Operational optimization on large‐scale combined heat and power units with low‐pressure cylinder near‐zero output

Many power plants have been transformed and operated using low‐pressure cylinder near‐zero output (LP‐ZO) technology in China to make way for or provide necessary guarantees for renewable energy power generation. This research focuses on utilizing numerical simulation tools to improve the operational economy of units with LP‐ZO. The LP‐ZO thermodynamic and energy‐consumption model of a reference plant is established, and the momentum particle swarm optimization method is used to optimize the unit operations. Moreover, a model to determine the economic operational range of the combined heat and power plant with LP‐ZO is proposed. The results show that the LP‐ZO mode operations of the unit can reduce the minimum peak shaving rate and improve the flexibility of operations. The energy efficiency of the LP‐ZO mode is increased due to the recovery of a significant portion of the cold source losses. However, the irreversible exergy loss caused by large temperature difference heat exchange reduces the exergy efficiency of the unit. After operational optimization, the reference plant can save approximately 68.15 t d−1 of coal and reduce carbon dioxide emissions by about 27 t tce−1 d−1. This study provides an operating approach to the CHP plant with LP‐ZO unit in enhancing flexibility and economy.


| INTRODUCTION
With peak carbon dioxide emissions before 2030 and achieving carbon neutrality before 2060, China's energy structures have entered an important period of in-depth adjustments. Energy conservation is key to constructing new development patterns for comprehensive intelligent energy. The combined heat and power (CHP) plants recover a portion of the cold source loss; therefore, they are more economical than pure condensing units. 1,2 However, with the acceleration of urbanization in China, the demand for heat supply services is growing rapidly. 3 Thus, the contradiction between heat and power is becoming more and more prominent, especially in the deep peak shaving period during the winter heating season. Bypass systems for heating, 4,5 high backpressure for heating, 6,7 and other technologies 8,9 are used to improve the peak shaving capacity and operational flexibility of CHP units.
Due to limitations in equipment operating conditions, the use of high-pressure cylinder/low-pressure cylinder (HP-LP) bypass heating and heat pumps to improve the peak shaving capacity of CHP units is limited. 10,11 Although using high backpressure for heating can increase the peak shaving capacity of CHP units, this forms a mode of determining the electrical generation from the heat, which reduces the flexibility of CHP plants. 12,13 The lowpressure cylinder near-zero output (LP-ZO) technology overcomes the disadvantage of inflexibility in high backpressure operations. The CHP units transformed by LP-ZO can switch flexibly between the high backpressure and extraction condensing operational modes, which increases the flexibility of CHP operations compared with high backpressure for heating. 14 Research on the economic performance and safety of CHP units with LP-ZO is equally important. Ge et al. 15 developed a 330 MW unit model with LP-ZO using the Ebsilon software to analyze the thermodynamic characteristics and heating capacity of the unit after transformation. After the transformation, the maximum heating capacity increased by approximately 37.1%, and the peak shaving capacity increased by around 34.8%. Yang et al. 16 used the Ebsilon software to construct an LP-ZO model for a 300 MW unit, and they studied the peak shaving capacity and operational economic benefits of the reconstructed unit. The results suggest that the depth peak-shaving capacity of the unit increased by 52.76 MW, and the operational benefit increased by 7800 CNY h −1 . Liu et al. 17 analyzed the thermal performance of a 320 MW unit with LP-ZO. The maximum heating steam extraction capacity of the unit increased, and the coal consumption for the electrical supply was reduced after the LP-ZO transformation. Wang 18,19 elaborated on the zero output modification scheme for a low-pressure cylinder of a 650 MW unit. The heat supply capacity of the modified unit increased by 52%, and the coal consumption for power generation under rated conditions was reduced by about 36.5 g kW −1 h −1 . In addition, the results showed that reducing the operating backpressure of the unit allows for increasing the volume flow for the low-pressure cylinder, thereby enhancing the unit's safety. Tian et al. 20 studied the safety of an LP-ZO unit and analyzed the flow field characteristics of the low-pressure cylinder and the dynamic stress of the final stage blades. Ju et al. 21 comparatively analyzed the advantages and disadvantages of HP-LP bypass for heating, heat storage tanks, electrode boilers, and LP-ZO technology in heating units. The results showed that using LP-ZO technology can reduce the costs of transformation technologies, which gives the best thermal economy. However, when the unit operates in the LP-ZO mode, it forms thermoelectric coupling. To date, power plants usually operate two or more units in parallel. Therefore, improving the flexibility of unit operations uses one unit operating in the high backpressure mode and the other in the extraction condensing mode. 22 The above research focuses on the transformation of a single unit and is limited to analyzing a few working conditions. The optimal operations of a single unit do not represent the optimal operating conditions of the power plant. In addition, relevant research has not deeply analyzed the economic operation range of plants with LP-ZO.
To boost the economy and flexibility of CHP plants with LP-ZO, an optimized operation method based on the thermodynamic characteristics of LP-ZO units is established in this study. Then, the economic operational range model of the LP-ZO unit is further proposed based on the above optimization algorithm. A thermodynamic model of the extraction condensing and LP-ZO modes are built using the Ebsilon software based on a 600 MW coal-fired CHP unit. Moreover, the data for the fully operating regions of the unit are obtained by calling the simulation data through Matlab. The associated thermodynamic characteristics of the zero-output unit were analyzed. Based on the load data of a characteristic day for the reference plant, the operations of two units are optimized using the momentum particle swarm optimization (MPSO) method. Finally, the economic operational range of the LP-ZO mode is further discussed for the power plant. This work guides the LP-ZO reconstruction and its operation for CHP plants.

