Configuration strategy and operation mode design under multi-objective optimisation framework in an off-grid optical storage microgrid

Under the grid-connected state, the following points need to be considered when the optical storage microgrid is in operation:

Reasonable scheduling of power generation units and energy storage equipment, efficient use of solar energy as much as possible under the support of the large power grid, and reduction of the loss of each device.

In the off-grid state, the photovoltaic storage microgrid needs to meet the following requirements when operating independently [19]:

A stable bus voltage and frequency need to be established within the microgrid system. At this time, the voltage and frequency within the system are controlled by the power generation unit and energy storage-related equipment, which need to meet the rated operating conditions of the power equipment.

The electricity cost of the optical storage microgrid consists of four parts: the cost of PV power generation, the cost of regulation of the storage device, the cost of exchanging electricity between the micro energy network and the main grid, and the cost of the policy subsidy, which is expressed as [20]: (1) W=ww∑t=1TPw,tΔt+ws∑t=1TPs,tΔt+wg∑t=1TPg,tΔt+wb∑t=1TΔPb,tΔt+∑t=1TΔPtΔt⋅Ct−wδ⋅δE\[\begin{align} & W={{w}_{w}}\sum\limits_{t=1}^{T}{{{P}_{w,t}}}\Delta t+{{w}_{s}}\sum\limits_{t=1}^{T}{{{P}_{s,t}}}\Delta t+{{w}_{g}}\sum\limits_{t=1}^{T}{{{P}_{g,t}}}\Delta t \\ & +{{w}_{b}}\sum\limits_{t=1}^{T}{\Delta }{{P}_{b,t}}\Delta t+\sum\limits_{t=1}^{T}{\Delta }{{P}_{t}}\Delta t\cdot {{C}_{t}}-{{w}_{\delta }}\cdot \delta E \end{align}\]

In the formula, P_w,t, P_s,t, P_g,t, ΔP_b,t for t moments of photovoltaic, gas unit, energy storage devices involved in the dispatch of the output value; w_w, w_s for photovoltaic, gas unit, respectively, the average cost of power generation, w_b for the average cost of energy storage device. ΔP_t is the exchange power between the photovoltaic and storage microgrid and the main grid at Δt time, C_t is the power purchase and sale price at t time, δE is the self-consumption power generation in the region, w_δ is the policy subsidy, and T is the scheduling cycle.

The exchange power between the optical storage microgrid and the main grid is the difference between the load power and the power provided by the optical storage microgrid. When the energy storage device is in the discharge state, it is equivalent to the power supply. When it is in the charging state, it is equivalent to the load, i.e: (2) ΔPt=Pload,t−Pw,t−Ps,t−Pg,t+Pb,t\[\Delta {{P}_{t}}={{P}_{load,t}}-{{P}_{w,t}}-{{P}_{s,t}}-{{P}_{g,t}}+{{P}_{b,t}}\]

Where P_load,t is the load power of the optical storage microgrid at t moments, P_b,t is the charging and discharging power of the energy storage device involved in scheduling at t moments. And meet: (3) Pb,t=ΔPb,tXYt={ Pcha,t,XYt=10,XYt=0−Pdis,t,XYt=−1${{P}_{b,t}}=\Delta {{P}_{b,t}}X{{Y}_{t}}=\left\{ \begin{array}{*{35}{l}} {{P}_{cha,t}},X{{Y}_{t}}=1 \\ 0,X{{Y}_{t}}=0 \\ -{{P}_{dis,t}},X{{Y}_{t}}=-1 \\ \end{array} \right.$

Where P_cha,t and P_dis,t are the charging and discharging power of the energy storage device at the moment of t, respectively. XY_t is the charging and discharging state of the energy storage device at the moment of t, XY_t ∈ {–1,0,1}.

When the photovoltaic power and the charging and discharging power of the energy storage device is not enough to supply power to all loads, it is necessary to purchase power from the main grid at the time of purchase price. When the sum of the photovoltaic and energy storage power of the microenergy network exceeds the load power, electricity is to be sold from the main grid at the electricity sales tariff for that time period. With the optimisation objective of minimising the average electricity consumption cost of the loads in the microenergy network area, the economic scheduling objective function is constructed based on the multi-objective optimisation algorithm as [21]: (4) w¯=W∑t=1TPload,tΔt\[\bar{w}=\frac{W}{\sum\limits_{t=1}^{T}{{{P}_{load,t}}}\Delta t}\]

