Multi-unit multi-mode combined heat and power generation power station load distribution method based on improved Harris eagle optimization algorithm

Abstract

The invention discloses a multi-unit multi-mode cogeneration power station load distribution method based on an improved Harlisia eagle optimization algorithm, and belongs to the technical field of energy management and optimization. The method comprises the specific steps that an HPHHO model is established, and characteristic equations of units in different heat supply modes (single extraction, absorption heat pump and cylinder switching) are analyzed; fitting an energy consumption characteristic polynomial of each unit through variable working condition calculation; taking the minimum total coal consumption as an objective function, and constructing an electric heating load balance constraint and a unit operation domain non-balance constraint; and optimizing multi-unit load distribution by using HPHHO. The method is compatible with the thermoelectric coupling relation of a multi-mode unit, the optimization problem under complex constraints is effectively solved, and it is ensured that a load distribution result meets external requirements and is in a feasible operation domain. Experiments show that compared with traditional HHO, the improved algorithm has higher convergence speed and stability, the total coal consumption of the system is remarkably reduced, and the economical efficiency and the operation efficiency of the combined heat and power generation power station are improved.

Description

Technical Field

[0001] This invention relates to the field of energy management and optimization technology, specifically to a load allocation method for multi-unit, multi-mode combined heat and power (CHP) power plants based on the improved Harris Hawk Optimizer (HPHHO) algorithm. By integrating a differential evolution algorithm, a reverse learning strategy, and dynamic parameter adjustment technology, this method addresses the complex heat-electric coupling relationship and load optimization issues in multi-unit, multi-mode CHP systems, minimizing total coal consumption. Background Art

[0002] Combined heat and power (CHP) systems improve energy efficiency by simultaneously generating electricity and providing heat. However, load distribution is challenging due to the diverse types of units and complex operating modes. Conventional optimization algorithms (such as genetic algorithms and particle swarm optimization) have the following problems:

[0003] Multi-mode coupling is difficult to handle: There are significant differences in the thermal and electric load equations under different modes such as single extraction heating, coupled absorption heat pump heating, and cylinder cutting heating, making it difficult to unify the modeling using traditional methods; local optimal trap: under complex constraints, the algorithm is prone to fall into local optimality, causing the allocation results to deviate from the global optimality; insufficient convergence speed and stability: when optimizing large-scale multiple units, the computational efficiency is low and the stability is poor. Summary of the Invention

[0004] In order to solve the problems existing in the above-mentioned prior art, the purpose of the present invention is to provide a load allocation method for a multi-unit multi-mode cogeneration power station based on an improved Harris Hawk optimization algorithm. Through the improved Harris Hawk optimization algorithm (HPHHO), the total coal consumption is minimized, and the electricity and heat loads are ensured to meet external demand and unit operation constraints, thereby improving the economy and operation efficiency of the cogeneration power station.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] A load dispatching method for a multi-unit multi-mode cogeneration power station based on an improved Harris Eagle optimization algorithm comprises the following steps:

[0007] Step 1: Establish the improved Harris Hawk optimization algorithm HPHHO;

[0008] Step 2: Conduct system analysis on the cogeneration power plant using different heating modes to obtain the characteristic equations of each unit under different heating modes;

[0009] Step 3: Perform variable operating condition calculations on each unit to obtain the energy consumption of the unit within its operating domain, and express the energy consumption characteristics of each unit in the form of a fitting function;

[0010] Step 4: Establish the objective function of the problem to be optimized as well as the equilibrium constraints and non-equilibrium constraints;

[0011] Step 5: Use the improved Harris Hawk optimization algorithm HPHHO to optimize the load of each unit in the cogeneration power station and obtain the load distribution optimization result.

[0012] The improved Harris Hawk optimization algorithm HPHHO described in step 1 simulates the hunting process of a Harris Hawk in nature. The specific improvements are as follows:

[0013] The reverse learning strategy OBL generates a new position relative to the current population position. For each individual position in the population, i,j , its position after reverse learning OBLPositions i,j The calculation is as follows:

[0014] When the boundary number Boundary_no = 1:

[0015] OBLPositions i,j =LB+UB-Positions i,j (1)

[0016] When the number of boundaries Boundary_no>1:

[0017] OBLPositions i,j =LB j +UB j -Positions i,j (2)

[0018] In the formula, UB and LB represent the upper and lower boundaries of the variable respectively, UB j , LB j Represent the upper and lower boundaries of the j-th dimension variable of the prey respectively; merge the original population after reverse learning, calculate the individual fitness, record the best individual, and select the top N better individuals to form a new population;

[0019] Introducing dynamic escape probability S p , dynamically adjust the algorithm's search strategy according to the number of iterations, S p Update using formula (3):