| LP-ZO TECHNOLOGY
The LP-ZO technology was recently proposed by the Xi'an Thermal Engineering Research Institute of China. Some domestic power plants have been transformed and operated using this mode. The LP-ZO transformation technology and the thermoelectric conversion characteristics of the transformed unit are introduced in this section.

| Key technologies and transformation scheme
The LP-ZO technology is also known as "cutting off the low-pressure cylinder for heating technology". The core approach is to retain only a small amount of cooling steam in the low-pressure cylinder to achieve near-zero output to allow more steam to enter the heating system, improve the heating capacity, and reduce the electrical load of the CHP unit. This technical operational mode can realize flexible switching between the extraction condensation mode and the high backpressure mode without shutting down the unit, which increases the flexibility of peak shaving. A principal thermodynamic system diagram for LP-ZO is shown in Figure 1. This technology mainly transforms the heating butterfly valve and heating pipeline, replaces the heating butterfly valve with a completely-sealed hydraulic butterfly valve, and increases the diameter of the heating pipeline to satisfy the heating extraction volume. In addition, when the low-pressure cylinder is operating in small volume flow conditions, the final stage and next-final stage of the low-pressure cylinder are under blast conditions. This causes the temperature of the final stage blades to increase, which requires water spray cooling to maintain the temperature of the low-pressure cylinder within a safe range. The last blade of the low-pressure cylinder should be sprayed with a metal wear-resistant layer to prevent water erosion.

| Operationally feasible area of the LP-ZO unit
Due to the influence of the operating characteristics of CHP units, when the external thermal load is constant, the electrical generation range of the unit is limited, which forms an electric-heat coupled situation that reduces the operational flexibility. The gray area enclosed by lines AB, BC, CD, and DA in Figure 2A is the operationally feasible area of the traditional CHP unit. When the external thermal load is Q 1 , the electrical load of the extraction condensing unit can only change between P 1 and P 2 . After the LP-ZO transformation of the steam turbine, the operationally feasible area of the LP-ZO unit is shown in Figure 2B. The LP-ZO unit can flexibly switch between the extraction condensing mode and the LP-ZO mode. Thus, the electrical load range changes from P min -P max to P min_z -P max , and the thermal load changes from 0-Q max to 0-Q max_z , which increases the peak shaving capacity of the unit. However, when the unit operates in the LP-ZO mode, the electrical load is P 3 when the external thermal load is Q 1 , and the unit operates completely by determining the electric by heat. Therefore, to improve the peak shaving capacity and operational flexibility of units in a CHP plant, one unit adopts the traditional extraction condensing mode, and the other adopts the LP-ZO mode.

