 in a COP value of 2。8 at the design condition, and the uncertainty of COP was 3。2%。 

在式（1），西是与独立变量，δ西是变西的不确定性，n是与警察测量数。的平方和的平方根误差（δCOP COP（RMS））由于评估为0。09一2。8的COP值在设计条件的个体变量的不确定性，不确定性和COP为3。2%

2。2。 Development of chiller model 

2。2冷水机组模型的开发

The chiller model was developed using the simulation program TRNSYS based on mass and energy conservation laws。 Classical heat exchanger efficiency method was used to model the evaporator and condenser。 The whole model consisted of a set of equations。 The general equations for energy balance and the assumptions made in the chiller model have been reported [5, 9]。 The procedure to determine the operating variables of the chiller model referred to the flow chart reported [9]。 The programme started with the model initialization using the input data。 As the cooling load of the chiller could be shared within the refrigeration circuits randomly, the strategy for the CLS should be specified first。 Then, the evaporating temperature and pressure of circuits 1 and 2 (Tev1, Tev2, Pev1 and Pev2) and the cooling loads of the three sections of the heat exchangers (Q11, Q12 and Q2) were calculated through an iterative procedure by assuming an initial value of Q11。 Once the model had determined the evaporating temperature and pressure of circuits 1 and 2, the model evaluated the state variables of each refrigeration circuit。 As the condensing temperature interacted between the compressor and condenser components, an iterative procedure was implemented to solve the operating variables of the two components simultaneously。 To control the condensing temperature, there was another iterative loop for determining the number of staged condensing fans。 The number of staged condenser fans and the corresponding airflow were computed according to the set point of condensing temperature。