[0020] S p =w·S p ·(1-S p ) (3)

[0021] Where w is a constant, set to 4;

[0022] The prey's escape energy is calculated as follows:

[0023]

[0024] Where E represents the escape energy of the prey, E0 is the initial escape energy, which is a random number in [-1,1]; t represents the number of iterations, and T is the maximum number of iterations;

[0025] By combining the global search capability of the differential evolution algorithm DE and the local search capability of the Harris Hawk algorithm HHO, the overall search performance of the improved Harris Hawk optimization algorithm is improved, and the optimization capability for complex optimization problems is enhanced; the population N is divided into two subgroups: X DE and X HHO , the number of subgroups are:

[0026] Using smoothing technology, X HHO and X DE The intermediate dimensions of the subgroups are smoothed:

[0027] (X HHOi,j -X HHOi,i-j )·(X HHOi,j+1 +1-X HHOi,j )<0 (4)

[0028] (X DEi,j -X DEi,i-j )·(X DEi,j+1 +1-X DEi,j )<0 (5)

[0029] When any of the above conditions are met:

[0030] X HHOi,j =0.5·(X HHOi,i-j +X HHOi,j+1 ) (6)

[0031] X DEi,j =0.5·(X DEi,i-j +X DEi,j+1 ) (7)

[0032] (1) For X DE Part, the update formula for the j-th dimension position of individual i is:

[0033]

[0034] Among them, Rabbit (j) represents the position of the prey in the jth dimension, F is the scaling factor; CR is the crossover rate; kkk is the index of the random permutation, j rand is a randomly chosen dimension index;

[0035] (2) For XHHO Part, in the exploration phase, when |E|≥1 and q<0.5:

[0036] X HHO,i =X HHO,i +F·(Rabbit-X HHOkk(1) +X HHOkk(2) -X HHOkk(3) ) (9)

[0037] Where q is a parameter that controls the exploration phase strategy and is a random number in [0,1]; kk is the index of the random permutation;

[0038] (3) For X HHO Part of the development phase is to introduce Levy flight to enhance local escape capabilities and global search diversity.

[0039] When 0.5≤|E|<1, and S p When the lag is less than 0.5, Harris's hawk will surround the prey with a gradual rapid dive and soft encirclement method, gradually correcting its direction and position according to the prey's deceptive behavior, and finally choosing the best position to hunt:

[0040]

[0041] When |E|<0.5, and S p When the value is <0.5, Harris's Hawks surround their prey by using a hard encirclement method with progressive rapid dives, and try to shorten the average position distance between them and the target prey:

[0042]

[0043] Where, X (t+1) is the position of the Harris Hawk at the t+1th iteration, X (t) is the position of Harris Hawk at the tth iteration, X m(t) is the average position of the Harris Hawk population at the tth iteration; Y and Z are the updated positions of Harris Hawk individuals; X rabbit (t) is the prey position, J is the random jump intensity, which is a random number in [0,2]; S is a D-dimensional random vector with a value between [0,1]; F(.) is the fitness function; Levy(D) represents the Levy flight strategy.

[0044] In step 2, the characteristic equations of each unit under different heating modes are different:

[0045] There is a coupling relationship between the main steam, heat load, and electric load of the single-extraction heating cogeneration unit, which can be expressed as:

[0046] F(D 0,cn,i ,P e,cn,i ,Q h,cn,i)=0 (12)

[0047] Among them, D 0,cn,i is the main steam flow rate of the i-th single extraction unit, t / h; P e,cn,i , Q h,cn,i are the electric load and thermal load of the i-th single extraction heating unit, MW;

[0048] The heating unit coupled with an absorption heat pump extracts high-grade steam from the steam turbine to drive the heat pump. The heat pump recovers the low-temperature waste heat of the circulating condensate and performs primary heating of the return water from the heating network. The return water from the heating network is then secondary heated to the temperature required by the heat user through the heating network heater. The main steam, heat load, and electrical load of the unit also have a coupled relationship, which can be expressed as a function:

[0049] F(D 0,AHP,i ,P e,AHP,i ,Q h,AHP,i )=0 (13)

[0050] Among them, D 0,AHP,i is the main steam flow rate of the i-th coupled absorption heat pump heating unit, t / h; P e,AHP,i , Q h,AHP,i are the electrical load and thermal load of the i-th coupled absorption heat pump heating unit, in MW. For a heating unit coupled with an absorption heat pump, its thermal load consists of two parts:

[0051] Q h,AHP,i =Q h,hp,i +Q h,cq,i (14)

[0052] Among them, Q h,hp,i , Q h,cq,i are the heat pump heat load and extraction steam heat load of the i-th coupled absorption heat pump heating unit, MW;

[0053] The heat pump heat load satisfies the following equation:

[0054] Q h,hp,i =Q g,i +Q n,i (15)

[0055] Among them, Q g,i , Q n,i Divided into heat load for driving heat pump and waste heat recovery of circulating condensing water, MW;

[0056] When the water flow rate of the heating network, the pressure and temperature of the water entering and leaving the heat pump are known, the heat load of the heat pump can be calculated by the following formula:

[0057] Q h,hp,i =D user ·(h hp,in -hhp,out ) / 3.6 (16)

[0058] Among them, D user is the water flow rate of the heating network, t / h; h hp,in , h hp,out The enthalpy values ​​of the water entering and leaving the heat pump are obtained by calculating the pressure and temperature. The temperature of the water entering the heat pump is 50°C and the pressure is 0.1 MPa. The temperature of the water leaving the heat pump is 70°C and the pressure is 0.1 MPa.

[0059] The coefficient of performance of an absorption heat pump is expressed by the following equation:

[0060]

[0061] Where COP,i is the heat pump coefficient of performance of the i-th coupled absorption heat pump heating unit. When the heat pump coefficient of performance COP,i and the calculated heat pump heat load value are known, the heat load of the driving heat pump is obtained:

[0062] Q g,i =Q h,hp,i / COP ,i (18)

[0063] Cut off the low-pressure cylinder heating unit, and its operating mode is to determine the electricity by heat. Therefore, the heat load and electricity load are expressed by functions as follows:

[0064] F(Q h,LPCC,i ,P e,LPCC,i )=0 (19)

[0065] Among them, Q h,LPCC,i 、P e,LPCC,i are the heat load and electricity load of the i-th low-pressure cylinder heating unit, respectively, in MW.

[0066] The energy consumption characteristics of each unit in step 3 are expressed by polynomials:

[0067]

[0068] Among them, B i is the standard coal consumption of the i-th unit, t / h; g is the coefficient before each term of the polynomial; j and f are the polynomial coefficients; m and n represent the upper limits of j and f respectively.

[0069] The objective function in step 4 is to minimize the total coal consumption of the cogeneration power station, which is expressed as:

[0070]

[0071] The balance constraint in step 4 is that the total electrical load and thermal load of the cogeneration power station must meet external demand:

[0072]

[0073] P e , Q h are the total electricity load and total heat load supplied by the cogeneration power station respectively; P e,i , Q h,i represent the electrical load and thermal load of the i-th cogeneration unit respectively.

[0074] The unbalanced constraint in step 4 is that the electrical load and thermal load of each unit must be limited to its feasible operating region:

[0075] P e,i,min ≤P e,i ≤P e,i,max (twenty four)

[0076] Q h,i,min ≤Q h,i ≤Q h,i,max (25)

[0077] Among them, P e,i,max 、P e,i,min , Q h,i,max , Q h,i,min are the upper and lower limits of the electric load and thermal load of the i-th unit respectively.

[0078] The proposed method combines the global search capabilities of the differential evolution algorithm with the population diversity enhancement mechanism of the reverse learning strategy to effectively improve the load distribution efficiency and accuracy of complex cogeneration systems. Compared with existing technologies, it has the following advantages:

[0079] (1) Improved global optimization capability of Harris Hawk optimization algorithm: combining DE and HHO algorithms to avoid falling into local optimality and improve the quality of solution;

[0080] (2) Efficient convergence: reverse learning and smoothing techniques enhance population diversity, and dynamic parameter adjustment accelerates the convergence process;

[0081] (3) Multi-mode compatibility: supports unified optimization of multiple heating modes such as single extraction, absorption heat pump, and removal of low-pressure cylinders to adapt to complex operating scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] Figure 1 It is a flowchart based on the improved Harris Hawk optimization algorithm HPHHO.

[0083] Figure 2 This is the feasible operating domain of the cogeneration unit, where (a) is steam extraction heating, (b) is coupled absorption heat pump heating, and (c) is cylinder cutting heating.

[0084] Figure 3This is a convergence process diagram of the improved Harris Hawk optimization algorithm HPHHO, where (a) is the fitness value convergence diagram of HHO and HPHHO algorithms as the number of iterations increases from 0 to 200, and (b) is a comparison diagram of the best fitness values ​​of HHO and HPHHO algorithms after 100 iterations. DETAILED DESCRIPTION

[0085] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0086] The present invention proposes a load allocation method for a cogeneration power plant that is adapted to multi-unit multi-mode operation. By using an improved Harris Hawk Optimization (HPHHO) algorithm, the method minimizes the total coal consumption of the cogeneration power plant while ensuring that the electrical and thermal loads meet external demands and meet the boundary constraints of unit safe operation, thereby improving the economic efficiency of the cogeneration power plant. The method specifically comprises the following steps:

[0087] Step 1: If Figure 1 As shown, an improved Harris Hawk optimization algorithm (HPHHO) is established to improve the efficiency and stability of the algorithm in solving complex optimization problems;

[0088] The improved Harris Hawk Optimization algorithm (HPHHO) simulates the hunting process of a Harris Hawk in nature. The algorithm consists of two phases: global exploration and local exploitation. In this algorithm, the Harris Hawk's position is considered as a candidate solution, and the best candidate solution in the iteration is the prey. The specific process is as follows:

[0089] In order to enhance the global search capability, the reverse learning strategy (OBL) is used to generate new positions relative to the current population position. For each individual position in the population, i,j , its position after reverse learning OBLPositions i,j The calculation is as follows:

[0090] When the boundary number Boundary_no = 1:

[0091] OBLPositions i,j =LB+UB-Positions i,j (1)

[0092] When the number of boundaries Boundary_no>1:

[0093] OBLPositions i,j =LB j +UB j -Positions i,j (2)

[0094] In the formula, UB and LB represent the upper and lower boundaries of the variable respectively; UB j , LB j Represent the upper and lower boundaries of the j-th dimension variable of the prey respectively; merge the original population after reverse learning, calculate the individual fitness, record the best individual, and select the top N better individuals to form a new population;

[0095] In order to enable the algorithm to conduct extensive exploration in the early stage and to conduct detailed development in the later stage, and to improve the search efficiency and accuracy of the algorithm, the dynamic escape probability S is introduced. p , dynamically adjust the algorithm's search strategy according to the number of iterations, S p Update using formula (3):

[0096] S p =w·S p ·(1-S p ) (3)

[0097] Here, w is a constant, set to 4.

[0098] The escape energy of the prey is calculated as follows:

[0099]

[0100] Where E represents the escape energy of the prey, E0 is the initial escape energy, which is a random number in [-1,1]; t represents the number of iterations, and T is the maximum number of iterations;

[0101] By combining the global search capability of the differential evolution algorithm (DE) and the local search capability of the Harris Hawk algorithm (HHO), the overall search performance of the improved Harris Hawk optimization algorithm is improved, and the algorithm's ability to find the best solution for complex optimization problems is enhanced. The population N is divided into two subgroups: X DE and X HHO , the number of subgroups are:

[0102] In order to optimize individual positions, reduce local fluctuations, and improve the convergence speed and stability of the improved Harris Eagle optimization algorithm, a smoothing technique is used to HHO and X DE The intermediate dimensions of the subgroup (2nd to D-1th dimensions) are smoothed:

[0103] (X HHOi,j -X HHOi,i-j )·(X HHOi,j+1 +1-X HHOi,j )<0 (4)

[0104] (X DEi,j -X DEi,i-j )·(X DEi,j+1+1-X DEi,j )<0 (5)

[0105] When any of the above conditions are met:

[0106] X HHOi,j =0.5·(X HHOi,i-j +X HHOi,j+1 ) (6)

[0107] X DEi,j =0.5·(X DEi,i-j +X DEi,j+1 ) (7)

[0108] (1) For X DE Part, the update formula for the j-th dimension position of individual i is:

[0109]

[0110] Among them, Rabbit (j) represents the position of the prey in the jth dimension, F is the scaling factor, which is 0.5; CR is the crossover rate, which is 0.9; kkk is the index of the random permutation, j rand is a randomly chosen dimension index;

[0111] (2) For X HHO Part, in the exploration phase, when |E|≥1 and q<0.5:

[0112] X HHO,i =X HHO,i +F·(Rabbit-X HHOkk(1) +X HHOkk(2) -X HHOkk(3) ) (9)

[0113] Where q is a parameter that controls the exploration phase strategy and is a random number in [0,1]; kk is the index of the random permutation;

[0114] (3) For X HHO Part of the development phase is to introduce Levy flight to enhance local escape capabilities and global search diversity.

[0115] When 0.5≤|E|<1, and S p When the lag is less than 0.5, Harris's hawk will surround the prey with a gradual rapid dive and soft encirclement method, gradually correcting its direction and position according to the prey's deceptive behavior, and finally choosing the best position to hunt:

[0116]

[0117] When |E|<0.5, and S pWhen the value is <0.5, Harris's Hawks surround their prey by using a hard encirclement method with progressive rapid dives, and try to shorten the average position distance between them and the target prey:

[0118]

[0119] Where, X (t+1) is the position of the Harris Hawk at the t+1th iteration, X (t) is the position of Harris Hawk at the tth iteration, X m(t) is the average position of the Harris Hawk population at the tth iteration; Y and Z are the updated positions of Harris Hawk individuals; X rabbit (t) is the prey position, J is the random jump intensity, which is a random number in [0,2]; S is a D-dimensional random vector with a value between [0,1]; F(.) is the fitness function; Levy(D) represents the Levy flight strategy;

[0120] In summary, the differential evolution DE algorithm is used in the standard Harris Hawk optimization algorithm to improve the overall search performance of the algorithm; the reverse learning strategy (OBL) is used to increase population diversity; the smoothing technology is used to optimize individual positions, reduce local fluctuations, and improve the convergence speed and stability of the algorithm; the dynamic escape probability is introduced to improve the search efficiency and accuracy of the algorithm, forming an improved Harris Hawk optimization algorithm (HPHHO) model.