| REFERENCE CASE
The thermodynamic and energy-consumption model of a reference CHP plant is established in this section. The basic parameters of the reference unit and methods to establish a thermodynamic model are briefly introduced.

| Disposition of unit
The reference CHP plant contains two 600 MW CHP units; the two steam turbines are the primary intermediate reheat and the heating steam exhaust from the intermediate pressure cylinder. One of the two units in the case power plant is a traditional CHP unit, and the other is an LP-ZO unit. Two boilers as the supercritical once-through with a F I G U R E 1 The principal thermodynamic system diagram of low-pressure cylinder near-zero output.
F I G U R E 2 Operationally feasible area of combined heat and power unit. single furnace and primary intermediate reheat. The maximum load of the boiler is 1900 t h −1 and the minimum stable combustion load is 503.37 t h −1 . Table 1 shows the  basic performance parameters for the unit, and Table 2 shows the characteristic day loads for a certain month during the heating period.

| Thermodynamic model development
The Ebsilon software was developed by the STEAG electric power company in Germany, as shown in Figure 3. It is a general thermodynamic modeling configuration software used primarily to calculate and simulate the thermal balance of thermal systems. The software can establish the thermal system model by combining related modules. The thermodynamic model for the reference unit was established using the Ebsilon software. The end difference between the heat exchanger and the backpressure of the steam turbine remains unchanged under off-design conditions. The heating extraction pressure is constant, and the steam temperature after heat exchange drops to the saturated water temperature under this pressure. In the LP-ZO mode, the minimum cooling steam flow for the LP is assumed to be 40 t h −1 . The correctness of the thermodynamic model is verified by the thermal balance diagram of the unit in the plant; the relative errors between the numerical results of electric load and the values of the heat balance diagram are less than 1%, as shown in Table 3.

| Energy-consumption model development
The energy consumption of the unit is affected by operational mode and load. Therefore, the unit coal consumption is functionally related to the electrical generation and thermal supply. 23 This can be obtained under off-design models, which can be expressed as: where C represents the coal consumption of the unit (t); Q t represents the heat input to the turbine (kW); LHV represents the lower heating value of the fuel (17.31 MJ kg −1 ); η b represents the boiler efficiency (%); η p represents the pipe efficiency of the boiler (%). The standard coal consumption of the unit can be expressed as: where C s represents the standard coal consumption of the unit (t); β represents the standard coal conversion factor (29.27 MJ kg −1 ). The carbon dioxide emissions of the unit is calculated as follows: where OP is the carbon dioxide emissions of the unit (t); ε is the carbon dioxide emission factor of standard coal. The recommended value from the Institute of Energy Economics, Japan (IEEJ) is 0.68 t tce −1 , from the U.S. Energy Information Administration (EIA) is 0.69 t tce −1 , and from the Energy Research Institute National Development and Reform Commission of China is 0.67 t tce −1 which is adopted in this paper. The exergy and energy efficiency are introduced to discuss the energy conversion of the unit from quantity and quality of energy perspectives. The exergy efficiency is the ratio of the output exergy of the unit to the input exergy, which can be expressed as Formula (4). The electrical exergy and thermal exergy constitute the output exergy of the unit, and the chemical exergy of the fuel is the input exergy.
where η ex represents the exergy efficiency of the unit (%); E p and E Q represent the electrical exergy and thermal exergy, respectively; E f represents the chemical exergy of the fuel.
Energy efficiency is the ratio of the output energy to the input energy, which is calculated as Formula (5).  Electrical energy and thermal energy constitute the output energy, and the chemical energy of the fuel is the input energy.
where η en represents the energy efficiency (%); P p and Q s represent the electrical energy and thermal energy of the unit, respectively (kW).