Off-grid state, the optical storage microgrid as a whole, through the scheduling of its internal optical-electricity-storage devices of the output value, so that the overall benefits to achieve the optimal. The overall benefits include not only economic benefits, but also environmental benefits brought by the use of renewable energy, i.e. the higher the utilisation rate of renewable energy, the better the environmental benefits. Considering the balance of economy (f₁), reliability (f₂) and environmental benefits (f₃), the configuration optimisation model of off-grid optical storage microgrid is established [22]: (5) cx=Px,1Δt∑t=1rPx,,Δt0≤cx≤1\[{{c}_{x}}=\frac{{{P}_{x,1}}\Delta t}{\sum\limits_{t=1}^{r}{{{P}_{x,}}},\Delta t}0\le {{c}_{x}}\le 1\]

Where c_x is the utilisation rate of wind power and photovoltaic energy respectively.

To improve the overall efficiency of off-grid optical storage microgrid scheduling, it is necessary to comprehensively consider the economic and environmental benefits of optical storage microgrids, so as to achieve the optimal values of w¯max\[{{\overline{w}}^{\max }}\] (maximising economic benefits), cwmax\[{{c}_{w}}^{\max }\] (maximising the reliability of microgrid operation), and csmax\[{{c}_{s}}^{\max }\] (maximising environmental benefits), respectively. As the objective function is not uniform, in order to reflect the degree of optimisation of the multi-objective function, it needs to be normalised, i.e.: (6) f1=w¯−w¯minw¯max−w¯min¯\[{{f}_{1}}=\frac{\bar{w}-{{{\bar{w}}}^{\min }}}{\overline{{{{\bar{w}}}^{\max }}-{{{\bar{w}}}^{\min }}}}\] (7) f2=cwmax−cwcwmax−cwmin\[{{f}_{2}}=\frac{c_{w}^{\max }-{{c}_{w}}}{c_{w}^{\max }-c_{w}^{\min }}\] (8) f3=csmax−cscsmax−cs\[{{f}_{3}}=\frac{c_{s}^{\max }-{{c}_{s}}}{c_{s}^{\max }-{{c}_{s}}}\]

The economic benefits (f₁), reliability (f₂) and environmental benefits (f₃) as the objective functions to be sought can reflect the degree of optimisation of each optimisation objective, whose optimal value is the minimum value, and the value range is [0, 1].

Due to the optimisation of each objective function in the optimisation process of the factors related to each other, in the optimisation calculation will interact with each other, resulting in the overall degree of optimisation can not reach the optimal solution of each objective optimisation. The satisfaction weight coefficient method is too subjective, which may lead to unsatisfactory multi-objective optimisation results. In order to reflect the overall optimisation degree of multiple objectives, an overall satisfaction model is established, with fmin=[f1min,f2min,f3min]\[{{f}^{\min }}=[{{f}_{1}}^{\min },{{f}_{2}}^{\min },{{f}_{3}}^{\min }]\] as the ideal objective point, and the vector value with the smallest distance from this objective point is the optimal solution, thus converting the above multi-objective optimisation problem into a single-objective optimisation problem for solving. In addition, it is also possible to add the range of constraints desired by subjective decision-making to obtain a balanced optimisation model of economy, reliability and environmental benefits: (9) maxS=1−‖ f−fiin ‖=1−(f1−f1min)2+(f2−f2min)2+(f3−f3min)2s.t.f1≤F1,f2≤F2,f3≤F3\[\begin{align} & \max S=1-\left\| f-{{f}^{iin}} \right\| \\ & =1-\sqrt{{{\left( {{f}_{1}}-f_{1}^{\min } \right)}^{2}}+{{\left( {{f}_{2}}-f_{2}^{\min } \right)}^{2}}+{{\left( {{f}_{3}}-f_{3}^{\min } \right)}^{2}}} \\ & s.t.{{f}_{1}}\le {{F}_{1}},{{f}_{2}}\le {{F}_{2}},{{f}_{3}}\le {{F}_{3}} \\ \end{align}\]

where F₁, F₂, and F₃ are the upper limits of the variables set for subjective decision making, respectively.

The total energy storage capacity of the optical storage microgrid is limited. Even if it can meet the above operational objectives, the capacity of the system is still much smaller compared to the large power grid, and the output of the power generation units and other equipment contained within it needs to meet certain constraints, mainly in the following aspects:

Equipment output power constraints

For photovoltaic power supply, the output power is constrained by environmental conditions and equipment carrying capacity, and its output power range is: (10) Ppvmin≤Ppv(t)≤Ppvmax\[{{P}_{pv\min }}\le {{P}_{pv}}\left( t \right)\le {{P}_{pv\max }}\]

Where P_pvmin is the minimum output power of photovoltaic power generation, generally 0. P_pr(t) is the t moment power generation, P_pvmax is the maximum output power of photovoltaic power generation.

