Assessment of recoverable oil and gas resources by in-situ conversion of shale-Case study of extracting the Chang 73 shale in the Ordos Basin

Lin-Hu Hou ，Xi Luo ，Sen-Hu Lin ，Yong-Xin Li ，**，Li-Jun Zhng ，Wei-Jio M

a Research Institute of Petroleum Exploration &Development，PetroChina，Beijing，100083，China

b National Energy Tight Oil &Gas R&D Center，Beijing，100083，China

c School of Geosciences，China University of Petroleum，Qingdao，Shandong，266580，China

Keywords:Shale oil In-situ conversion processing Thermal simulation experiment Hydrocarbon generation mechanism Hydrocarbon resource Ordos basin

ABSTRACT The purpose of this study is to investigate the entire evolution process of shales with various total organic contents (TOC) in order to build models for quantitative evaluation of oil and gas yields and establish methods for assessing recoverable oil and gas resources from in-situ conversion of organic matters in shale.Thermal simulation experiments under in-situ conversion conditions were conducted on Chang 73 shales from the Ordos Basin in a semi-open system with large capacity.The results showed that TOC and Ro were the key factors affecting the in-situ transformation potential of shale.The remaining oil and gas yields increased linearly with TOC but inconsistently with Ro.Ro ranged 0.75%-1.25%and 1.05%-2.3%，respectively，corresponding to the main oil generation stage and gas generation stage of shale in-situ transformation.Thus a model to evaluate the remaining oil/gas yield with TOC and Ro was obtained.The TOC of shale suitable for in-situ conversion should be greater than 6%，whereas its Ro should be less than 1.0%.Shales with 0.75% (Ro) could obtain the best economic benefit.The results provided a theoretical basis and evaluation methodology for predicting the hydrocarbon resources from in-situ conversion of shale and for the identification of the optimum“sweet spots”.The assessment of the Chang 73 shale in the Ordos Basin indicated that the recoverable oil and gas resources from in-situ conversion of organic matters in shale are substantial，with oil and gas resources reaching approximately 450×108 t and 30×1012 m3，respectively，from an area of 4.27×104 km2.

1.Introduction

With the future increase in the global population and economy，the energy demand will grow substantially in the next few decades(Ma et al.，2020a).The global oil consumption has already risen from 41.43×108t in 2008 to 46.62×108t in 2018 (BP，2019).Energy security has drawn the attention of many countries to their oil and gas supply.The successful development of shale oil and gas in the US has reshaped the energy landscape in the world (Jia et al.，2012;Zendehboudi and Bahadori et al.，2017;Yang and Zou et al.，2019;Zhao et al.，2020).In 2019，the crude oil production reached 44.65 billion barrels in the US，of which，29.14 billion barrels were from shale oil，accounting for 65.12%of the total oil production that year(EIA，2020-01).Shale oil is playing an increasingly important role in the dramatic growth of American oil production(EIA，2018，2019a，2019b;Zhou et al.，2019;Zou et al.，2019) and exerts a significant influence on the global energy and geopolitical layout.

The technologies of horizontal wells and multi-stage fracturing have driven the rapid growth of oil and gas production from shales with medium-high maturity in the US(IHS，2016;Jin et al.，2019a;Perrin，2019-06-06).Consequently，China has attached great importance to shale oil and gas(Zou et al.，2010;Lu et al.，2012;Xue et al.，2015;Du et al.，2019;Ju et al.，2020;Kuang et al.，2020;Song et al.，2020).Surveys indicate that two types of shale，i.e.，lacustrine shale and marine shale，are developed in China.Marine shales are mainly distributed in South China with relatively high maturity，in which the vitrinite reflectance (Ro) is generally greater than 2.0%(Zou et al.，2011，2015;Zhao et al.，2012，2016;Nie et al.，2016).These areas contain abundant shale gas resources and some have been developed successfully with a production of 150×108m3in 2019.However，lacustrine shales are primarily distributed in North China，including the Ordos，Songliao，Bohai Bay and Junggar Basins，covering an area of more than 20×104km2(Dang et al.，2015;Zhao et al.，2018a，b;Yang and Jin，2019;Zou et al.，2020) and characterized by relatively low maturity for which the Rois normally less than 1.0%.China has been attempting to exploit the shale oil in its lacustrine formations for more than a decade by drilling nearly 500 horizontal wells.However，the economic development of shale oil with low and medium maturity has not yet been realized，and the undeveloped organic pores，poor connectivity，low gas-oil ratio(GOR) and high viscosity of oil are the main reasons，all of which lead to an extremely low recovery factor of only 1%-3%(Saif et al.，2017;Jin et al.，2019b;Hou et al.，2020a) by simply relying on volume fracturing in horizontal wells.Prospecting history in the US also shows that conventional volume fracturing in horizontal wells is unable to realize the economic development of shale oil with low to medium maturity.Instead，in-situ conversion must be used to achieve economic development.Shell，ExxonMobil，Total，and many other giant oil companies have conducted relevant research and established four types of in-situ heating technologies，including electric heating，convection heating，radiant heating，and combustion heating(Wang et al.，2013).