| OPERATIONAL OPTIMIZATION MODEL DEVELOPMENT
The objective of operation optimization for CHP plants with LP-ZO units is to minimize coal consumption for a characteristic day under the given constraints. This section provides the fitness function and constraints of the operation optimization model. Moreover, the MPSO algorithm is used to find the optimal value of the fitness function for unit operational optimization. The optimization model and steps are briefly introduced.

| Fitness function
Taking coal consumption of the reference power plant as the fitness function, the objective is to minimize the total coal consumption, which can be expressed as: where C total represents the total coal consumption of the reference plant (t); C ij represents the coal consumption of unit i in hour j (t).

| Thermal and electrical load balance for the cogeneration plant
The thermal load for the two units in the power plant is balanced with the demand of the heating network as: where Q j represents the demand load from the heat supply network in hour j, kW; Q ij represents the thermal load of unit i in hour j, kW.
The electrical load for the two units in the power plant is balanced with the demand of the electric grid as: where P j represents the demand load from the electric grid in hour j (kW); P ij represents the electrical load of unit i in hour j (kW).

| Thermal load constraints for the cogeneration plant
The thermal load for the cogeneration unit should not be greater than the demand load from the heat supply network and less than the maximum thermal load of the unit. The unit operates in extraction condensing mode as: where Q max represents the maximum thermal load of the unit when operating in extraction condensing mode (kW). The unit operates in LP-ZO mode as: where Q max_z and Q min_z are the maximum and minimum thermal load, respectively, of the unit when operating in LP-ZO mode (kW).

| Electrical load constraints of the cogeneration plant
Various operational modes are adopted when the thermal load is constant and the limit range of the unit electrical load differs, as shown in Figure 2B. When the unit operates in the extraction condensing mode, the electrical load must change within the operationally feasible area. When the unit operates in the LP-ZO mode, the electrical and thermal loads must meet the constraints relationship. The unit operates in the extraction condensing mode as: where f Q ( ) ij DCB and f Q ( ) ij AB are the boundary constraints when the thermal load is Q ij .
The unit operates in the LP-ZO mode as: i jEF (14) where f Q ( ) ij EF represents the constraint relationship when the thermal load is Q ij .

| Optimization model
In recent years, particle swarm optimization (PSO) has been gradually applied to CHP plant operations. Zhao et al. 24 optimized the operation of a cogeneration system with compressed air energy storage through the PSO algorithm. Lai et al. 25 optimized the operation of CHP systems with heat storage tanks using the PSO algorithm. Yuan and Yang 26 established the dispatching optimization model for five units, which was optimized using the PSO. However, the PSO algorithm readily converges to local optima, which reduces the quality of the optimization results. Therefore, the PSO algorithm has been improved by many researchers. [27][28][29][30] In this paper, the operation optimization of the reference CHP plant adopts the MPSO algorithm. 31 Compared with the traditional PSO algorithm, the MPSO algorithm solves the problem where the PSO algorithm easily falls into local optima, which improves the computational efficiency and results' quality. In the MPSO algorithm, the particle position is randomly initialized, and the individual and global optimal values are updated continuously with an iterative process. The optimized results can be obtained after a limited number of iterations. The velocity and position update functions for the particles in the optimization algorithm can be expressed as follows: where v represents the particle velocities; x represents the particle positions; c 1 and c 2 are constants set to1.49445; rand() is a randomly generated function from 0 to 1; P best represents the individual optimal value of the particles; g best represents the global optimal value of the particles; t is the number of iterations; λ is the coefficient set to 0.3. The velocity and position update functions of PSO algorithm method for the particles in the optimization algorithm can be expressed as follows: where w represents the inertia weight.