Energy storage converter is a bi-directional power mobility converter, mainly used for energy storage battery access system, due to its internal DC/AC converter circuit has a maximum power limit, the exchange power of the energy storage converter also has its own power limit: (11) PLimin≤PLi(t)≤PLimax\[{{P}_{Li\min }}\le {{P}_{Li}}\left( t \right)\le {{P}_{Li\max }}\] (12) PScmin≤PSc(t)≤PScmin\[{{P}_{Sc\min }}\le {{P}_{Sc}}\left( t \right)\le {{P}_{Sc\min }}\]

Where P_Limin and P_Limax are the permissible minimum and maximum power of the storage converter when the storage battery is discharged, and P_Li(t) is the output power of the storage converter at the time of t discharge. P_Scmin and P_Scmax are the permissible minimum and maximum power of the SCV when charging the storage battery, and P_Sc(t) is the input power of the SCV during charging.

In a photovoltaic and storage microgrid system, photovoltaic power units and storage batteries are connected to the bus via a converter, and the system avoids extreme operating conditions of local equipment by controlling the flow of energy. The loads studied in this paper are charging piles and lighting, and the analysis of the system is somewhat simplified. In developing a reasonable system for energy management, the first step is to determine the objectives to be controlled by this system and the principles it needs to follow, which mainly include the following three points:

Reducing light abandonment and improving energy utilisation and efficiency where system operating conditions allow.

The optical storage microgrid system follows the principle of conservation of energy, as shown in the following: (17) Ppv+Pb=PL\[{{P}_{pv}}+{{P}_{b}}={{P}_{L}}\]

Where P_pv is the output power of the photovoltaic power generation unit; P_b is the output power of the energy storage battery, a negative value means that the battery is in the charging state; P_L is the power required by the load.

According to the current operating state of the PV power generation unit and the energy storage battery, the converters connected to both are controlled to work in conjunction with each other to keep the system in an energy-balanced state, and its corresponding three-layer control architecture of the optical storage microgrid is shown in Fig. 1:

Physical layer

It mainly consists of local controllers and protection devices for each part of the optical storage microgrid. The local controller of the system is the equipment controller, which can complete the primary regulation of the system voltage and frequency, and, together with the protection equipment, it can achieve rapid recovery after failure.

Figure 1.

The hierarchical structure of the photovoltaic and energy storage microgrid system

With a good prioritisation of power usage, the power balance equation for the system is elicited: (18) PPV+Pbat+Pgrid=Pload\[{{P}_{PV}}+{{P}_{bat}}+{{P}_{grid}}={{P}_{load}}\]

Where P_PV, P_bat and P_grid refer to the power output or absorbed by the PV module, Li-ion battery and the public grid respectively to the system, and P_load refers to the power consumed by the load connected to the system. Among them:

When the value of P_PV is positive, it is considered that the photovoltaic module unit outputs power to the system.

In summary, there is when the system operates in the independent state: (19) PPV+Pbat=Pload\[{{P}_{PV}}+{{P}_{bat}}={{P}_{load}}\]

There are when the system is running in grid-connected state: (20) PPV+Pbat+Pgrid=Pload\[{{P}_{PV}}+{{P}_{bat}}+{{P}_{grid}}={{P}_{load}}\]

In order to make the system work smoother and the energy exchange more coordinated, the PV module output power, Li-ion battery SOC and Li-ion battery charging and discharging power are explained as follows:

When the PV module output power is less than the minimum output power threshold of the PV module, i.e., P_PV < P_{PV_min}, it is considered that there is no power output from the PV module, and the Boost converter on the PV side should be turned off at this time.

In actual operation, the lithium battery charging and discharging power will change according to the photovoltaic output power and load power changes, the range between zero and the lithium battery maximum charging and discharging power, that is, 0 < P_bat < P_{bat_max}.