In-situ conversion of shale oil，including physical and chemical process，enables the organic matter(including solid organic matter and asphalt)in shale or oil shale to be rapidly converted into light oil and natural gas by in-situ heating.It is then exploited，while coke and other residues are left in the subsurface.This process can be regarded as an“underground refinery”(Zhao et al.，2018a，b).Insitu transformation of shale has four technical advantages:(1) the oil obtained through the in-situ heating and conversion is of good quality(35-49°API);(2)the overpressure fluid and micro-fracture network generated by the in-situ conversion process increase the driving force，the permeable channels，and the drainage system，resulting in an ultimate recovery surpassing 60%;(3) there are no tailings or waste produced，with little groundwater pollution and no hydraulic fracturing needed，which attains actual clean mining by minimizing the damage to the ecosystem caused by harmful byproducts during conventional production;and (4) the occupied area is very small，which helps to protect farmland.The in-situ conversion technologies can successfully tackle the challenges in exploiting the shale with low-medium maturity，and lead to efficient and environmentally friendly utilization (hereafter we use shale to represent both the shale and oil shale).Nonetheless，some key issues must be addressed to achieve commercial development of the shale oil with low-medium maturity，including but not limited to criteria for the selection and evaluation of the recovery targets.To address these issues，we took the shale of the 3rd section in the Chang 7 member of the Tertiary in the Ordos Basin(hereafter referred to as the Chang 73shale) as an example to simulate the hydrocarbon generation and expulsion process under in-situ conditions with a semi-open system.The purpose was to establish an approach with which to evaluate the recoverable oil and gas resources from in-situ conversion of shale，to provide a theoretical basis and evaluation methodology for predicting the hydrocarbon resources from in-situ conversion of shale，and to identify the optimum “sweet spots.”

2.Samples and experimental conditions

2.1.Samples

The Chang 73shale developed in the Late Triassic sediment in semi-deep to deep lacustrine facies.It is characterized by continuous sedimentation with large thickness，high organic matter abundance(average total organic content(TOC)is 14.32 wt%)，and a good organic matter type (TypeII kerogen (Ma et al.，2020b;Hou et al.，2020b)).A series of hydrocarbon generation and expulsion thermal simulations were conducted by analyzing various Chang 73shale outcrop samples with different TOC values.Altogether，nine samples were acquired from 5 m below the surface (to avoid the effects of epigenesis，such as weathering and leaching) in the southeast of the Ordos Basin(Fig.1).The Roof the samples was less than 0.5%.The samples were crushed into 40-60 mesh size，mixed，and divided evenly into several parts.Then，some samples were randomly chosen for TOC，Rock-Eval，and vitrinite reflectance (Ro)tests to obtain fundamental geochemical information(see Table 1).

Table 1Basic geochemical data of the original unheated shale samples.

Table 2Characteristics of the residual samples after pyrolysis experiments.

2.2.Experimental device and process

Pyrolysis experiment is one of the most direct and efficient ways to investigate the oil and gas generation process.Previous researchers have conducted a large number of simulation studies by changing experimental conditions，including the sample feed(from 10 mg (You et al.，2019) to 200 g (Tang et al.，2015))，heating rate(from 0.5 to 25°C/s to 12°C/h (Doan et al.，2013;Lan et al.，2015;Siramard et al.，2017)，pyrolysis temperature (max temperature from 420°C (Siramard et al.，2017;Ma et al.，2020c) to 1000°C(Shao et al.，2019))，and pressure (0.06-68 MPa (Lan et al.，2015;Sun et al.，2019;Yan et al.，2019)).Based on these studies，kinetic models have been built of hydrocarbon generation of different types of organic matter (Han et al.，2014)，impacts of temperature and pressure on hydrocarbon generation have been discussed，and it was concluded that a rapid heating rate reduces oil generation，increase gas generation(Siramard et al.，2017)，and exerts very little influence on the kinetic parameters (Bai et al.，2015).Thermal simulation experiments deepen our understanding of the hydrocarbon generation process.However，they also have disadvantages like other technologies.For instance，the error in assessing oil and gas generation can increase if the sample feed is small，and the heating rate is rapid.Additionally，there has not been much discussion regarding the oil and gas expulsion process constrained by experimental instruments.In this paper，thermal simulations with large sample feed and a slow heating rate were designed in a semiopen system using newly developed reactors with a large capacity.A total of 99 experiments were completed with nine samples at 11 temperatures to study the hydrocarbon generation features of the low maturity shale with various TOC values.