| Optimization process
The LP-ZO unit is different from the high backpressure and extraction condensing units. The unit, after the LP-ZO modification, can flexibly switch between that mode and the extraction condensing mode. Thus, the two operational modes should be considered when optimizing the unit operations. The specific calculation process is shown in Figure 4. First, it is judged whether the external thermal load meets the thermal constraints. If so, the MPSO method for the extraction condensing mode and the MPSO method for the LP-ZO mode are used to optimize the operations between units. Finally, the optimization results for the two modes are compared, the minimum fitness function (coal consumption) is selected as the optimal result, and the operation mode of the unit is the output. The running mode outputs 1 when the unit is operating in the extraction condensing mode, and the running mode outputs 2 when the unit is operating in the LP-ZO mode. If the external thermal load does not meet the thermal constraints, the program ends, and "the thermal load is not within the operationally feasible area" is the output. Figure 5 shows the optimization process of the MPSO method for the LP-ZO mode. First, the initial particle population of the thermal load Q 1 for the LP-ZO mode is generated, and the electrical load P 1 of the associated mode is calculated based on Q 1 . Second, the electrical load P 2 and thermal load Q 2 of the extraction condensing unit is calculated based on the constraints of the thermal and electrical load balance. Then, the fitness function is calculated based on the load data for two units, and the individual and global optimal values for the particles are initialized. Finally, the positions of particles are continuously updated by iterating the velocity and position functions to update the individual and global optimal values for the particle swarm. The optimization results are output when the maximum number of iterations is reached. It is noted that the constraints of the units are realized through the penalty function. When the unit loads do not meet the constraints, the fitness function becomes infinite. Compared with the MPSO method of the LP-ZO mode, the MPSO method for the extraction F I G U R E 4 Operational optimization process for combined heat and power (CHP) units.
F I G U R E 5 The momentum particle swarm optimization method for the low-pressure cylinder near-zero output mode. condensing mode is similar except for the different constraints, so the optimization process is not described in detail here.

| Thermoelectric characteristics of the LP-ZO unit
The entire operating condition dataset for the unit is obtained by changing the boiler and steam turbine thermal loads in the Ebsilon software model. Then, Matlab transfers the Ebsilon software simulation data to construct the operationally feasible area and energyconsumption model for the reference cogeneration unit. The operationally feasible area of the LP-ZO unit is shown in Figure 6. The area enclosed by the lines AB, BC, CD, and DA indicate the operationally feasible area for the unit to operate in the extraction condensing mode, and line C′D′ is the operationally feasible area when the unit operates in the LP-ZO mode. When the thermal load of the unit is 400 MW and operates in the extraction condensing mode, the electrical load ranges from P 1 to P 2 at approximately 542.9-193.6 MW. When the unit operates in LP-ZO mode, the electrical load of the unit is P 3 at approximately 94.1 MW. Compared with the extraction condensing mode, the minimum peak shaving of LP-ZO mode is reduced from 32.2% to 15.7%.
The energy efficiency and exergy efficiency of the reference cogeneration unit when operating in the LP-ZO mode are shown in Figure 7. The thermal load range of the unit is from approximately 326-927 MW. The exergy efficiency of the unit increases from around 32.93% to 38.98% as the thermal load increases, and the maximum energy efficiency is about 89.3%, with a minimum of approximately 87.5%. Compared with previous research results, 23 the maximum energy efficiency of the unit is about 78.8% when operating in the extraction condensing mode. After the LP-ZO transformation, the range of energy efficiency is increased as the unit recovers a large amount of condensation loss. When the unit operates in the LP-ZO mode, high-quality steam is transformed into low-quality thermal energy. Therefore, large temperature difference heat exchange causes irreversible exergy loss, which results in a reduced exergy efficiency range for the reference unit.