Based on the above analysis and combined with the multi-objective configuration optimisation model for off-grid optical storage microgrids established in subsection 2.1, the flow of the proposed energy management algorithm for off-grid optical storage microgrids is shown in Fig. 2. Here, it should be noted that when the system satisfies the two constraints of P_PV ≤ P_{PV_min} and SOC<10%, the power is supplied to the load by the grid and charges the energy storage battery. If no judgement is made on the working state (charging or discharging mode) of the energy storage battery, it will cause the system’s working mode to switch to the battery discharging mode when the SOC of the energy storage battery is just greater than 10%. Then, it quickly enters the battery charging mode again (storage battery SOC < 10%), which in turn results in an oscillating state with frequent switching of operating modes (the same will happen in P_{PV_min} ≤ P_PV ≤ P_load with storage battery SOC < 10%). In order to avoid this phenomenon, the communication between the DSP and the battery management system (BMS) module can be used to determine the state of the storage battery when the system was working in the previous cycle, and decide whether the storage battery will continue to be in such a state in the next cycle. This effectively prevents the frequent switching of system operating modes and ensures that the energy management strategy of the system can enable the system to operate efficiently and stably in a certain operating mode.

Figure 2.

System energy management algorithm flow

According to the optimisation process of energy storage capacity configuration of off-grid photovoltaic storage microgrid, the more typical lead-acid energy density batteries (VRLA-B), lead-acid power density batteries (VRLA-CAP), lithium iron batteries (LFP), all-vanadium liquid current batteries (V-redox), and sodium-sulphur batteries (NaS) are five types of storage batteries as the configuration object of off-grid photovoltaic storage microgrids, the battery parameters will be accurate, and its battery specific parameters are shown in Table 1. Among them, C_E is the price of the battery, C_P is the price of DC/DC converter, C_B is the price of auxiliary devices, C_fp is the price of battery recycling, the unit is (¥/KW·h). 5 kinds of energy storage battery discharge efficiency between 0.7 ~ 0.85, the battery’s state of charge SOC value in the range of 0.1 ~ 0.9.

Where the initial SOC of the battery is set = 0.4, the battery undergoes complete charging and discharging once a day and operates for 365 days per year, resulting in a 3D spatial diagram of the configuration of energy storage and PV under known loads as shown in Fig. 3. From the figure, it can be seen that the load shortage rate σ_LPSP and the energy spillover ratio σ_EXC present a strong coupling relationship. Therefore, the optimal configuration effect of PV and energy storage can only be satisfied by reasonably selecting the values of σ_LPSP and σ_EXC.

Figure 3.

Optical storage joint configuration

From the PV energy efficiency maximisation model proposed in this paper, the load deficit rate σ_LPSP and energy spillover ratio σ_EXC of the system should be as small as possible. At the same time, according to the new energy power plant construction indicators in the load deficit rate σ_LPSP is less than 2%, the energy overflow ratio σ_EXC is generally 5% ~ 30%. From this, the appropriate values of σ_LPSP and σ_EXC are selected in order to obtain the optimal configuration capacity of PV and energy storage under specific load demand.

Since the energy storage capacity is affected by the storage charging and discharging efficiency, etc., different storage batteries have different capacities in the same PV installed system. So in this paper, we will first set σ_LPSP =0.015, σ_ESC =15%. In order to determine the minimum capacity of PV when this setting value is met, Figure 4 shows the effect of PV capacity on σ_LPSP, σ_EXC. From the figure, when the PV capacity is small, the size of the storage capacity will not change the value of σ_EXC, only when the PV capacity reaches a certain size, σ_EXC will gradually increase, and the size of its value will be reduced by the increase of the storage capacity. However, the value of σ_LPSP is opposite to the value of σ_EXC, due to the fact that the energy storage capacity is only affected by the PV and the grid in the long-time operation of the photovoltaic storage microgrid. When charging it by the grid is not considered, it means that the energy storage system only acts as a relay device for PV energy when the photovoltaic storage microgrid is in an off-grid state. Therefore, the magnitude of the σ_LPSP value is affected by the magnitude of the energy storage capacity in essentially the same way, and its value decreases with the increase of the PV capacity. Thus, in determining σ_LPSP =0.015 and σ_EXC =0.15, the minimum capacity of PV per hour is taken as 77kW.

Figure 4.

The influence of photovoltaic capacity on σ_LPSP and σ_EXC

Figure 5 shows a typical PV curve with an installed PV capacity of 285 kW. Since the PV output energy is time-limited, the total energy released during the power generation phase should satisfy its capacity requirements for each hour of the day.

Figure 5.

Typical photovoltaic curve with installed capacity of 285kW

Fig. 6 shows the curves of energy storage capacity E_rat versus σ_LPSP and σ_EXC for an installed PV capacity of 285kW fixed capacity. It can be seen that after the PV fixed capacity of 285kW, the increase of the energy storage capacity will decrease the values of σ_LPSP and σ_EXC. The values of σ_LPSP and σ_EXC will remain unchanged when the energy storage capacity increases to a certain value, and the configured energy storage capacity varies due to the different discharge intervals and discharge efficiencies of different energy storage systems.

