The optimal planning model is formulated as Eqs 11–14 to minimize the total annualized net present cost (NPC) of the project, in which the investment cost and total annual operation cost are involved. min C Total = j ( 1 + j ) N ( 1 + j ) N − 1 ∑ y = 0 N C y inv ( 1 + j ) y + C ope + C co 2 , (11) C y inv = ∑ i ∈ Ω device [ c i inv ψ i cap ( y = 0 ) + c i rep ψ i cap ( device   i   is   replaced   in   year   y ) − c i rep ψ i cap l i whole l i rest ( y = N ) ] , (12) C ope = ∑ t = 1 T c t e , buy P t grid , buy Δ t + ∑ t = 1 T c t g , buy F t CHP Δ t + ∑ t = 1 T c t g , buy F t GB Δ t + ∑ i ∈ Ω device c i O&M ψ i c a p − ∑ t = 1 T c t e , sale P t grid , sale Δ t , (13) C co 2 = c co 2 [ ∑ t = 1 T μ CHP co 2 F t CHP + ∑ t = 1 T μ GB co 2 F t GB + ∑ t = 1 T μ grid co 2 P t grid , buy ] , (14) where C y inv is the investment cost in the corresponding year y, ranging from the initial capital cost in year 0, the devices replacement cost at their end of lifespans, and the salvage value at the whole project life cycle end. Ω device represents the device set. c i inv and c i rep denote the unit investment cost and unit replacement cost of the device i, respectively. ψ i cap denotes the capacity of the device i. l i whole indicates the whole lifespan of device i, and l i rest represents the rest lifespan of device i at the end of the project lifetime. C ope is the annual total operation cost that contains the cost of electricity purchase from grid, natural gas consumption cost for CHP and GB service, operation and maintenance (O&M) cost of all devices, and revenues from electricity sells to the grid. c t e , buy is the electricity price at timeslot t. P t grid , buy is the electricity purchase from the grid at moment t. c t g , buy is the natural gas purchase price at timeslot t. c i O&M is the unit O&M cost of device i. In addition, c t e , sale is the feed-in tariff, and P t grid , sale is the electricity sale to the grid at timeslot t. The term C co 2 denotes the total CO₂ penalty caused by CHP and GB natural gas consumption as well as grid emission. c co 2 is the carbon tax. μ co 2 represents the different carbon emission intensity factors of the emitters.

Since the investment and operation are both considered, the constraints should cover the whole investment consideration plus each time interval during the optimization. Above all, the planning capacity of devices is limited by the geographical reality and other factors expressed as φ i cap , min ≤ φ i cap ≤ φ i cap , max , (15) where φ i cap , min and φ i cap , max denote the lower and upper limit of devices’ installing capacity.

The electricity output of each energy source of PV, WT, CHP, electrolyzer, and fuel cell should stay at a certain output range for concerning the safety, reliability, and their internal performance, as formulated in { 0 ≤ P t pv ≤ P t , pv out 0 ≤ P t wt ≤ P t wt , out ( v ( t ) ) P min CHP ≤ P t CHP ≤ P max CHP P min in ,electro ≤ P t in ,electro ≤ P max in ,electro P min FC ≤ P t FC ≤ P max FC , (16) where P min CHP and P max CHP are the minimum and maximum limits of CHP electricity generation. P min in ,electro and P max in ,electro are the upper and lower limits of the power input of the electrolyzer. P min FC and P max FC are the upper and lower limits of fuel cell generation.

There are two kinds of storage being used during the optimization, and a similar change and discharge constraints should be taken into account, as shown in { 0 ≤ P t in , BESS ≤ u ( t ) P max in , BESS 0 ≤ P t out , BESS ≤ ( 1 − u ( t ) ) P max out , BESS E min BESS ≤ E t BESS ≤ E max BESS , (17) { v ( t ) m min , H 2 s , in ≤ m t , H 2 s , in ≤ v ( t ) m m a x , H 2 s , in ( 1 − v ( t ) ) m min , H 2 s , o u t ≤ m t , H 2 s , out ≤ ( 1 − v ( t ) ) m max , H 2 s , out m min , H 2 s ≤ m t , H 2 s ≤ m max , H 2 s , (18) where P max in , BESS and P max out , BESS are the maximum charging and discharging powers of BESS. E min BESS and E max BESS are minimum and maximum energies that can be stored in BESS. m min , H 2 s , in ( m min , H 2 s , out ) and m max , H 2 s , in ( m max , H 2 s , out ) are the lower and upper limits of hydrogen injection (output) to (from) the hydrogen storage tank. m min , H 2 s and m max , H 2 s are the upper and lower limits of stored hydrogen. u ( t ) is a binary variable that ensures the charging and discharging will not happen at the same time, and v ( t ) is set for the same purpose.

The grid purchase and sale capacity constraint is formulated in Eq. 19. Notice that the electricity purchase and sale at the same time is not allowed in the modeling of this article. { 0 ≤ P t grid , sale ≤ P max grid P t grid , buy P t grid , sale = 0 , (19)

The energy balance can be formulated in P t grid , buy + P t pv + P t wt + P t CHP + P t out , BESS = L t electricity + P t grid , sale + P t in , electro + P t in , BESS , (20) H t CHP + H t GB + H t F C = H t heat , (21) where the electricity demand L t electricity and heat demand H t heat should be met by the sum of different energy suppliers in the whole system.

There are groups of required input parameters including annual solar irradiance, wind speed, annual electric load, and heat load, as shown in Figures 2–5. These data are all with a time slot of 1 h. In addition, the technical parameters of candidate energy devices are listed in Table 1. The energy prices in each time period are shown in Table 2. The project lifetime, discount rate, and carbon tax are set to 25 years, 6%, and 50$/ton, respectively.