The experimental devices are diagramed in Fig.2.The reactor was made of a special alloy，resistant to corrosion by H2S，CO2，and H2.A special gasket and coating were used between the reactor body and cap to ensure long-term operation without leakage.The highest temperature that the reactor could bear was 700°C，the maximum pressure was 40 MPa，the container capacity was 1.36 L，and samples weighing approximately 2 kg could be accommodated.The instrument meets the experimental requirements of thermal simulation of large-volume shale under different temperatures and pressures.The reactor has two heating systems，one at the bottom and the other on the side.They can operate simultaneously.Additionally，multiple thermocouples and pressure sensors are equipped to precisely control the temperature and pressure inside the reactor，such that the temperature error is below 1°C and the fluid pressure error is less than 0.1 MPa during experiments.

Fig.1.Shale distribution in the Ordos Basin and the sampling point.

Fig.2.Schematic diagram of the pyrolysis device.(1) autoclave，(2) sealing cover，(3) outer cover，(4) electric heater，(5) thermo couple，(6) pressure meter，(7) piston valve controlled by three-way solenoid valve，(8) needle valve，(9) heating tape，(10) synchronous collector，(11) floating piston，(12) liquid collector，(13) condenser，(14) gas bag，(15)measuring cylinder.

The experimental steps are discussed below.

(1) Sample loading.The crushed and evenly mixed shale samples were loaded to the reactor.

(2) Heating simulation.①Leak test.After sealing the reactor，vacuumed it，filled it with 20 MPa helium，and placed it for leak test.When there was no leakage，released the gas.Repeated the process for 3-4 times，and vacuumized the reactor for the last time.②Heating.5 MPa and 7 MPa were set as the fluid pressure and hydrocarbon expulsion pressure respectively.A total of 11 thermal simulation temperature points were designed (Table 2)，covering the whole process of oil and gas generation.The heating procedure was as follows:the sample was heated from room temperature to the temperature before each set point at the heating rate of 20°C/d (the temperature before the first temperature point is 200°C)，and then the sample was heated to the set temperature at the heating rate of 5°C/d，and kept at this temperature for 10 h.

(3) Simulation，product collection and quantification.①hydrocarbon expulsion.The process of hydrocarbon expulsion was controlled by the three-way solenoid valve.At the beginning，the piston valve was closed and the whole reactor was sealed.When the generated products accumulated continuously in the container and the fluid pressure exceeded the set value，the piston valve would open automatically，and the products were released from the top of the reactor until the pressure dropped to the preset value.After that，the piston valve would close automatically and the reactor would be closed again.②Collection and quantification of pyrolysis products.When products were discharged，their release rate was controlled by the needle valve，so that the pressure in the container was not reduced rapidly，and the synchronous collector could also play a buffer role，so as to ensure that the oil and gas could be separated in time when they were generated and discharged in large quantities.The hydrocarbon expulsion pipeline in the apparatus was wrapped with a heating belt，and the temperature of the heating belt was synchronized with the temperature in the reactor to prevent the condensation of products from blocking the pipeline.When the product passed through the condenser (water circulation at 20°C)，water and hydrocarbon liquid(including C5) were collected in the liquid collector.Hydrocarbon gas(C1-C4) and non-hydrocarbon gas were collected in the air bag immersed in the water tank.The volume of the gas bag was determined by the volume of water discharged from the expansion of the air bag.Comparing the mass or volume of produced oil or gas before thermal simulation，the oil and gas yields of shale sample could be obtained.③Collection and quantification of retained oil and solid residual sample.After the autoclave cooled down to room temperature，the simulated source rock sample was taken out，weighed and extracted with dichloromethane，which was regarded as retained oil.The gas obtained by purging was the retained gas(the retained oil and gas were not discussed here，as they were not involved in our product modelling).The residual samples were tested for TOC and Roafter dichloromethane extraction (Table 2).

(4) Reliability analysis of simulation results.Theoretically，the mass of the reactants before and after the thermal simulation was conserved.Nevertheless，measurement errors and equipment leakage may cause product loss and result in disequilibrium.The mass balance rate，the ratio of the product mass after simulation(including oil，gas，water，and residuals)to that of the shale sample before simulation，was used to evaluate the reliability of the experimental results(Equation (1)).

where mass balance is the mass balance rate，residual rock is the mass of the residual sample after simulation，oil is the mass of the produced and remaining oil after simulation，water is the mass of the produced and remaining water after simulation，gas is the mass of the produced and remaining gas after simulation，and unheated rock is the mass of the sample in the reactor before simulation.

The mass balance rate should be no less than 99.5% if there is only a measurement error.Otherwise，leakage occurred during the simulation and the results are not reliable.Under such circumstances，the thermal simulation must be conducted again.In this study，the mass balance rates of all 99 simulations exceeded 99.75%(Fig.3)，indicating the results of the pyrolysis experiments were reliable.