[0121] Step 2: Conduct system analysis on the cogeneration power plant using different heating modes to obtain the characteristic equations of each unit under different heating modes;

[0122] There is a coupling relationship between the main steam, heat load, and electric load of the single-extraction heating cogeneration unit, which can be expressed by the function:

[0123] F(D 0,cn,i ,P e,cn,i ,Q h,cn,i )=0 (12)

[0124] Among them, D 0,cn,i is the main steam flow rate of the i-th single extraction unit, t / h; P e,cn,i , Q h,cn,i are the electric load and thermal load of the i-th single extraction heating unit, MW;

[0125] A heating unit coupled with an absorption heat pump extracts high-grade steam from the turbine to drive the heat pump. The heat pump recovers the low-temperature waste heat from the circulating condensate and performs primary heating of the return water from the heating network. This return water is then secondary heated to the user's required temperature by the heating network heater. The unit's main steam, heat load, and electrical load are also coupled and can be expressed as:

[0126] F(D 0,AHP,i ,Pe,AHP,i ,Q h,AHP,i )=0 (13)

[0127] Among them, D 0,AHP,i is the main steam flow rate of the i-th coupled absorption heat pump heating unit, t / h; P e,AHP,i , Q h,AHP,i are the electrical load and thermal load of the i-th coupled absorption heat pump heating unit, in MW. For a heating unit coupled with an absorption heat pump, its thermal load consists of two parts:

[0128] Q h,AHP,i =Q h,hp,i +Q h,cq,i (14)

[0129] Among them, Q h,hp,i , Q h,cq,i are the heat pump heat load and extraction steam heat load of the i-th coupled absorption heat pump heating unit, MW;

[0130] The heat pump heat load satisfies the following equation:

[0131] Q h,hp,i =Q g,i +Q n,i (15)

[0132] Among them, Q g,i , Q n,i Divided into heat load for driving heat pump and waste heat recovery of circulating condensing water, MW;

[0133] When the water flow rate of the heating network, the pressure and temperature of the water entering and leaving the heat pump are known, the heat load of the heat pump can be calculated by the following formula:

[0134] Q h,hp,i =D user ·(h hp,in -h hp,out ) / 3.6 (16)

[0135] Among them, D user is the water flow rate of the heating network, t / h; h hp,in , h hp,out The enthalpy values ​​of the water entering and leaving the heat pump are obtained by calculating the pressure and temperature. The temperature of the water entering the heat pump is 50°C and the pressure is 0.1 MPa. The temperature of the water leaving the heat pump is 70°C and the pressure is 0.1 MPa.

[0136] The coefficient of performance of an absorption heat pump can be expressed by the following equation:

[0137]

[0138] Where COP,i is the heat pump coefficient of performance of the i-th coupled absorption heat pump heating unit, which is taken as 1.45 according to the design value. After knowing the heat pump coefficient of performance COP,i and the calculated heat pump heat load value, the driving heat pump heat load can be obtained:

[0139] Q g,i =Q h,hp,i / COP ,i (18)

[0140] The operation mode of the heating unit with the low-pressure cylinder cut off (referred to as cylinder cut off) is to determine the electricity by heat. Therefore, the heat load and electricity load can be expressed by the function:

[0141] F(Q h,LPCC,i ,P e,LPCC,i )=0 (19)

[0142] Among them, Q h,LPCC,i 、P e,LPCC,i are the heat load and electricity load of the i-th cylinder-cutting heating unit, MW;

[0143] Step 3: Perform variable operating condition calculations on each unit to obtain the energy consumption of the unit within its operating domain, and express the energy consumption characteristics of each unit in the form of a fitting function;

[0144] The energy consumption characteristics of each unit can be expressed by a polynomial:

[0145]

[0146] Among them, B i is the standard coal consumption of the i-th unit, t / h; g is the coefficient before each term of the polynomial; j and f are the polynomial coefficients; m and n represent the upper limits of j and f respectively.