Four cases are developed for comparison by considering different forms of energy sources subjected to various factors such as geographical resource endowment, actual area, and available technology of the integrated energy systems.

Scenario #1 (base scenario): considering investing CHP as the coupling unit of electricity and heat generation. Thus, the whole electric demand is catered by utility grid supply and CHP electricity generation whereas the heat demand is supplied by the CHP heat generation and gas boiler.

Scenario #2: assuming the area is endowed rich resources of solar energy but lack resources of wind energy. The electric load is supplied by PV, BESS, CHP, and the utility grid, and the CHP and gas boiler are the two heat demand suppliers.

Scenario #3: on the basis of scenario #2, assuming that there is enough land for facility construction and abundant resources endowed of wind energy in this IES area, and WT has been added as a new electricity supplier. The rest structure remains the same as Scenario #2.

Scenario #4: based on Scenario #3, hydrogen production and storage are considered in the integrated energy system. Similarly, the electric load is supplied by PV, WT, BESS, CHP, and utility grid. At the same time, the hydrogen fuel cell can generate part of the electricity. The heat load is supplied by CHP and the gas boiler.

The final optimal planning results of wind–photovoltaic–hydrogen storage in each scenario are shown in Table 3. The more specific cost breakdown is depicted in Figure 6. It can be observed that with the integration of different resources, the total cost and emissions significantly decreased. The result comparison between scenarios #2 and #3 indicates that though the higher initial investment cost occurred with the WT installation, the annualized total cost still dropped. This is because WT integration alleviates the dependence on grid purchases with a relatively high price during the operation (the proportion of grid operation in O&M cost dropped from 61.21 to 54.01% from the pie chart in Figure 6) and relieves the pressure of CHP electricity generation (less fuel purchase cost). In addition, the total carbon emission reduced about 560 tons due to new clean power supplier integration (i.e., WT). Scenarios #3 and #4 are attempted to compare what effects can be brought from the hydrogen energy system. Results show that the total investment capacity of PV and WT increased, and the corresponding BESS capacity increased. It can be conducted that integrating the hydrogen for carbon-free electricity supply and storage into an IES would stimulate more renewable energy investment and such variability of solar and wind power would be endured by adding a new hydrogen storage and conventional BESS capacity. Consequently, the annualized total cost would rise while total emission reduces.

Despite the fact that hydrogen integration would lead to more carbon reductions, it is not cost competitive from the corresponding optimal result. From Figure 6, it can be learned what exact contribution hydrogen can bring. As seen, the total cost consists of investment, O&M cost of energy devices, fuel cost, and carbon emission cost. From the comparison between scenarios #3 and #4 in O&M cost, it is found that PV and WT operation cost increases from 35.25 to 39.38% of the total operation fee, which indicates that with the HES integration the more renewable generation would be adopted and contribute to much carbon reduction. Inevitably, the initial investment cost together with the O&M cost would increase. From the sole perspective of scenario #4, noticing that the hydrogen power generation and storage cost only occupies less than 4%, implying that the HES could not take over the role as the main carbon-free electricity supplier or electrical energy storage in the whole system. This explains why the HES is not expected to invest on a large scale at present.

Figure 7 shows the carbon emission in each scenario. Carbon emissions in this IES are mainly generated by the grid and gas-fired devices (i.e., CHP and GB). As seen, scenario #1 held the most emission due to the lack of carbon-free technologies. From scenarios #2 to #4, the total carbon emission gradually decreased. Compared to scenario #2, the carbon emission reduced in total in scenario #3 because of the WT investment, which makes grid supply dependence weaker and CHP electricity generation drop more. Thus, the corresponding carbon emissions are reduced from 1,390 and 1,794– to 1,136 tons and 1,490 tons, respectively. The same trend can also be observed from the comparison of scenario #3 and #4. When investing in HES in scenario #4, the total carbon emission would be further reduced owing to its stimulation to PV and WT investment. As depicted, though the carbon emissions from CHP and GB generation are barely unchanged, the emission from the grid reduced by 132 tons, resulting in the total carbon emission reduction, eventually in scenario #4.

Main conclusions that can be drawn from the above results are as follows. Knowing the fact that the integration of HES can certainly reduce carbon emissions, the carbon reduction effect does not depend on hydrogen power generation via the fuel cell itself but also on HES’s positive impact on promoting more renewable energy generation investment. In other words, integrating an HES can be considered as a great motivation to invest more PV panels and WTs; hence, the carbon reduction is realized.

To specifically learn about the economy during the whole project lifetime of each scenario, the life cycle–based cost–benefit analysis is carried out in this section. The cash flows based on the net present value (NPV) of each 4 scenarios are shown in Figure 8. It is observed that in the initial stage (year 0), the investment cost from scenario #1 to scenario #4 increased due to more energy devices being decided to install. Though it takes much initial investment cost in scenarios #2, #3, and #4, the situation reversed after the year 0. The operation costs in each year of scenarios #2, #3, and #4 are all lower than those in scenario #1. From the sight of the cumulative cash flow shown in Figure 9, it can be learned that breakeven points occurred in year 8, 9, and 11 for scenarios #2, #3, and #4 compared to scenario #1, which indicates that the investment in scenarios #2, #3 and #4 are starting to become cost competitive from then on.

There is no doubt that more types of new energy device integration do help achieve some economical goals in IES investment from the final cost–benefit analysis results. However, the ideal investment capacity is not easy to find. The three breakeven points in Figure 9 have proved that the high initial investment cost would certainly delay the final payback years. The cumulative cash flow curve in scenario #4 is always seated above than that in scenarios #2 and #3 and with no trend to drop, informing that whether trying to invest in numerous new decarburization technologies to pursue the environmental benefits would face the dilemma of the poor economy of the overall project lifetime.