3.Experimental results

3.1.Relationship between pyrolysis and Ro

Thermal maturity is an important parameter with which hydrocarbon generation and expulsion (Tissot and Welte，1984) can be assessed.Of the various evaluation indexes，Rois the only indicator put forward as an international measurement standard(Stach et al.，1982) and is widely used.The Roof the residual samples at each simulation temperature was analyzed，and the results indicated that there were very highly positive correlations between Roand temperature was analyzed，and the results indicated that there were very highly positive correlations between Roand the thermal simulation temperature (Fig.4).Furthermore，the measured Rovalues of different samples at the same temperature were very close.Therefore，the average Roof the nine samples at each temperature was used to establish the model to express the relationship between thermal simulation temperature and Ro(Equation(2)).

where Rois vitrinite reflectance，T is thermal simulation temperature，and a1and a2are empirical coefficients，equal to 0.1380 and 0.0057，respectively.

3.2.Characteristics of hydrocarbon yield from in-situ conversion of shale

Fig.5 shows the yields of oil and gas with increasing Ro.Generally，the oil and gas yields of shale increased linearly with TOC，that was，the higher the abundance of organic matter(TOC)in source rock，the higher the yields.For example，the oil and gas yields of the sample with TOC of 25.988 wt% are 66.85 mg/g·rock and 33.30 mL/g·rock at Ro=3.65%.However，the variation trend of oil and gas yields were different from that of Ro.For oil yield，it increased rapidly at the beginning of thermal evolution and reached the peak of oil generation at Ro?1.2%，which was similar with that of Lucaogou shale (Hou et al.，2021)，another typical lacustrine type-II shale in China.After that，it almost remained the same.For gas yield，the value first increased rapidly with Roand then increased slowly，showing that it was affected by the cracking of crude oil and wet gas during high evolution stage.However，for samples with TOC ≤13.34 wt%，the gas production increased rapidly before Ro?1.67%，and then increased very slowly.While for the two samples with TOC＞20 wt%，the trend of increasing gas production was very obvious before Ro≈2.4%，and then it showed a gentle increasing trend.

4.Discussions

4.1.Characteristics of the remaining hydrocarbon yield

Fig.3.Evolution of mass fractions of major products with increasing thermal maturity.

Fig.4.Correlation between pyrolysis temperature and Ro.

Fig.5.Yields of oil and gas with increasing Ro.(a) yield of oil，(b) yield of gas.

Fig.6.Yield of the remaining hydrocarbons with increasing Ro.(a) remaining yield of oil，(b) remaining yield of gas.

Fig.7.Yield of remaining hydrocarbons with original TOC.(a) remaining yield of oil，(b) remaining yield of gas.

Fig.8.Yield of remaining hydrocarbons with residual TOC.(a) remaining yield of oil，(b) remaining yield of gas.

The hydrocarbons generated but retained in shale，together with the organic matters that were not yet degraded constituted the major source of oil and gas during the in-situ conversion of shale.Subsurface shales have generally undergone thermal evolution.Some oil and gas may have been generated and discharged already.Therefore，in this paper，the hydrocarbon yield from in-situ conversion of shale refers to the remaining hydrocarbon yield.With this in mind，we have estimated the hydrocarbon yield from in-situ conversion of shale by calculating the obtained oil and gas at each experimental temperature (yields of the remaining oil and gas are plotted in Fig.6).It shows that the remaining oil and gas yields exhibit consistent trends as Roincreases and both were negatively correlated with Ro，whereas regarding specific performance，these data reduced slowly at first and then decreased dramatically when Roreached a certain value.However，oil and gas have different evolution stages.Regarding the remaining oil yield，it decreased slowly when Roranged from 0.58%to 0.75%，then it dropped almost linearly when Rowas between 0.75% and 1.25% and decreased slowly again until it reached zero when Rowas greater than 1.25%.In contrast，the remaining gas yield declined slowly when Rowas from 0.58% to 1.05% and diminished rapidly in a near power function pattern when Rowas between 1.05%and 2.3%.Additionally，the major oil and gas generation stages were different，and it can be concluded from the yield rates that the stage between 0.75% and 1.25%was the major oil generation period，which accounted for 75%of the total oil yield.However，the amount of oil generated during the high evolution stage(Ro1.7%-2.3%)was small and approached to zero when Roexceeded 2.3%.On the contrary，the major gas generation period was 1.05%-2.3%.The remaining gas yield declined slowly to a constant level when Ro＞2.3%.At this stage，the gas was mainly generated from thermal expansion because of increasing temperature.

TOC was another key factor influencing the potential of shale transformation，as well as Ro.The organic matter provided a material basis for hydrocarbons，and the original TOC in shale determined its ultimate yield of oil and gas.Fig.7 shows that both the remaining oil yield and gas yield increased with the original TOC for the samples with the same Ro，and there were strong linear correlations between the yield and the original TOC (Fig.7).Furthermore，the remaining oil yield and gas yield decreased as Roincreased under the same original TOC，which was consistent with previous findings.

