Combining reinforcement learning with mathematical programming:An approach for optimal design of heat exchanger networks

Heat integration is important for energy-saving in the process industry.It is linked to the persistently challenging task of optimal design of heat exchanger networks(HEN).Due to the inherent highly nonconvex nonlinear and combinatorial nature of the HEN problem,it is not easy to find solutions of high quality for large-scale problems.The reinforcement learning(RL)method,which learns strategies through ongoing exploration and exploitation,reveals advantages in such area.However,due to the complexity of the HEN design problem,the RL method for HEN should be dedicated and designed.A hybrid strategy combining RL with mathematical programming is proposed to take better advantage of both methods.An insightful state representation of the HEN structure as well as a customized reward function is introduced.A Q-learning algorithm is applied to update the HEN structure using theε-greedy strategy.Better results are obtained from three literature cases of different scales.

Scale-up and thermal stability analysis of fluidized bed reactors for ethylene polymerization

A set of hydrodynamic similarity laws is applied to the scale-up of ethylene polymerization fluidized bed reactors(FBRs)under the condensed mode operation.The thermal stability of open-loop controlled FBRs is investigated by the homotopy continuation method.And the Hopf bifurcation point is selected as an index of the thermal stability similarity.The simulation results show the similarity in state variables,operation parameters,the space-time yield(STY),and the thermal stability of FBRs with different scales.Furthermore,the thermal stability behaviors and similarity of the closed-loop controlled FBRs with different scales are analyzed.The observed similar trend of Hopf bifurcation curves reveals the similarity in the thermal stability of closed-loop controlled FBRs with different scaling ratios.In general,the results of the thermal stability similarity confirm that the hydrodynamics scaling laws proposed in the work are applicable to the scale-up of FBRs under the condensed mode operation.

Modelling and simulation of two-bed PSA process for separating H2 from methane steam reforming

Methane steam reforming is the main hydrogen production method in the industry. The product of methane steam reforming contains H2, CH4, CO and CO2 and is then purified by pressure swing adsorption(PSA) technology. In this study, a layered two-bed PSA process was designed theoretically to purify H2 from methane steam reforming off gas. The effects of adsorption pressure, adsorption time and purgeto-feed ratio(P/F ratio) on process performance were investigated to design a PSA process with more than99.95% purity and 80% recovery. Since the feed composition of the PSA process changes with the upstream process, the effect of the feed composition on the process performance was discussed as well.The result showed that the increase of CH4 concentration, which was the weakest adsorbate, would have a negative impact on product purity.

Acoustic emission detection of slag performance in coal gasifiers

This work first investigated the detection of slags,slag pool liquid level,and slag accumulation height in laboratory scale based on acoustic emission(AE)detection,and further tried the feasibility of this method in an industrial scale coal gasifier.Results show that the energy and variance of acoustic signals can realize the accurate detection of large slag(criterion:E>1.5 E0,S>1.2 S0),and the average relative error is only 0.28%.The acoustic energy in the frequency range of 20–40 k Hz is defined as the characteristic energy,which can realize the accurate detection of slag accumulation height and slag pool liquid level,and the average relative error is only 3.94%.Furthermore,AE detection also realize accurate detection of large slag in an industrial scale gasifier and the acoustic signals at slag screen can be used to realize the early warning of the slag collapse(5 h earlier).

Polar-Coded Modulation Based on the Amplitude Phase Shift Keying Constellations

In this paper, we investigate the polarization effect of the amplitude phase shift keying(APSK) constellations. We find that the polarization effect of the APSK constellations is affected by the bit mapping and the bit loading. Traditionally, the Gray mapping is usually adopted in APSK constellations. Based on the given Gray mapping, we firstly propose the bit interleaved coded modulation polar-APSK(BICM-PA) scheme, which neglects the correlations between the bit levels by using the bit interleaver. In the BICM-PA scheme, a new metric called cumulative Bhattacharyya parameter is introduced to optimize the bit loading of the APSK constellations under the Gray mapping. Second, the multilevel coded modulation polar-APSK(MLCM-PA) is proposed to further improve the performance. A twostage optimization approach is adopted to select the bit mapping and the bit loading in the MLCM-PA scheme. The semi-set partitioning mapping is introduced to achieve a better system performance in the MLCM-PA scheme. Simulation results verify the effectiveness of all the proposed metrics. In addition to this, BICM-PA and MLCM-PA both outperform the coded modulation Turbo-APSK scheme up to 1 dB.