| Thermodynamic characteristics of the extraction condensing mode
When the two units of the reference CHP plant operate in the extraction condensing mode, the loads they bear differ, and the thermodynamic characteristics of the power plant vary correspondingly. Based on the thermodynamic characteristics of the units, the effect of the thermal-electrical load distribution on the coal consumption and average exergy efficiency of the power plant is studied using a quantitative analysis method. Assuming that the external thermal and electrical loads demand for the reference power plant is 720 and 740 MW, respectively, the distribution of the coal consumption and average exergy efficiency of the power plant can be obtained by adjusting the thermal and electrical loads of a single unit, as shown in Figures 8 and 9.
F I G U R E 6 Operationally feasible area of the low-pressure cylinder near-zero output unit.
F I G U R E 7 Thermodynamic characteristics of the low-pressure cylinder near-zero output mode.
The influence of load changes for a single unit on the coal consumption of the cogeneration plant is shown in Figure 8. When the external thermal and electrical loads are determined, the coal consumption of the power plant presents a nonlinear centrosymmetric distribution structure due to the parallel operation of two units. When the electrical and thermal loads are evenly distributed between the two units, the coal consumption is at an average level of approximately 460 t h −1 . The maximum value of the coal consumption of the power plant is distributed at the maximum or minimum electrical load in the operationally feasible area of the units at approximately 462 t h −1 , and the minimum is distributed at the maximum or minimum thermal loads in the operationally feasible area of the units at about 458 t h −1 . When the electrical and thermal loads are 740 and 720 MW, respectively, the coal consumption of the power plant varies by about 4 t h −1 within the operationally feasible area.
The average exergy efficiency distribution of the plant exhibits a saddle shape, as shown in Figure 9. Under the same external load demand, the maximum average exergy efficiency for the units in the operationally feasible area is approximately 39.65%, with a minimum of approximately 38.29%, which gives an extreme difference from the average exergy efficiency of about 1.4%. Comparing the two figures indicates that the maximum coal consumption is the minimum average exergy efficiency point, and the maximum average exergy efficiency corresponds to the minimum coal consumption point. In conclusion, when the external thermal and electrical loads are certain, optimizing the operation between units can reduce the coal consumption of power plants to alleviate carbon dioxide emissions and improve the economy of plant operations.

| Comparison of optimization models
When the number of particle swarms is set to 100, the number of iteration steps is 150, and the external thermal and electrical loads are 407.2 and 522.4 MW, respectively. The MPSO and PSO method are used to optimize the unit operation three times, and the results are shown in Figure 10. It can be seen from the figure that the two optimization methods have the same optimization results under the same conditions, but the PSO method has not obtained the optimal value at about 40 steps, while the MPSO method has obtained the optimal value. Therefore, the MPSO method can reduce the number of iteration steps and improve the calculation efficiency when there are many load data. Thus, the MPSO method is used to optimize the unit operation in this paper.