When we evaluated the recoverable petroleum resources from in-situ conversion of shale under geologic conditions，we had to understand that the evaluated objects had generally experienced thermal evolution.Because of this，the organic matter in shale had already been transformed making it difficult to obtain the original TOC value.Therefore，the relationship between the remaining oil/gas yield and the residual TOC was a focus in this paper(Fig.8).The remaining oil yield from shale samples with different TOC values increased as the residual TOC increased when Rowas below 1.7%and changed consistently with TOC when Rowas below 1.1%(Fig.8a).The remaining gas yield increased as the residual TOC increased when Roremained the same.However，when TOC was unchanged，the remaining gas yield increased at first and then decreased as Roincreased，and the turning point was at Ro=1.1%(Fig.8b).This feature could be attributed to the evolution process because first，the organic matter generated oil and gas as the temperature rised，then some oil was cracked into coke at high maturity，and these actions caused TOC to decline rapidly at first，then increased slightly，and ultimately remained stable.

Based on the correlations between the remaining oil/gas yield(per mass unit of shale and mass unit of TOC) versus TOC and Ro(Fig.9a)，it could be concluded that，as for the same shale sample，the remaining oil yield increased slightly at first，then decreased steadily and finally diminished rapidly as Roincreased.However，for the remaining gas yield，Roincreased (Fig.9b).

4.2.Models for evaluating the remaining oil/gas yield from in-situ shale transformation

As previously described，the experimental results demonstrated that there was a strong linear correlation between the remaining oil/gas yield and TOC.Accordingly，a model could be built to express this relationship.Some key parameters were extracted and then used to establish links with Ro，such that a model to evaluate the remaining oil/gas yield with TOC and Rocould be obtained (Equation (3)).

where Qpo is the remaining oil yield per mass unit of shale，Rois the vitrinite reflectance，and TOC is the total organic carbon.a1，a2，a3，a4，a5，and a6are empirical parameters，where a1and a6equal 0.99892 and 0.01538，respectively.When Ro≤0.76%，a2and a3equal 0.4265 and 0.7516，when 0.76%＜Ro≤0.95%，a2and a3equal 0.4593 and 1.41，when Ro＞0.95%，a2and a3equal 4.164 and 5.3161，when Ro≤0.77%，a4and a5equal 0.068 and 1.1297，when 0.77% ＜Ro≤1.06%，a4and a5equal 2.6881 and 3.2629，and when Ro＞1.06%，a4and a5equal 3.5488 and 4.1449，respectively.

where Qpg is the remaining gas yield per mass unit of shale，Rois the vitrinite reflectance，and TOC is the total organic carbon.b1to b15are empirical parameters and equal 1.0062，0.9478，0.5744，-0.0997，-1.1745，3.4118，2.1756，1.5235，-2.3651，-0.2334，2.9012，-2.9174，-0.0967，0.5035，and -0.4776，respectively.

To test the confidence of the models，the remaining oil and gas yields were estimated by Equations (3) and (4) using TOC and Rodata obtained from this study.Through comparisons with experimental results，the correlation coefficients of both the oil and gas yields were greater than 0.997 (Fig.10).Specifically，the average absolute error between the calculated remaining oil yield and the experimental yield was 0.0001 mg/g·rock，and the absolute values of the absolute errors averaged 0.4796 mg/g·rock.The average absolute error between the calculated remaining gas yield and the experimental one was-0.0002 m3/t·rock，and the absolute values of the absolute errors averaged 0.3707 m3/t·rock.Such figures showed that the models built in this paper were reliable.

Fig.9.Yield of remaining hydrocarbons with TOC and Ro.(a) remaining yield of oil，(b) remaining yield of gas.

Fig.10.Relationship between calculated yield and experimental yield.(a) oil yield，(b) gas yield.

Fig.11.Relationship of the cutoffs of the remaining oil yield versus Ro and TOC.

Fig.12.Relations of the remaining oil yield cutoff versus minimum TOC cutoff and Ro

4.3.Cutoffs of TOC and remaining oil yield

Shale to be developed by in-situ transformation into oil have normally experienced some thermal evolution，and thus，have different remaining oil yields.Furthermore，shale with commercial development value also differs in TOC.Taking the shale samples used in our experiments as an example，the cutoffs of the remaining oil yield of shale were 50 mg/g·rock，45 mg/g·rock，40 mg/g·rock，35 mg/g·rock，30 mg/g·rock，25 mg/g·rock，20 mg/g·rock，and 15 mg/g·rock，respectively.The TOC cutoffs under different Rovalues could be obtained by using Equation (1).The results are shown in Fig.11.When Roremained unchanged，the TOC cutoff declined as the yield increased.When the cutoff of the remaining oil yield remained the same，the TOC cutoff declined slightly at first and then increased as Roincreased.The minimum value of the TOC cutoff occurred when Rowas approximately 0.75%.Therefore，the shale with Roof approximately 0.75% offered the best commerciality for development by in-situ transformation，and it was untrue that lower Rowas better.The reasons might be that after oil and/or gas are generated，they are absorbed first by organic matter(Baker，2016;Tissot et al.，1971;Stainforth and Reinders，1990;Larsen and Li，1997;Ertas et al.，2006;Kelemen et al.，2006;Han et al.，2017)and minerals (Schettler and Parmely，1991;Li et al.，2020).Only when the amount of the generated hydrocarbons is large enough，can microfractures be formed within the source rocks because of increased pressure (Jarvie，2012)，and the oil and/or gas generated from kerogens can then move and be discharged from the source rocks.In summary，the source rock needs to reach a certain maturity to generate enough oil and gas and form an effective hydrocarbon expulsion channel.At this time，the oil and gas generated during in-situ transformation can be directly exploited from shale.As a result，the time to produce oil and/or gas is advanced，which helps to shorten the investment recovery period and improve project economics.