[0147] Step 4: Establish the objective function of the problem to be optimized as well as the equilibrium constraints and non-equilibrium constraints;

[0148] Taking the minimum total coal consumption of the cogeneration power station as the objective function, it can be expressed as:

[0149]

[0150] The balance constraint is that the total electrical load and thermal load of the cogeneration power station must meet external demand:

[0151]

[0152] P e , Q h are the total electricity load and total heat load supplied by the cogeneration power station; P e,i , Q h,i represent the electrical load and thermal load of the i-th cogeneration unit respectively.

[0153] The unbalanced constraint condition is: the electrical load and thermal load of each unit must be limited to Figure 2 Within the feasible operating domain shown: Figure 2 (a) is the operating domain of a single-extraction cogeneration unit, and ABCD encloses the flexible adjustment area for thermal load under the intermediate extraction steam heating mode. Points A and D are the minimum and maximum output electrical load conditions under pure condensing conditions, respectively; points C and B are the maximum extraction steam flow conditions under minimum fuel flow and full load conditions, respectively; DC and AB are the thermal load characteristics corresponding to different extraction steam flows under minimum fuel flow and full load conditions, respectively; and BC represents the thermal load characteristics achieved by maintaining the minimum condensing steam flow rate in the low-pressure cylinder under different loads. Figure 2 (b) is the flexible adjustment area of ​​thermal power load after the single extraction unit is coupled with the absorption heat pump. Figure 2 Based on (a), the unit operation domain is expanded from ABCD to AEFCD; Figure 2 Middle (c) is the operating domain of the cut-cylinder unit. When the unit is in the cut-cylinder mode, the heat is used to determine the electricity, and the electric heating load is limited to the GH boundary line. Points G and H are the maximum heating load conditions under full boiler load and minimum fuel quantity, respectively.

[0154] P e,i,min ≤P e,i ≤P e,i,max (twenty four)

[0155] Q h,i,min ≤Q h,i ≤Q h,i,max (25)

[0156] Among them, P e,i,max 、P e,i,min , Q h,i,max , Q h,i,min are the upper and lower limits of the electric load and thermal load of the i-th unit respectively.

[0157] Step 5: Use the improved Harris Hawk optimization algorithm (HPHHO) established in step 1 to perform load economic allocation on the cogeneration system with multiple units and multiple modes.

[0158] Taking the electrical load and thermal load of each unit as optimization variables, we select an appropriate number of optimization variables X to form the X-dimensional position vector of each prey, which represents the current electrical and thermal load of each unit. The Harris Hawk searches for the optimal position in the X-dimensional search space. The specific steps are as follows:

[0159] (1) Randomly generate N search individuals, each of which represents a load distribution scheme:

[0160] X i =[P e1 ,Pe2 ,...,Q h1 ,Q h2 ,...]

[0161] Among them, P e is the electrical load, Q h is the heat load, satisfying the constraints in step 4; at the same time, generate the OBL population according to equations (1) and (2) in step 1;

[0162] (2) The total coal consumption of the cogeneration power plant is used as the objective function to evaluate the fitness;

[0163] (3) Merge the main population and the OBL population and select the top N best solutions;

[0164] (4) Smoothing the load of each individual according to equations (4)-(7);

[0165] (5) Update the prey's escape energy E and the escape probability S of formula (3) p ;

[0166] (7) To X DE The subpopulation performs crossover and mutation operations according to step 1 (8);

[0167] (8) To X HHO The subpopulation updates its position according to step 1 (9)-(11);

[0168] (8) Merge X DE and X HHO The two populations re-evaluate their fitness values ​​and update the global optimal solution (prey position) and minimum coal consumption (prey escape energy);

[0169] (9) If the maximum number of iterations T is reached or the optimal solution is stable, output the result; otherwise, return to (1).

[0170] The convergence process of the standard Harris Hawk optimization algorithm (HHO) and the improved Harris Hawk optimization algorithm (HPHHO) is as follows Figure 3 As shown in (a) and (b), the population size of both algorithms is 50, and the number of iterations is within 200. Figure 3 From the results in (a) and (b), it can be seen that the improved Harris Hawk optimization algorithm can improve the convergence speed of the algorithm, and the HPHHO algorithm has better convergence effect and stability than the HHO algorithm.

[0171] This paper addresses the problems of low load distribution efficiency and susceptibility to local optimization in existing cogeneration systems due to the diverse types of units and complex operating modes. By integrating the global search capabilities of the differential evolution algorithm (DE) with the local development capabilities of the Harris Hawk algorithm (HHO), this method enhances population diversity through an outward-facing learning strategy (OBL), employs smoothing techniques to reduce local fluctuations, and dynamically adjusts escape probabilities to balance the algorithm's exploration and development phases, significantly improving global optimization capabilities and convergence speed.