| Operational optimization of reference plant
The unit transformed by LP-ZO can be flexibly switched between the high backpressure and extraction condensing modes. Therefore, when optimizing the unit's operation, the two modes of LP-ZO and extraction condensing should be considered simultaneously. Here, the operational modes of the units are coupled in the MPSO algorithm. The optimization algorithm in Section 4.5 is used to optimize the operations of the two units using the characteristic day loads data.
The thermoelectric load borne by the two units after the optimized operations of the cogeneration plant is shown in Figure 11. After the optimized operations, one  unit in the power plant operates at Point C in Figure 6; that is, the maximum thermal working condition of the steam turbine unit when the boiler is under a stable combustion load, and the other unit performs peak shaving operations with the external thermal and electrical loads demand. When the unit operates at Point C, the irreversible exergy loss corresponding to the same steam extraction volume is minimal due to the low steam parameters. At the same time, Figure 7 indicates that when two units operate in parallel, the average exergy efficiency of the power plant presents a parabolic distribution, and changes in the exergy efficiency gradually slow down as the load moves to the average distribution point.
The operational modes of the power plant units after optimization are shown in Figure 12, in which 1 represents that the two units operate in the extraction condensing mode, and 2 represents that one unit operates in the LP-ZO mode and the other operates in the extraction condensing mode. Both units operating in the extraction condensing mode are more economical in the above characteristic day of the power plant. Before optimization, the average distribution mode of loads is adopted in the reference cogeneration plant. The coal consumption of the power plant after optimization is compared with that before optimization. The results shown in Figure 13 indicate that the hourly coal consumption of the optimized power plant is less than that of the average load distribution. When the average distribution mode of loads is adopted between two units, the coal consumption on the characteristic day of the cogeneration plant is approximately 7827.69 tons.  The above analysis indicates that it is more economical for both units of the reference cogeneration plant to operate in the extraction condensing mode under the characteristic day loading. Therefore, it is important to study the economic operational range of the LP-ZO mode, which can provide theoretical support for the transformation and guide its economic operations. In this section, an economic operation range model for the LP-ZO mode is established by improving the operational optimization model, as shown in Figure 14. First, the external load is input into the model, and the unit operations are optimized using the MPSO method for the extraction condensing mode and LP-ZO modes. Second, the fitness function results are compared. When the fitness function of the LP-ZO mode is greater than that of the extraction condensing mode, the model outputs the operational optimization results of the extraction condensing mode; otherwise, the operational F I G U R E 13 Coal consumption of power plant. optimization results of the LP-ZO are output. Finally, it is determined whether the electrical and thermal loads of the two units are in the operationally feasible area. If the electrical and thermal loads are in the operationally feasible area, the operational mode will be output (1 represents the extraction condensing mode and 2 represents the LP-ZO mode). If the electrical and thermal loads are not in the operationally feasible area of the plant, the operation mode will output 0; that is, both the extraction condensing mode of two units and the LP-ZO mode of a single unit cannot meet the external load demands.
The economic operational analysis interval of the LP-ZO mode takes the electrical load area of 0-500 MW and the thermal load of 500-1000 MW as an example. The economic operational range model of the LP-ZO mode is used to simulate this load area, and the results are shown in Figure 15. There are many discrete points between areas B and C, as this region is located at the boundary of the operationally feasible area of the extraction condensing mode. When using the MPSO algorithm, the step size of the particle velocity and the randomly initialized particle locations do not allow the particles to accurately capture the boundary points of the operationally feasible area; thus, discrete points will appear near the boundary of the operationally feasible area. Therefore, a straight line is fit near the discrete points as the dividing line, shown as the red line in the figure. According to the model results, the load region is divided into three areas: A, B, and C. Area A is when there is an external demand load that cannot be met, even if a single unit adopts LP-ZO mode. Area C is when the external demand load within this area makes it more economical for both units to adopt extraction condensing mode. When the external demand load is in area B, it is more economical for one unit to operate in the LP-ZO mode and the other with the extraction condensing mode. Therefore, when the external load is in area B, it is more economical to perform an LP-ZO transformation and operate in that mode for the reference plant.

| CONCLUSION
This paper established a thermodynamic model for a 600 MW unit with extraction condensing and LP-ZO modes using the Ebsilon software and analyzed the thermodynamic characteristics of the unit in the LP-ZO mode. In addition, a quantitative analysis was performed for the thermodynamic characteristics of the two units operating in parallel with the extraction condensing mode. The plant operations were then optimized using the load data for a characteristic day. Finally, an economic operational range model for the LP-ZO mode was proposed to determine the associated operating load range, which provides an operating approach to the CHP plant with an LP-ZO unit in enhancing flexibility and economy. The following research findings can be obtained: 1. The thermoelectric characteristics of a 600 MW CHP unit with an LP-ZO was investigated. The results show that the minimum peak shaving capacity of the unit increases after the LP-ZO transformation. In addition, the recovery of the cold end loss increases the maximum energy efficiency of the unit. However, the exergy efficiency of the unit decreases due to the transformation of high-quality steam into low-quality thermal energy. is divided into three areas: A, B, and C. When the external demand load is in area B, the reference plant that operates in the LP-ZO mode is more economic. This model provides a theoretical reference for the LP-ZO transformation and its operation in power plants.