Fig.13.Relationship between the maximum recoverable oil volume and the maximum recoverable gas volume of the shales.

Fig.12 plots the relationship of the remaining oil yield cutoffs versus minimum TOC cutoffs and Ro，where it can be observed that the minimum TOC cutoff increased as the remaining oil yield cutoff increased，and there was a strong linear correlation between them.In contrast，the Rocorresponding to the minimum TOC cutoff decreased as the cutoff of the remaining oil yield increased.For instance，when the cutoffs of the remaining oil yield were 15 and 50 mg/g·rock，respectively，the Rovalues corresponding to the minimum TOC cutoff were 0.76% and 0.74%，respectively.

5.Assessment of recoverable oil and gas resources in Chang 73 shale，Ordos Basin

5.1.Geological settings and exploration

Fig.14.Scheme of TOC low limit vs Ro.

The Ordos Basin is located in the central part of North China(Fig.1)，covering an area of approximately 37×104km2and contains 146.5×108t and 15.68×1012m3of conventional oil and gas resources，respectively.The Triassic Yanchang Formation.and the Jurassic Yan’an Formation are the major oil-bearing series in the basin.The oil source rocks belong to the Yanchang Formation deposited in semi-deep to deep lacustrine facies (Zhang et al.，2006;Li et al.，2019).Specifically，organic-rich shale is mainly developed in the Chang 73section.During the sedimentary period of this section，the maximum water depth of the ancient lake basin could reach 60-120 m，and the distribution area of organic-rich shale is approximately 5×104km2.By the end of 2019，the total resources in the Ordos Basin were more than 5.6 trillion tons，and the annual oil production was more than 36 million tons.The reservoirs of the Yanchang Formation are tight with low permeability.The average porosity and permeability were 9.9% and 1.23 mD，respectively.Approximately 81.6% of the resources were in reservoirs with a permeability lower than 5 mD，among which 54.3%of the reservoirs featured permeability lower than 1 mD.The estimated ultimate recovery factor of the resources was 15.8%.

The Chang 73shale has low maturity with a Rovalue average of 0.82%.Tests indicate that it is impossible to realize commercial development using the existing volume fracturing of horizontal wells.The initial oil rates of the seven vertical wells were only 0.60-3.23 t/d after volume fracturing and declined rather rapidly.Taking Well G295 with the slowest decline rate as an example，the daily oil production was 1.09 t/d after producing for 1 year，and the annual oil production was 415.03 t.After producing 912 d，its cumulative oil production was only 623.6 t.All other wells were shut after producing 1-8 months.Estimation of GOR showed that the peak GOR of the Chang 73shale ranged from 60 m3/m3to 98 m3/m3，and averaged 83 m3/m3.There was liquid petroleum in the shale.In-situ conversion was feasible，although the existing technologies could not realize commercial development.

5.2.Determination of key parameters

The methods for determining the key parameters，such as TOC，Ro，and net pay of shale for assessing the recoverable oil/gas resources from in-situ conversion of shale are introduced in this paper using the Chang 73shale in the Ordos Basin as an example.

5.2.1.Determination of TOC cutoff

To evaluate whether or not the target is economical，it is necessary to determine the cutoffs of recoverable oil and gas volumes of the shale.Because there was a strong linear correlation between the maximum recoverable oil volume and the maximum recoverable gas volume (Fig.13)，it was enough to determine only the cutoff of the recoverable oil volume.

During the commercial development via in-situ shale transformation，the cutoff of recoverable oil volume could be determined with a model (Equation (5)) based on the oil production cutoff of a well group and the mass of the shale in the effective heating area controlled by this group.

Where Qpo_limitis the cutoff (or lower limit) of recoverable oil volume per mass unit of shale，Qoil_limitis the lower limit of the cumulative oil produced by a well group，and Wtrockis the mass of shale in the effective heating area controlled by the well group.