Claims

1. A load distribution method for a multi-unit multi-mode cogeneration power station based on an improved Harris Hawk optimization algorithm, characterized by: The following steps are involved: Step 1: Establish the improved Harris Hawk optimization algorithm HPHHO; Step 2: Conduct system analysis on the cogeneration power plant using different heating modes to obtain the characteristic equations of each unit under different heating modes; Step 3: Perform variable operating condition calculations on each unit to obtain the energy consumption of the unit within its operating domain, and express the energy consumption characteristics of each unit in the form of a fitting function; Step 4: Establish the objective function of the problem to be optimized as well as the equilibrium constraints and non-equilibrium constraints; Step 5: Use the improved Harris Hawk optimization algorithm HPHHO to optimize the load of each unit in the cogeneration power station and obtain the load distribution optimization result.

2. The load distribution method for a multi-unit multi-mode cogeneration power station based on an improved Harris Hawk optimization algorithm according to claim 1, characterized in that: The improved Harris Hawk optimization algorithm HPHHO described in step 1 simulates the hunting process of a Harris Hawk in nature. The specific improvements are as follows: The reverse learning strategy OBL generates a new position relative to the current population position. For each individual position in the population, i ,j, its position after reverse learning OBLPositions i,j The calculation is as follows: When the boundary number Boundary_no = 1: OBLPositions i,j =LB+UB-Positions i,j (1) When the number of boundaries Boundary_no>1: OBLPositions i,j =LB j +UB j -Positions i,j (2) In the formula, UB and LB represent the upper and lower boundaries of the variable respectively, UB j , LB j Represent the upper and lower boundaries of the j-th dimension variable of the prey respectively; merge the original population after reverse learning, calculate the individual fitness, record the best individual, and select the top N better individuals to form a new population; Introducing dynamic escape probability S p , dynamically adjust the algorithm's search strategy according to the number of iterations, S p Update using formula (3): S p =w·S p ·(1-S p ) (3) Where w is a constant, set to 4; The escape energy of the prey is calculated as follows: Where E represents the escape energy of the prey, E0 is the initial escape energy, which is a random number in [-1,1]; t represents the number of iterations, and T is the maximum number of iterations; By combining the global search capability of the differential evolution algorithm DE and the local search capability of the Harris Hawk algorithm HHO, the overall search performance of the improved Harris Hawk optimization algorithm is improved, and the optimization capability for complex optimization problems is enhanced; the population N is divided into two subgroups: X DE and X HHO , the number of subgroups are: Using smoothing technology, X HHO and X DE The intermediate dimensions of the subgroups are smoothed: (X HHOi,j -X HHOi,i-j )·(X HHOi,j+1 +1-X HHOi,j )<0 (4) (X DEi,j -X DEi,i-j )·(X DEi,j+1 +1-X DEi,j )<0 (5) When any of the above conditions are met: X HHOi,j =0.5·(X HHOi,i-j +X HHOi,j+1 ) (6) X DEi,j =0.5·(X DEi,i-j +X DEi,j+1 ) (7) (1) For X DE Part, the update formula for the j-th dimension position of individual i is: Among them, Rabbit (j) represents the position of the prey in the jth dimension, F is the scaling factor; CR is the crossover rate; kkk is the index of the random permutation, j rand is a randomly chosen dimension index; (2) For X HHO Part, in the exploration phase, when |E|≥1 and q<0.5: X HHO,i =X HHO,i +F·(Rabbit-X HHOkk(1) +X HHOkk(2) -X HHOkk(3) ) (9) Where q is a parameter that controls the exploration phase strategy and is a random number in [0,1]; kk is the index of the random permutation; (3) For X HHO Part of the development phase is to introduce Levy flight to enhance local escape capabilities and global search diversity. When 0.5≤|E|<1, and S p When the lag is less than 0.5, Harris's hawk will surround the prey with a gradual rapid dive and soft encirclement method, gradually correcting its direction and position according to the prey's deceptive behavior, and finally choosing the best position to hunt: When |E|<0.5, and S p When the value is <0.5, Harris's Hawks surround their prey by using a hard encirclement method with progressive rapid dives, and try to shorten the average position distance between them and the target prey: Where, X (t+1) is the position of the Harris Hawk at the t+1th iteration, X (t) is the position of Harris Hawk at the tth iteration, X m(t) is the average position of the Harris Hawk population at the tth iteration; Y and Z are the updated positions of Harris Hawk individuals; X rabbit (t) is the prey position, J is the random jump intensity, which is a random number in [0,2]; S is a D-dimensional random vector with a value between [0,1]; F(.) is the fitness function; Levy(D) represents the Levy flight strategy.