Figure 6.

Influence of energy storage capacity of optical microgrid on σ_LPSP and σ_EXC

Table 2 shows the minimum configured capacity of each energy storage under the fixed capacity of PV, combined with the construction cost minimisation function. Where the energy storage discharge time, by the intersection of the load average and PV output curve to take the value, thus set (i = 0.1,γ = 0.2,t = 15.0) can be obtained for each energy storage system economic comparison results. In the same PV energy storage system, the energy storage capacity is satisfied, lead-acid energy density battery (VRLA-B) > all-vanadium liquid current battery (V-redox) = lead-acid power density battery (VRLA-cap) = lithium battery (LFP) = sodium-sulfur battery (NaS). In terms of average annual cost, all-vanadium liquid current batteries (V-redox) > sodium-sulfur batteries (NaS) > lithium batteries (LFP) > lead-acid energy density batteries (VRLA-B) > lead-acid power density batteries (VRLA-cap).

In order to investigate the capacity allocation optimisation strategy proposed in this paper and the effective line of the operation model, simulation verification is carried out in the following two scenarios:

Scenario 1

Using the method proposed in this paper, the microgrid is configured with an optical storage system, where the energy storage regulation space is set according to the real-time demand of the base station load. The base station optical storage micro-network, and base station micro-network and distribution network for energy sharing operation between the base station optical storage micro-network operator and the distribution network operator, through the multi-objective optimization function of multi-base station optical storage micro-network system for multi-body joint investment in the construction of the micro-network system.

In order to analyse the necessity and economy of the joint investment approach between the optical storage microgrid operator and the distribution network operator, the results of the optical storage microgrid system configuration of Scenario 1 and Scenario 2 are analysed. The results of the operational benefits of Scenario 1 and Scenario 2 are shown in Table 3. It can be seen that Scenario 2 is configured with higher capacity of both PV system and energy storage system, in which the PV system is configured to the maximum capacity due to the limitation of site area. As the photovoltaic storage microgrid operator, as a single investment body, only takes into account its economic benefits, and in order to obtain higher operating returns, it increases the investment cost, and the annual return from selling electricity to the distribution grid reaches 8,220,251,000 yuan. Scenario 2, despite the obvious effect of peak shaving, still makes the distribution grid operator lose 5,848,180,000 yuan per year on average, while the annual operating income of Scenario 1 is a profit of 2,250,560,000 yuan.

Figure 7 shows a dot plot of the daily gains for scenarios 1 and 2, where the gains are negative when they represent over-the-grid expenditures with power purchase and energy sharing, leaving aside investment and O&M costs. The two scenarios gain more during the PV output-rich time period, where the effect is more pronounced in the summer when the PV output is highest. A comparison of the gains per moment in each season between the two graphs reveals that Scenario 2 obtains greater gains. However, Scenario 2 only considers the PV storage microgrid operator as a single investor, sacrificing the interests of the distribution grid operator in order to improve the returns of the PV storage microgrid operator.

Figure 7.

Daily revenue results for Scenario 1 and Scenario 2

Figure 8 shows the comparison of typical summer destination load curves. Comparing the typical summer destination load curves of Scenario 1 and Scenario 2 with the regional load curves, it is concluded that Scenario 2 can achieve good peak shaving during the peak load period due to the large capacity of the configured photovoltaic storage system. But at the same time, it causes a new load trough, which still increases the peak-to-valley difference under the premise of reducing the regional load peak. The peak-valley difference is increased from 24.06 MW to 33.59 MW, which is mainly due to the optical storage microgrid operator’s efforts to maximise the operational revenue.

Figure 8.

Typical daily load curve in summer

The energy storage system discharges a large amount of electricity during the flat and peak hours of the tariff to generate revenue from the sale of electricity, which causes the load curve to fall steeply. Although the peak shaving effect of Scenario 1 is not as obvious as that of Scenario 2, it reduces the regional load variance, making the load curve smoother, which is more conducive to improving the stability of the distribution network, thus ensuring that the optical storage microgrid receives a more reliable power supply. Therefore, the joint investment of the distribution network operator and the optical storage microgrid operator is more conducive to guiding the optical storage microgrid to carry out rational allocation, participate in peak shaving and valley filling of the distribution network, improve the reliability of the distribution network, and at the same time, realise the win-win situation for the interests of all the parties involved in the investment.