Horizontal wells were used in this paper.Specifically，there were 10 heating wells and one production well in the well group.The spacing between two heating wells was 15 m，and the lateral length of a horizontal well was 1，200 m.Based on these data，the rock mass in the effective heating area was 7，080，000 tons，the cutoff of the oil production was 100，000 tons，and the cutoff of recoverable oil volume of shale was 14 mg/g·rock as estimated with Equation (5).

Fig.15.Distribution of TOC in Chang73 of Triassic in Ordos basin.

The TOC cutoff corresponding to the cutoff of recoverable oil resources per mass unit of shale could be determined with Equation (3) by using the cutoff of recoverable oil volume determined above (Equation (6)).

where TOClimitis the TOC lower limit (i.e.，cutoff) of the effective shale，Qpo_limitis the lower limit of recoverable oil volume per mass unit of shale，Rois the shale to be tested，and c1，c2，c3，c4，c5，and c6are empirical parameters.c1and c4are 1.0011 and -0.0154，whereas c2and c3are 0.068 and 1.1297 when Ro≤0.77%，and c2and c3equal 2.6881 and 3.2629 when 0.775%＜Ro≤1.06%，and c2and c3are 3.5488 and 4.1449 when Ro＞1.06%.

Regarding c5 and c6，they were 0.4265 and 0.7516 when Ro≤0.76%，equaled 0.4593 and 1.41 when 0.76% ＜Ro≤1.0%，and they are 4.164 and 5.3161 when Ro＞1.0%.

Assuming the cutoff of recoverable oil volume per mass unit of rock is 14 mg/g·rock，then the plot of Roversus the TOClimitof the effective shale(Fig.14)could be obtained using Equation(5)，from which it can be seen that the cutoff of TOC is approximately 6%when Rois approximately 0.8%.

5.2.2.Determination of other key parameters

5.2.2.1.TOC.The analysis of 8690 cores taken from the Chang 73 shale from 271 wells indicated that their TOC ranged from 6 to 39 wt%，and averaged 14.32 wt%.In this paper，TOC and Ro obtained from the laboratory tests were used to calibrate the log data using the ΔlogR method(Passey et al.，1990).Then，the TOC values of the target intervals in 791 wells were obtained by the interpretation model built by zoning.It should be noted that the TOC used here refers to the average TOC of the shale intervals satisfying the TOC cutoff criteria.To test the reliability of the calculation results，referring to the standard of effective shale section，the area and thickness weighted statistics of TOC values in the control area of each well point showed that the variation range was from 6 to 32.4 wt%，and averaged 14.12 wt%，which was very close to the value obtained from laboratory cores.This result indicated that the calculations were reliable.As Fig.15 shows，the Chang 73 shale with TOC greater than 6 wt% covered an area of approximately 5.1×104 km2，and the organic-rich region (TOC＞18 wt%) was mainly distributed in Zhengning-Miaowan，Ningxian-Jingchuan，Heshui-Qingcheng，Maling-Huachi-Changguanmiao，and Maling-Gengwan-Jiyuan.Overall，organic-rich shale sections were distributed in succession，but there was still strong heterogeneity，which reflected the effects of the paleoenvironment on the development of high-quality source rocks.Nevertheless，the Chang 73 shale had a good material basis for in-situ transformation.

Fig.16.Distribution of Ro in Chang73 of Triassic in Ordos Basin

5.2.2.2.Ro.Altogether，2200 shale samples were taken from 321 wells to test and analyze the vitrinite reflectance of the Chang 73 shale.Based on the analysis results，the Ro distribution map was drawn for the Chang 73 shale.According to the results，the Chang 73 shale experienced uneven thermal evolution.Its Ro was mainly distributed in 0.52%-1.25%，with an average value of 0.82%.Among these，93% of the areas had a Ro less than 1.0% and 54.3% had a Ro less than 0.82%;these were mainly distributed in the south of Qincheng，Yanwu-Wuqi，and Gengwan-Yanchi.According to the previous analysis，shales in the above areas have good hydrocarbon generation capacity and should be used as primary targets for insitu transformation (Fig.16).

Fig.17.Distribution of effective shale thickness in the Chang 73 of the Triassic in Ordos Basin.

5.2.2.3.Net paythickness.The effective shale sections suitable for in-situ transformation were divided using the TOC cutoff of effective shale and the TOC value of logging interpretation to obtain the effective shale thickness and determine the favorable shale distribution area.The principles were as follows.①the TOC of shale was higher than its lower limit value (6 wt%)，and the continuous thickness was greater than 2 m;or ②the cumulative thickness of shale section was greater than 2 m，and the accumulated thickness of TOC less than 6 wt%was less than 20%of the thickness of this section，and the single-layer thickness with average TOC less than 6 wt% was less than 1 m.According to the above principles，the plane distribution of effective shale thickness of the Chang 73member in the study area was drawn using the logging interpretation results of 791 wells(Fig.17).The results showed that the effective thickness of the Chang 73shale ranged from 2 m to 49 m，with an average of 17.5 m.Among these，the distribution area with an effective thickness greater than 20 m exceeded 1.5×104km2was mainly located in Jiyuan-Maling，the east of Wuqi-Nanliang，and Zhengning-Miaowan，where the average thickness of effective shale is as high as 27.5 m.Additionally，from the profile，the spatial distribution of the effective shale of Chang 73was also very stable，especially in the long axis direction of the paleo deep lake semi-deep lake.The effective shale distribution was continuous and stable，and the thickness change was small (Fig.18).In the short axis direction，affected by the ancient landform，the effective thickness of the shale changed substantially，but the main body was still relatively continuous with a large thickness (Fig.19).