Scaling Factor Optimization of Turbo-Polar Iterative Decoding

In this paper, we investigate the weighted iterative decoding to improve the performance of turbo-polar code. First of all, a minimum weighted mean square error criterion is proposed to optimize the scaling factors(SFs). Secondly, for two typical iterative algorithms,such as soft cancellation(SCAN) and belief propagation(BP) decoding, genie-aided decoders are proposed as the ideal reference of the practical decoding. Guided by this optimization framework, the optimal SFs of SCAN or BP decoders are obtained. The bit error rate performance of turbo-polar code with the optimal SFs can achieve 0.3 dB or 0.7 dB performance gains over the standard SCAN or BP decoding respectively.

高性能聚乙烯产品设计

聚烯烃是国民经济建设和日常生活必不可少的重要基础材料和战略物资.目前,我国聚烯烃行业严重依赖进口,高端聚烯烃产品占进口量的一半以上,行业发展严重受制于人.金属有机催化理论的迅猛发展极大地增强了对聚合物分子结构的调控能力,使人们有可能设计出更多性能优异的聚烯烃新产品.但是,如何实现聚烯烃新产品的工业生产,并使之转化成为商品,是制约聚烯烃产品高性能化的瓶颈,涉及化学工程和高分子科学的深度交叉.本文分别从催化剂结构设计、非均相催化剂结构解析、聚合反应工艺以及产品加工工艺的层面介绍了当前聚乙烯高性能化的重要进展;重点介绍了从活性中心聚合微环境设计出发,通过与催化剂聚合行为的匹配,开发的高性能聚乙烯产品制备新工艺.利用烃类冷凝液的喷射蒸发,在单一流化床反应器中构建温度、浓度差异化的持液颗粒聚合和气相聚合共存的复合流型,打破流化床固有的温度浓度均一性特点,相继开发了超低/极低密度聚乙烯高端产品数万吨,率先生产热收缩膜、拉伸套管膜、双向拉伸聚乙烯膜等高端产品;利用催化剂在惰性组分纳微气泡富集区域的穿梭和碰撞,在生长的聚乙烯颗粒周围营造反应组分的传质开关,使分子链生长和结晶有节奏地交替进行,降低了初生态聚乙烯的链缠结程度,开发了易加工的超高分子量聚乙烯产品.

Important mesoscale phenomena in gas phase fluidized bed ethylene polymerization

Gas phase fluidized bed processes have been widely applied to polyethylene production.In these processes,the flow,mass transfer,and reaction rate on the microscale and macroscale are strongly coupled because of the multiphase and multiscale nature of the fluidization system.Understanding mesoscale phenomena is therefore essential to the quantitative translation of the knowledge obtained from the microscale to the macroscale.This paper reviews the development of ethylene polymerization gas phase processes while focussing on studies regarding mesoscale phenomena.These include experimental characterizations,mathematical modelling and control strategies.Trends and future developments in this field are also discussed.

Hydrodynamics characterization of a choanoid fluidized bed bioreactor used in the bioartificial liver system： Fully resolved simulation with a fctitious domain method

Choanoid fluidized bed bioreactors （CFBBs） are newly developed core devices used in bioartificial liver- support systems to detoxify blood plasma of patients with microencapsulated liver cells. Direct numerical simulations （DNS） with a direct-forcing/fictitious domain （DF/FD） method were conducted to study the hydrodynamic performance of a CFBB. The effects of particle-fluid density ratio, particle number, and fil- ter screens preventing particles flowing out of the reactor were investigated. Depending on density ratio, two flow patterns are evident： the circulation mode in which the suspension rises along one sidewall and descends along the other sidewall, and the non-circulation mode in which the whole suspension roughly flows upward. The circulation mode takes place under non-neutral-buoyancy where the particle sedimentation dominates, whereas the non-circulation mode occurs under pure or near-neutral buoy- ancy with particle-fluid density ratios of unity or near unity. With particle-fluid density ratio of 1.01, the bioartificial liver reactor performs optimally as the significant particle accumulation existing in the non-circulation mode and the large shear forces on particles in the circulation mode are avoided. At higher particle volume fractions, more particles accumulate at the filter screens and a secondary counter circulation to the primary flow is observed at the top of the bed. Modelled as porous media, the filter screens play a negative role on particle fluidization velocities; without screens, particles are fluidized faster because of the higher fluid velocities in the jet center region. This work extends the DF/FD-based DNS to a fluidized bed and accounts for effects from inclined side walls and porous media, providing some hydrodynamics insight that is important for CFBB design and operation optimization.