3. The load distribution method for a multi-unit multi-mode cogeneration power station based on an improved Harris Hawk optimization algorithm according to claim 1, characterized in that: In step 2, the characteristic equations of each unit under different heating modes are different: There is a coupling relationship between the main steam, heat load, and electric load of the single-extraction heating cogeneration unit, which can be expressed as: F(D 0,cn,i ,P e,cn,i ,Q h,cn,i )=0 (12) Among them, D 0,cn,i is the main steam flow rate of the i-th single extraction unit, t / h; P e,cn,i , Q h,cn,i are the electric load and thermal load of the i-th single extraction heating unit, MW; The heating unit coupled with an absorption heat pump extracts high-grade steam from the steam turbine to drive the heat pump. The heat pump recovers the low-temperature waste heat of the circulating condensate and performs primary heating of the return water from the heating network. The return water from the heating network is then secondary heated to the temperature required by the heat user through the heating network heater. The main steam, heat load, and electrical load of the unit also have a coupled relationship, which can be expressed as a function: F(D 0,AHP,i ,P e,AHP,i ,Q h,AHP,i )=0 (13) Among them, D 0,AHP,i is the main steam flow rate of the i-th coupled absorption heat pump heating unit, t / h; P e,AHP,i , Q h,AHP,i are the electrical load and thermal load of the i-th coupled absorption heat pump heating unit, in MW. For a heating unit coupled with an absorption heat pump, its thermal load consists of two parts: Q h,AHP,i =Q h,hp,i +Q h,cq,i (14) Among them, Q h,hp,i , Q h,cq,i are the heat pump heat load and extraction steam heat load of the i-th coupled absorption heat pump heating unit, MW; The heat pump heat load satisfies the following equation: Q h,hp,i =Q g,i +Q n,i (15) Among them, Q g,i , Q n,i Divided into heat load for driving heat pump and waste heat recovery of circulating condensing water, MW; When the water flow rate of the heating network, the pressure and temperature of the water entering and leaving the heat pump are known, the heat load of the heat pump can be calculated by the following formula: Q h,hp,i =D user ·(h hp,in -h hp,out ) / 3.6 (16) Among them, D user is the water flow rate of the heating network, t / h; h hp,in , h hp,out The enthalpy values ​​of the water entering and leaving the heat pump are obtained by calculating the pressure and temperature. The temperature of the water entering the heat pump is 50°C and the pressure is 0.1 MPa. The temperature of the water leaving the heat pump is 70°C and the pressure is 0.1 MPa. The coefficient of performance of an absorption heat pump is expressed by the following equation: Where COP,i is the heat pump coefficient of performance of the i-th coupled absorption heat pump heating unit. When the heat pump coefficient of performance COP,i and the calculated heat pump heat load value are known, the heat load of the driving heat pump is obtained: Q g,i =Q h,hp,i / COP ,i (18) Cut off the low-pressure cylinder heating unit, and its operating mode is to determine the electricity by heat. Therefore, the heat load and electricity load are expressed by functions as follows: F(Q h,LPCC,i ,P e,LPCC,i )=0 (19) Among them, Q h,LPCC,i 、P e,LPCC,i are the heat load and electricity load of the i-th low-pressure cylinder heating unit, respectively, in MW.

4. The load distribution method for a multi-unit multi-mode cogeneration power station based on an improved Harris Hawk optimization algorithm according to claim 1, characterized in that: The energy consumption characteristics of each unit in step 3 are expressed by polynomials: Among them, B i is the standard coal consumption of the i-th unit, t / h; g is the coefficient before each term of the polynomial; j and f are the polynomial coefficients; m and n represent the upper limits of j and f respectively.

5. The load distribution method for a multi-unit multi-mode cogeneration power station based on an improved Harris Hawk optimization algorithm according to claim 1, characterized in that: The objective function in step 4 is to minimize the total coal consumption of the cogeneration power station, which is expressed as:

6. The load distribution method for a multi-unit multi-mode cogeneration power station based on an improved Harris Hawk optimization algorithm according to claim 1, characterized in that: The balance constraint in step 4 is that the total electrical load and thermal load of the cogeneration power station must meet external demand: P e , Q h are the total electricity load and total heat load supplied by the cogeneration power station; P e,i , Q h,i represent the electrical load and thermal load of the i-th cogeneration unit respectively.

7. The load distribution method for a multi-unit multi-mode cogeneration power station based on an improved Harris Hawk optimization algorithm according to claim 1, characterized in that: The unbalanced constraint in step 4 is that the electrical load and thermal load of each unit must be limited to its feasible operating region: P e,i,min ≤P e,i ≤P e,i,max (24) Q h,i,min ≤Q h,i ≤Q h,i,max (25) Among them, P e,i,max 、P e,i,min , Q h,i,max , Q h,i,min are the upper and lower limits of the electric load and thermal load of the i-th unit respectively.

Images