5.3.Evaluation method of recoverable oil and gas resources

The recoverable oil and gas resources from in-situ shale transformation were assessed using the following steps:(1)Estimate the remaining oil and gas yield per mass unit of shale through Equations (3) and (4) according to TOC and Rodistribution of effective shale distribution area.(2) Estimate the abundance (or density) of recoverable oil and gas resources in the assessment area by Equations (7) and (8) using the effective shale thickness and shale density (obtained through log data).(3) Estimate the recoverable oil and gas resources based on the control area of evaluation points.

Fig.18.Organic rich shale section of the Chang7 of Triassic in Ordos Basin (NW-SE)(Location is shown in Fig.1).

Fig.19.Organic rich shale section of the Chang7 of Triassic in Ordos Basin (SW-NE)(Location is shown in Fig.1).

Fig.20.Distribution of recoverable oil resource abundance in Chang73 shale of Triassic in Ordos Basin.

where AOR is the abundance of recoverable oil resources in the assessment area，AGR is the abundance of recoverable gas resources in the assessment area，Qpois the oil recovery volume per mass unit of shale within the effective interval，Qpgis the gas recovery volume per mass unit of shale in the effective interval，and ρ is the shale density in the effective interval.

The evaluation results showed that the recoverable oil resources of the Chang 73shale by in-situ conversion were approximately 450×108t，natural gas was approximately 30×1012m3，and the distribution area was approximately 4.27×104km2.The abundance of recoverable oil resources in the effective shale distribution area was 50×104t/km2-595×104t/km2，with an average value of 146×104t/km2.Among these，the area with technically recoverable oil resource abundance greater than 160×104t/km2was approximately 1.55×104km2，mainly distributed in Jiyuan-Gengwan-Maling，Wuqi and Qingcheng-Miaowan (Fig.20)，which were the key areas of the Chang 73shale in-situ conversion.The distribution of recoverable natural gas resource abundance was consistent with this finding.The abundance of recoverable natural gas resources was 3×108m3/km2-33×108m3/km2，with an average value of 8.1×108m3/km2.The area with the resource abundance greater than 8×108m3/km2was approximately 1.66×104km2，mainly distributed in Jiyuan-Gengwan-Maling，Huachi and Qingcheng-Miaowan (Fig.21).It was demonstrated that the Chang 73shale in Ordos Basin has a large amount of recoverable oil and gas resources.If it could be effectively utilized，the crude oil production of in-situ transformation could have a resource base of 100 t/year and stable production for more than 100 years.

6.Conclusions

(1) In this paper，the hydrocarbon generation and expulsion process of shale under in-situ conversion were simulated using a large-capacity reactor，slow heating rate，and a semiopen system.It was revealed that TOC and Rowere the key factors affecting the in-situ transformation potential of shale.The remaining oil and gas yield of shale increased linearly with TOC，and the remaining oil yield increased slightly at first and then declined steadily and finally decreased rapidly as Roincreased;however，the remaining gas yield increased as Roincreased.

Fig.21.Distribution of recoverable gas resource abundance in Chang73 shale of Triassic in Ordos Basin.

(2) Roranged 0.75%-1.25% and 1.05%-2.3%，respectively，corresponding to the main oil generation stage and gas generation stage of shale in-situ transformation.The TOC of shale suitable for in-situ conversion should be greater than 6%，whereas its Roshould be less than 1.0%.Shales with 0.75%(Ro) could obtain the best economic benefit.An evaluation model of the remaining oil/gas yield was established.

(3) The organic-rich intervals in the Triassic Chang 73shale of the Ordos Basin were characterized as being continuous and having a large thickness and high TOC.They were the optimal targets for in-situ conversion.The resource assessment results of Chang 73shale showed that the recoverable oil resources of in-situ transformation were approximately 450×108t，whereas the recoverable gas resources were 30×1012m3，and the distribution area was approximately 4.27×104km2.If it could be used effectively，the crude oil output of in-situ transformation in the Ordos Basin has a resource base of 100 t/year and stable production for more than 100 years.

Acknowledgments

This work was supported by PetroChina Co Ltd.(Grant number:2015D-4810-02;2018YCQ03;2021DJ52) and National Natural Science Foundation of China(Grant number:42172170).We thank the two anonymous reviewers for their valuable suggestions.